That gap is no longer theoretical. Recent research shows 59% of organizations reporting bandwidth issues, up from 43% a year earlier. Latency challenges have nearly doubled, rising from 32% to 53%. Those numbers point to a broader problem: enterprises are underinvesting in the network conditions required by AI, then paying for the gap through poor performance, damaged user experience and unnecessary spending on the wrong parts of the technology stack.

Key facts

• 59% of organizations report bandwidth issues, up from 43% the previous year.
• Latency problems have risen from 32% to 53%.
• A typical AI agent can generate up to 450% more total network traffic per task than a human performing equivalent work.
• Local pilots can perform well when users are close to cloud regions and bandwidth is abundant, then deteriorate when deployments go global.
• Bandwidth is about volume; latency is about time, and latency is the harder challenge for chained AI workflows.
• Compute and GPU availability are rarely the primary bottleneck when end-to-end performance fails.

Why AI failures often start below the model

An AI conversation from two years ago and one today would be almost unrecognizable. The market has moved from asking whether the technology is real to declaring full commitment and demanding that it work at scale. During that shift, most of the attention has gone to model structures, model choice and model performance. But a model is only as effective as the infrastructure connecting it to users, applications, agents and the other models on which it depends.

Traditional enterprise networks were built around a predictable human transaction. A person searches for something, the system fetches it, the transaction ends. AI breaks that pattern. Requests are distributed, bursty and concurrent. They are generated at machine speed rather than human speed. An AI agent may call a model, retrieve enterprise data, invoke an API to another system and trigger further model calls in turn. Each step adds distance, delay and dependency.

That is why a successful pilot proves only that the model works in a controlled setting. It does not prove that the enterprise can run the model across its real geography, security architecture and production traffic patterns.

The data movement problem in numbers

The scale of the shift is easy to underestimate. One statistic remains striking: a typical AI agent can generate up to 450% more total network traffic per task than a human performing the equivalent work. This is not simply more traffic of the same kind. It is a different traffic profile: distributed, bursty, concurrent and generated at machine speed. Networks designed for human browsing patterns can struggle under that load even when raw bandwidth appears sufficient on paper.

The research on bandwidth and latency reinforces the point. Bandwidth issues are now reported by 59% of organizations, up from 43% a year earlier. Latency challenges have climbed from 32% to 53%. These are not marginal changes. They suggest that many AI rollouts are colliding with infrastructure limits that were never designed for agentic workloads, rich data types and global user bases.

The pilot trap: local success, global slowdown

A familiar pattern plays out across large enterprises. In the initial environment, the AI deployment performs exactly as expected. Users are close to the cloud region. Bandwidth is ample. Usage is still modest because the organization has not yet rolled the tool out broadly. Everyone declares success.

Then the deployment goes global. Employees in Latin America, parts of Asia or Africa see dramatically slower responses. The instinctive conclusion is that the AI platform is underperforming. The logical response seems to be more spending: more compute, more licenses, more capital expenditure aimed at expanding the platform. But that can be the wrong lever. When the full end-to-end transaction is examined, the model usually is not the problem.

Consider the path of a single AI request. It may move from a branch office to a centralized security stack, then to a cloud region, then to a data-hosted backend somewhere else. Add inconsistent local internet performance, packet loss and congestion, and the AI experience deteriorates. The model may still be functioning as designed, but the data movement around it is failing the user.

The challenge also varies widely by geography. In some regions, the central question is not which model to use but how to deploy it at all. In others, the dominant concern is latency, especially when local infrastructure routes traffic back to major cloud regions. Silicon Valley and São Paulo do not present the same infrastructure conversation. Treating them as though they do is exactly how well-intentioned rollouts stall.

Bandwidth is volume; latency is time

Bandwidth and latency are often treated as the same problem, but they are not. Bandwidth is volume. AI increases it because organizations are moving richer data: documents, images, video and telemetry. Latency is time, and it is where the harder problem lives.

AI workflows are chained. An agent calls a model, retrieves enterprise data, invokes an API to another system and that system may call another model in turn. The cumulative delay across that chain is not predictable, because no one knows in advance what will be asked or how far the request must travel. A bigger circuit does not solve that problem by itself. What matters is congestion routing, packet loss, jitter and where security inspection happens along the path.

Even hyperscale infrastructure does not eliminate the issue. The system’s speed depends on whether a request can traverse the full chain, from user to network to security to execution and back, without delay compounding at every hop. When teams are certain that a cloud provider must be the bottleneck, the infrastructure on their own end is rarely the issue. More often, that infrastructure is excellent because it is built to deliver exactly what it promises. The real question is whether the request can survive the full round trip without losing time at every handoff.

Why compute and GPUs are rarely the bottleneck

Organizations have become accustomed to solving AI performance problems by adding compute. If a model is slow, buy more processing power. If usage is growing, buy more licenses. If the pilot succeeded but production is uneven, expand the platform. That reflex is understandable, but it can direct money toward the wrong layer of the stack.

In many failing AI experiences, compute and GPU availability are not the primary constraint. The model may be healthy. The application may be well designed. The problem may be congestion, routing, packet loss, jitter or security inspection along the path. Leaders who rely on average network metrics can miss these problems because averages conceal trouble along specific routes or in particular locations. A network can look healthy overall while a critical region, office or user group experiences degraded performance.

Why geography changes the AI performance equation

AI performance is not uniform across a global enterprise. It depends on where users, models and data sit, and on the legal and technical ability to move traffic across regions. Data sovereignty rules can limit where information may travel. Local internet performance can vary sharply. Routing decisions can send traffic thousands of miles away from the nearest cloud region. Each of these factors changes the experience of an AI workflow.

That is why leaders should evaluate performance in their hardest geography, not their best one. A rollout that works near a major cloud hub tells little about how it will perform in a remote office, an emerging market or a region with constrained connectivity. The organizations that plan for the hardest cases will be better prepared to scale; those that extrapolate from the easiest cases will encounter surprises after the budget has already been committed.

Questions leaders should ask before spending more

There is no one-size-fits-all fix, but there are questions every leader should be able to answer before scaling an AI deployment:

• Where do AI users, models and data sit: locally, regionally or centrally?
• What is the data sovereignty picture, and can traffic legally and technically move across the regions the rollout depends on?
• Are teams assessing the actual end-to-end network conditions experienced by AI users, including latency, jitter, packet loss and routing, or relying on average metrics that conceal problems?
• Do the team running the AI strategy and the team running the network infrastructure communicate, or do they operate in separate silos?
• What does performance look like in the hardest geography, not the best one?

Until those questions can be answered, scaling will remain difficult regardless of how well the pilot performed. Enterprises risk spending billions on an elite racing engine and then trying to race it on a road never designed for that speed. They will not get the performance they paid for, not because the engine is wrong, but because the infrastructure underneath it cannot support what they are asking it to do.

AI strategy and network strategy can no longer be separate conversations. They must be part of the same conversation from the very beginning. The organizations that scale AI successfully will be those that treat the network as part of the product, not as an afterthought.

Source: Network World News

A generic top-level domain is the portion of an internet address that appears to the right of the final dot. It is the broad category under which second-level domains are registered. When the domain system began in 1985, there were only a few primary TLDs: .com, .net, .org, .edu, .gov and .mil. The number of TLDs grew in small fits and spurts until 2012, when 1,241 new gTLDs were added. That round transformed the namespace by introducing hundreds of new alternatives to .com and other legacy TLDs. The current application set could bring roughly 1,600 new names into the process, although not every application will survive evaluation, objection or delegation.

Key facts at a glance

• ICANN received 1,615 applications for the new gTLD round, the first since 2012.
• North America accounts for 864 applications, Europe for 506, Asia/Australia/Pacific for 218, Latin America and the Caribbean for 11, and Africa for 1.
• There are 333 brand applications, at least 29 artificial intelligence strings, at least 31 cryptocurrency and blockchain strings, at least 16 network and developer strings, and 15 self-identified geographic applications.
• Some 263 names have more than one applicant. .agent is the most contested string with 13 applicants, followed by .bit with 11, and .api, .brand and .hub with 10 each.
• String Confirmation Day is set for November 17, 2026. The Community Input and Objections Period runs to March 16, 2027, and a Prioritization Draw is scheduled for the first half of 2027.

What is different from the 2012 round

ICANN has changed several parts of the application process after the 2012 round. One major change is a standalone Registry Service Provider Evaluation Program. Registry service providers run the back-end operations for a registry operator. Some operators act as their own provider, while others outsource the work. Providers can now be pre-evaluated, so the technical review does not have to happen at the same time as the application evaluation. That change is intended to separate technical readiness from the broader application review and to give applicants more flexibility in how they structure their registry operations.

The Applicant Support Program offers financial support to applicants who need it. The program received 76 applications this round, and 56 qualified. Of those, 51 submitted a gTLD application. In 2012, the program received three applications and one applicant progressed. The difference shows a broader effort to widen participation beyond large companies and well-funded registry operators. Applicants can now also apply for variant gTLDs, in part because the ICANN community has developed mechanisms to manage these TLDs. A variant TLD is an Internationalized Domain Name that can be registered in different ways because of variations in the spelling of words in a given language.

There is also an evaluation of and commitment to trademark rights protections. The ICANN community engaged in an extensive policy review of the rules and protections from the 2012 round through the Review of Rights Protection Mechanisms Policy Development Process. That work matters because a larger namespace creates more opportunities for brand owners, but it also creates more room for confusion, cybersquatting and disputes over names that resemble established trademarks. The application process now includes checks and commitments intended to address those risks before new strings move toward delegation.

Applications by region and category

The regional distribution of applications is heavily weighted toward North America and Europe. North America accounts for 864 applications, and Europe accounts for 506. Asia, Australia and the Pacific account for 218. Africa accounts for 1, while Latin America and the Caribbean account for 11. Those numbers suggest that the next gTLD round remains concentrated among applicants in wealthier and more established internet markets, even as ICANN has tried to lower barriers through applicant support and process changes.

Corporate brands make up a large share of the filing set. The 333 brand applications include technology and cloud names such as .adobe, .cloudflare, .datadog, .salesforce, .shopify, .slack, .snowflake and .stripe. Consumer platforms include .facebook, .instagram, .whatsapp, .tiktok, .zoom, .airbnb and .uber. Other brand names include .adidas, .newbalance, .bankofamerica and .goldmansachs. For large companies, a branded gTLD can serve as a controlled namespace for products, campaigns, customer services and internal systems. It can also function as a defensive asset, giving the brand owner direct control over a string that might otherwise be claimed by another party.

Artificial intelligence strings draw attention

Artificial intelligence is one of the most visible themes in the application set. The file includes at least 29 strings tied to AI terms or AI company and product names. Generic terms include .agent, which drew 13 applicants, .agentic with two, and .agi, short for artificial general intelligence, with seven. Others include .llm, short for large language model, with two applicants, .mcp, short for Model Context Protocol, with five, plus .prompt and .superintelligence. Company and product names include .openai, .chatgpt, .codex, .anthropic, .claude, .copilot, .gemini, .llama and .qwen. Community-based applications include one each for .agent and .mcp. One of the 21 Internationalized Domain Name applications is for the Chinese term for artificial intelligence.

The contest over .agent is notable because it is the single most contested string in the entire round. That level of interest reflects both the broad appeal of agentic software and the difficulty of agreeing on a single owner for a generic term. A gTLD such as .agent could be used by AI developers, automation platforms, customer service tools or other software businesses. When a term is valuable across many industries, multiple applicants may believe they have a legitimate claim to operate it. ICANN's process must then sort through the applications, objections and community input before deciding which, if any, should proceed.

Cryptocurrency and blockchain names return

Cryptocurrency and blockchain are also well represented. The file includes at least 31 strings tied to cryptocurrency and blockchain. Names include .bitcoin with two applicants, .crypto with five, .ethereum, .ether with two, .solana, .tether, .ripple, .ledger, .wallet with four, and .token with three. Four community-based applications are for crypto strings: .bitcoin, .sol, .sui and .avax. The presence of these applications shows that crypto projects continue to see domain names as part of their identity, governance and user-facing infrastructure. A dedicated gTLD can provide a recognizable namespace for wallets, exchanges, decentralized applications and community resources.

At the same time, crypto-related strings raise questions about ownership, trust and consumer protection. A name such as .wallet or .crypto could be used legitimately, but it could also be used to confuse users if it is not managed carefully. ICANN's evaluation and objection process is designed to examine whether applicants have the operational and financial capacity to run a registry, as well as whether a proposed string conflicts with existing rights or creates a risk of confusion. Community objections can also play a role when a string is tied to a specific community or public interest.

Network, developer and geographic strings

At least 16 strings come from the network and developer stack. They include .lan for local area network, .vpn for virtual private network with two applicants, .dns for Domain Name System with two, .cdn for content delivery network, .proxy, .auth and .login. The string .api, for application programming interface, drew 10 applicants. These terms are common in technical environments, and their presence in the gTLD application set shows how internet infrastructure language has become part of the broader naming economy. They also create special challenges because some of these terms are already used inside private networks or by software systems that expect them to remain local.

The 15 self-identified geographic applications include .bali, .beijing, .bremen, .budapest, .salzburg and .vienna. Geographic gTLDs can be used by tourism boards, local governments, business associations or community groups to promote a place and organize local digital services. They can also be contentious when multiple parties claim to represent the same location or when the proposed string overlaps with a cultural or commercial identity. Internationalized Domain Names add another layer of complexity. Applications can be submitted in 26 scripts representing more than 380 languages, and applicants can seek variant gTLDs that account for different spelling conventions within a language.

Name collision and universal acceptance

Adding new gTLDs to the internet is not only about accepting proposals. It is also about validating that each new name will work without disrupting existing services. Two software readiness issues are especially important: name collision and universal acceptance. Name collision is the problem of a new public TLD matching a name that a network already uses privately. Many organizations use names inside their networks that do not exist in the public DNS, such as an internal domain ending in something like .corp or .lan. If ICANN delegates a new gTLD with the same string, lookups for those internal names can start resolving through the public DNS, or leaking out of the network, instead of staying internal. That can break applications or expose traffic.

ICANN's name collision framework subjects every applied-for string to a name collision risk assessment. What it requires of the applicant depends on the risk identified. A low-risk string may need limited measures, while a higher-risk string may require more detailed analysis, monitoring or mitigation. The goal is to prevent a new public namespace from interfering with private naming systems that were never designed to interact with the global DNS. Because many enterprises rely on internal domains for authentication, routing and application access, name collision review is a critical part of gTLD readiness.

Universal Acceptance is about whether software actually works with every valid domain name and email address, including new ones. The broader use of Internationalized Domain Names and longer TLDs depends on progress in Universal Acceptance, which ensures that all valid domain names and email addresses can be correctly accepted, validated, processed, stored and displayed across internet-enabled applications, devices and systems. If a new gTLD is delegated but software rejects it, the expansion creates little practical value. That is why Universal Acceptance has become a long-term concern for registries, software vendors, email providers and governments.

What happens next

So far, ICANN has only revealed which applications have been made. The next milestone is String Confirmation Day on November 17, 2026, when ICANN will publish the final list of strings that will proceed through the program. That date begins the Community Input and Objections Period, which runs to March 16, 2027. During that window, parties can raise concerns about specific applications, including issues related to trademarks, community impact, name collision or applicant qualifications. ICANN has also set a Prioritization Draw for the first half of 2027 to determine the order in which applications will be processed. The draw matters because a large application set cannot be evaluated all at once, and the order of review can affect timelines for applicants and the communities that may be affected by new strings.

Source: Network World News

That finding matters because observability has become a core operational capability rather than a niche monitoring function. Network, cloud, application, security, and service management teams all rely on telemetry to understand whether digital services are healthy. When those teams use separate toolchains, each group may see a narrow slice of reality. The result can be a frustrating paradox: every dashboard reports that its domain is healthy, while users still experience delays, outages, or degraded performance. Closing that gap requires more than adding another tool or layering AI on top of the existing stack.

Key Facts From the Research

• The survey included 356 enterprise IT professionals.
• Most organizations use six to eight observability tools.
• 51 percent had tool unification efforts under way, 32 percent were planning or evaluating unification, and 17 percent had completed it.
• 45 percent cited skills and staffing burdens from operating multiple tools.
• 44 percent said integration and API complexity consume engineering time.
• 41 percent said context switching among tools slows investigations and response.
• 40 percent reported increased manual effort.
• 34 percent said alert noise creates toil and cognitive overload.
• 32 percent use AI extensively across observability tools, 51 percent use it in selected areas, and 13 percent are still piloting it.
• 24 percent reported fully centralized observability tool governance, while 48 percent said ownership was centralized for the most part with some domain-specific exceptions.

Tool Sprawl Strains IT Operations

Tool sprawl is more than a procurement problem. Every additional platform introduces its own data model, interface, alerting logic, access controls, and upgrade cycle. Engineers must learn how each tool behaves, how to query it, and how to interpret its output. Integration work expands as teams try to connect dashboards, ticketing systems, automation workflows, and security controls. Even when individual tools are effective within their domains, the seams between them can slow incident response and obscure the end-to-end customer experience.

The survey captured several of these operational costs directly. Forty-five percent of respondents cited skills and staffing burdens from operating multiple tools. Forty-four percent said integration and API complexity consume engineering time. Forty-one percent pointed to context switching among tools as a drag on investigations and response. Forty percent reported increased manual effort, and 34 percent said alert noise creates toil and cognitive overload. These are not abstract concerns. They translate into longer mean time to detect and resolve incidents, more repetitive work, and less time for engineering teams to improve reliability.

Sprawl also affects cost and data quality. More tools mean more licenses, more training, more integration maintenance, and more overlapping capabilities. Data may be duplicated, delayed, or defined differently across platforms. Coverage gaps can emerge when one team assumes another tool is monitoring a dependency. End-to-end visibility suffers when no single group can correlate signals across network, cloud, application, and user experience domains. In one common scenario, network operations can show that connectivity is healthy, cloud teams can show that infrastructure is operating normally, and application teams can show that services are available, yet users still report a problem. The tools may all say everything is fine, but the organization lacks the means to connect the dots.

Why the Single Pane of Glass Remains Elusive

The idea of a single pane of glass has long appealed to IT leaders. In practice, it has been difficult to achieve because observability is not one market or one workflow. Network performance monitoring, application performance monitoring, cloud infrastructure monitoring, log analysis, security analytics, digital experience monitoring, and service management each have specialized requirements. Large organizations often inherit tools through mergers, regional operations, or domain-specific purchasing decisions. Central IT may standardize on one platform while application teams adopt another, and security teams may insist on separate controls for compliance reasons.

The survey reflects that reality. Half of respondents had unification efforts under way, but only 17 percent had completed unification. Nearly one-third were still planning or evaluating their approach. Larger organizations reported using more observability tools across their environments, which is consistent with the scale and diversity of their technology estates. Unification, therefore, is less about finding one product that does everything and more about creating a governed architecture in which data, workflows, and ownership are connected across tools. The goal is a coherent operating model, not necessarily a single platform.

AI Takes On More of the Workload

Enterprises are already applying AI to operational data. Thirty-two percent of respondents said they use AI extensively across observability tools, 51 percent use it in selected areas, and 13 percent are still piloting it. Common use cases include incident summaries and explanations, alert correlation and noise reduction, capacity forecasting, anomaly detection, predictive analytics, and root-cause analytics. These capabilities can reduce manual effort and help engineers make sense of large volumes of telemetry. AI can also help correlate information across tools when data is fragmented, surfacing patterns that would be difficult to spot manually.

But AI is not a substitute for fixing bad data or broken processes. If the underlying telemetry is incomplete, inconsistent, or poorly governed, AI models may produce confident but inaccurate conclusions. Organizations that reported extensive AI use also said they had greater confidence in their observability unification strategies. That correlation makes sense: teams with cleaner data, stronger integration, and clearer ownership are better positioned to train, operate, and trust AI systems. At the same time, respondents cited concerns about inaccurate outputs and security and compliance risks. Those concerns grow as AI moves from summarizing alerts to recommending actions and eventually taking action on infrastructure.

Governance Before Simplification

Tool governance has emerged as a prerequisite for successful unification. About one-quarter of organizations reported fully centralized observability tool governance, while 48 percent said ownership was centralized for the most part with some domain-specific exceptions. Organizations with centralized ownership reported greater confidence in their unification strategies, deeper integration between observability and IT service management, more effective incident-impact analysis, and more mature use of AI-powered observability.

Centralization, however, is not simply a matter of putting one group in charge. The research found that understaffed tool governance and engineering teams had less success with unification strategies. A centralized group needs authority, funding, engineering capacity, and executive support to define standards, manage integrations, and resolve conflicts between domain-specific needs. Without those resources, governance becomes a bottleneck rather than an enabler. The practical lesson is that organizations should establish clear ownership and then equip that owner to deliver. Simplification efforts are more likely to succeed when governance precedes consolidation rather than following it.

When AI Starts Acting on Infrastructure

AI governance becomes even more important when AI systems begin taking actions across network infrastructure. As enterprises give AI more autonomy, they need controls that match the probabilistic nature of the technology. Agentic AI systems can generate responses and take actions that are not always predictable, while many traditional cybersecurity controls are designed to operate deterministically by following predefined rules. That mismatch creates risk.

Consider a firewall rule that appears unnecessary to an AI system. The rule might allow a particular IP address to access a database process for one day each quarter in support of a legitimate business process. An AI system scanning for unused or redundant rules might identify it as a mistake and remove it without understanding the quarterly workflow it supports. What looks like cleanup could become an outage. This example illustrates why network teams must determine where human review remains necessary as AI takes on more operational functions.

AI governance requires human oversight for unusual exceptions, business-critical rules, and potentially harmful or inaccurate outputs. Controls should be tailored to an organization’s risk profile, with clear boundaries around what AI systems can and cannot do. For network operations teams, the increasing use of AI and automation requires integration of observability data and workflows, stronger tool governance, and controls over autonomous actions. AI is becoming another tool in the operations tool chest, and it will continue to evolve. The organizations that benefit most will be those that treat it as part of a governed operating model rather than as a shortcut around fragmentation. The survey’s central lesson is that AI can make complex environments more manageable, but it cannot make poor data, weak processes, or unclear ownership disappear.

Source: Network World News

Upscale introduced Token Fabric with general availability planned for early 2027. Early-access and joint-validation programs are already running. The company said the architecture is intended to give AI operators one control and observability plane across both networking tiers, rather than forcing them to stitch together separate fabrics as accelerator clusters grow from a few racks to hundreds or thousands.

Key facts at a glance

• Upscale has launched Token Fabric, an integrated scale-up and scale-out networking architecture for AI clusters.
• The design combines SkyFabriX scale-up switch silicon, scale-out systems built on Nvidia Spectrum-X Ethernet and a common software stack led by SkyOS and SkyCMD.
• General availability is planned for early 2027; early-access and joint-validation programs are running now.
• SkyFabriX supports ESUN and standard Ethernet/IP, is designed for UALink, offers 115.2 Tbps switch capacity and has a roadmap to multiple petabits per second.
• Upscale emerged from stealth in September 2025 with a $100 million seed round, then raised $190 million in June; Nvidia joined that round.
• The company has raised about $500 million, has more than 300 employees and is based in Santa Clara, Calif.
• Upscale is positioning tokens, not just packets, as a key metric for AI network performance.

Why AI networking needs a different approach

AI infrastructure has traditionally been described in terms of compute: GPUs, accelerators, memory bandwidth and power. But as clusters scale, networking has become a first-order constraint. Training and inference jobs are distributed across many accelerators, and their performance depends on how quickly data and synchronization messages move between those devices. Scale-up fabrics handle the high-bandwidth, low-latency communication inside a rack. Scale-out fabrics handle communication among racks. When those two fabrics are managed separately, operators can face inconsistent policies, fragmented telemetry and slower troubleshooting.

Barun Kar, CEO of Upscale, framed the problem as a widening mismatch between compute and networking. Compute is ramping up on a rapid scale, he said, while networking has fallen way behind. Token Fabric is Upscale's answer: a vertically integrated portfolio that spans silicon, systems and software rather than a single switch or a single operating system.

Inside Token Fabric: silicon, systems and software

Token Fabric is built from several components. At the hardware level, SkyFabriX is Upscale's scale-up switch silicon. It supports ESUN, or Ethernet for Scale-Up Networking, as well as standard Ethernet/IP. It is designed for evolving protocols such as UALink, the Ultra Accelerator Link. Upscale said SkyFabriX offers 115.2 Tbps of switch capacity, with a roadmap to multiple petabits per second.

The company also offers scale-up switch trays in custom and standard rack form factors. Those trays pool GPUs and other accelerators into large scale-up domains. For scale-out, Upscale builds systems around Nvidia Spectrum-X silicon at speeds ranging from 400G and 800G up to 1.6T. With SkyOS, the company said, those systems integrate into Nvidia and other accelerator clusters.

SkyOS is the network operating system for both fabrics. It is based on SONiC, the Software for Open Networking in the Cloud, and is optimized for AI protocols. SkyCMD is the orchestration and observability layer. It provides a single orchestration plane across scale-up and scale-out, which is central to Upscale's pitch. Instead of requiring network teams to operate one stack for in-rack accelerator communication and another for data-center-wide communication, Token Fabric aims to present one operational model.

Upscale is offering several consumption models. Customers can adopt silicon, systems, software or the full stack. The full-stack option targets neoclouds and enterprises that do not have hyperscaler-sized network engineering teams. An integrated-systems option targets XPU vendors, meaning makers of non-GPU accelerators that want validated hardware without building a complete network stack. A third option provides silicon or custom systems with a software development kit for hyperscalers that run their own network operating system. All three options draw on the same portfolio of silicon, systems and software.

Everything has to be integrated, Kar said, adding that this is where Upscale comes in and that the company is vertically stacked. The strategy reflects a broader reality in AI infrastructure: many customers want validated building blocks, but they do not all want the same level of integration. Hyperscalers may insist on controlling their own software stack, while neoclouds and enterprises may prefer a more complete system. XPU vendors need networking that works with their accelerators but may not want to become networking companies themselves.

Protocols and open-source contributions

Token Fabric is built from the protocol layer up. Upscale is extending existing protocols rather than replacing them. Scale-up traffic uses ESUN and standard Ethernet/IP. Scale-out traffic uses standard Ethernet with RoCE, which carries remote direct memory access over Ethernet. That choice matters because Ethernet is already the dominant foundation in many data centers, and RoCE has become a common way to deliver low-latency, high-throughput communication for AI and storage workloads.

Upscale is also contributing to open projects underneath the stack. Those include SONiC, extensions of the Switch Abstraction Interface, or SAI, for ESUN, the Ultra Ethernet Consortium, or UEC, and UALink. The involvement in standards and open-source communities is significant because AI networking is still evolving. Customers do not want to be locked into a proprietary protocol path if the industry converges on different approaches for scale-up or scale-out. By supporting ESUN, Ethernet/IP, RoCE and UALink while contributing to SONiC and SAI, Upscale is trying to position Token Fabric as compatible with both current and emerging designs.

Software ties the two fabrics together. SkyOS abstracts the underlying hardware and provides a control plane across the cluster. SkyCMD exposes that control plane through one layer. Multiple network elements sit behind it, and any compute platform in a heterogeneous cluster can use it to control the network. Kar described the operating system as needing to be lean and mean to be fast, flexible and secure, with an orchestration layer on top to enable capabilities such as heterogeneous compute. In practice, that means a cluster mixing different accelerator types or vendor platforms could use a common network control model rather than separate tools for each silo.

Tokens as the new networking metric

Network teams have traditionally tracked packets, throughput, loss and jitter. Those metrics remain important, but AI operators increasingly care about tokens. Tokens are the units that AI models generate and process, and they are also the basis for many commercial AI services. Kar argued that tokens are the right unit of optimization. Users pay based on tokens today, he noted, and the relevant measures include time to first token, tokens per second, tokens per dollar and tokens per watt.

Optimizing those measures requires networking, especially in clusters that scale to hundreds of thousands of accelerators. A delay in moving data between accelerators can increase time to first token or reduce tokens per second. Inefficient communication can raise cost per token or power per token. The challenge is that many networks today were not designed with token-level operations as the primary target. Kar identified latency, bandwidth and scale-up features in silicon as major components of the gap. On the software side, he pointed to predictive analytics and telemetry as tools that keep machines running.

The other side of the challenge is optimizing networking with agentic AI. Upscale has built agent-based operations directly into the platform. Kar said an agent is always looking at the network, trying to figure out whether cables will fail, optics will fail or congestion will appear, and the agents are looking for ways around those problems. That vision moves network operations from static dashboards and manual runbooks toward continuous, automated monitoring and remediation. For AI clusters, where downtime and congestion can directly affect token throughput and cost, that kind of closed-loop operations capability could become as important as raw switch capacity.

Token Fabric's early-access and joint-validation programs are intended to test those claims with customers before general availability in early 2027. The architecture also illustrates a wider shift in AI infrastructure: networking is no longer a supporting utility but a core part of how compute is pooled, scheduled and monetized. As accelerators become more powerful and clusters grow larger, the boundary between scale-up and scale-out will matter less to the application and more to the operator. Upscale's bet is that customers will want one fabric, one software layer and one operational view, even if the underlying silicon and protocols remain diverse.

Source: Network World News

Key Facts at a Glance

• Arista introduced Ethernet reference architectures and integration partnerships with AMD, Arm, Broadcom, Meta, Microsoft, Qualcomm, and others.
• The designs are based on standard Ethernet, with scale-up built on ESUN Ethernet and scale-out built on Ethernet with UEC features.
• Arista is contrasting its approach with Nvidia’s proprietary NVLink and InfiniBand stacks.
• The company will offer 7060EX7 Series switches along with fiber patch panels, integrated liquid-cooling manifolds, drip trays with active leak detection, rack management controllers, and power shelves with backup battery units.
• Arista’s Etherlink SU-144 scale-up architecture includes orthogonal chassis, cabled backplane, and cross-rack designs.
• The cross-rack design supports up to 1,024 accelerators while maintaining high-bandwidth, low-latency connectivity.
• Arista is not selling integrated racks; it is selling switches and software for joint development designs delivered through partners and system integrators.
• NetDI, Arista’s Network Diagnostics Infrastructure, provides hardware-level validation, telemetry, and failure isolation across EOS, SONiC, FBOSS, and other network operating systems.

Arista’s Ethernet Push Targets a Changing AI Networking Market

Arista’s latest move reflects a broader shift in AI infrastructure. For years, Nvidia’s proprietary scale-up networking technologies were the primary option for connecting accelerators inside high-performance AI systems. That limited the choices available to customers and gave Nvidia significant control over the networking layer. As the scale-up market expands and more non-Nvidia accelerators reach production, Arista sees a new opening for Ethernet-based designs that can support multiple accelerator families and give customers more flexibility.

Hardev Singh, vice president and general manager of Arista’s cloud and AI group, said the company’s strategy is built on the idea that customers want an open, standards-based alternative. He noted that Arista’s positioning contrasts with Nvidia’s proprietary NVLink and InfiniBand stacks. According to Singh, the growth of the scale-up total addressable market and the rise of the non-Nvidia accelerator ecosystem create a brand-new opportunity for Arista. The key differentiator, he said, is that the company’s approach is all Ethernet—strictly Ethernet.

Singh pointed to a range of accelerator programs as evidence that the market is broadening. He cited Microsoft Maia, Meta’s MTIA training and inference accelerator, AMD’s Helios, and Google’s Tensor Processing Units. Those technologies, he said, represent opportunities for Ethernet players such as Arista. The implication is that AI networking will not be dominated by a single proprietary fabric, but will instead support a mix of accelerators, switch silicon, optics, and software from multiple suppliers.

ESUN and UEC Provide the Standards Foundation

Arista’s scale-up designs are based on the Ethernet for Scale-Up Networking initiative, known as ESUN. The initiative was formed by the nonprofit Open Compute Project and is intended to advance Ethernet networking technology so it can handle scale-up connectivity across accelerated AI infrastructure. ESUN backers include AMD, Arista, Arm, Broadcom, Cisco, HPE Networking, Marvell, Meta, Microsoft, Nvidia, OpenAI, and Oracle.

According to Arista, ESUN gives network operators flexibility across accelerators, switching silicon, optics, physical rack designs, and network software while retaining the high bandwidth, low latency, and reliable in-order delivery required for collective communication. Those characteristics are essential for AI training and inference workloads, where accelerators must exchange data quickly and consistently across large clusters.

For scale-out networking, Arista is leaning on the Ultra Ethernet Consortium, or UEC. The UEC includes AMD, Arista, Broadcom, Cisco, Eviden, HPE, Intel, Meta, and Microsoft. Its charter is to bring together industry leaders to build a complete Ethernet-based communication stack architecture for high-performance networking. Arista’s scale-out designs incorporate UEC features and are intended to handle the demanding traffic patterns of AI token generation and other accelerated workloads.

Three Scale-Up Designs for High-Density AI Racks

Arista and its partners plan to offer designs that give customers a path to integrate compute, networking, power, and liquid cooling across evolving high-density AI deployments. The three primary designs are part of what Arista calls Etherlink SU-144, an Ethernet ESUN-based scale-up architecture.

Orthogonal Chassis

The orthogonal chassis design uses direct orthogonal connectivity between accelerator and switch blades. It supports accelerator density of 144 XPUs in 100 kW to 400 kW direct liquid cooling envelopes. That design is intended for customers that need very high density within a single chassis and are prepared to use direct liquid cooling to manage the thermal load.

Cabled Backplane

The cabled backplane design combines a high-density cabled backplane with a modular rack structure to enable a highly serviceable design. Serviceability matters in large AI deployments, where components must be replaced or upgraded without disrupting the entire cluster. A modular approach can help operators maintain availability while evolving the system over time.

Cross-Rack Architecture

The cross-rack design is a modular architecture that provides maximum flexibility to scale compute and networking across multi-rack deployments. It enables larger scaling domains across scale-up and scale-out. Bob Laliberte, principal analyst and founder at Liberte Research Group, noted that the cross-rack architecture is particularly noteworthy because it extends the scale-up domain beyond the physical limits of a single rack. It supports up to 1,024 accelerators while maintaining high-bandwidth, low-latency connectivity.

Laliberte also observed that Arista is addressing a broader infrastructure shift. As AI clusters become denser and move toward liquid cooling, the rack is replacing the individual switch as the practical unit of deployment. Power distribution, thermal management, fiber density, leak detection, and system validation must now be engineered as an integrated solution rather than as separate components.

Scale-Out Designs Focus on Resiliency and Load Balancing

On the scale-out side, Arista is defining UEC-based designs that include Multipath Reliable Connection fabric resiliency, multi-plane routing, intelligent dynamic load balancing, and cluster load balancing. These capabilities are intended to handle AI token generation efficiency and the variable workloads that characterize modern AI applications. As AI clusters grow, the network must be able to route traffic efficiently, recover from failures, and balance load across many paths without introducing bottlenecks.

Singh said customer designs vary by accelerator power, GPU count, bandwidth, thermal envelope, and rack configuration. He positioned Arista’s capabilities in chassis engineering, thermal design, and signal integrity as differentiators. Those engineering disciplines become more important as racks become denser and as liquid cooling moves from an exception to a standard requirement for high-performance AI infrastructure.

Bandwidth Density and Data Center Footprint

Arista said one rack configuration features bandwidth that scales with next-generation interconnects. A 16-switch configuration approaches 1.6 petabytes per second per rack. Other designs using XPO, CPO, and NPO optical technologies could quadruple bandwidth, according to Singh. A Tomahawk 7, 200G-based generation could support 6.5 petabytes per second in one rack while exceeding 100 kilowatts.

Compared with today’s OSFP-based switches, which deliver 1.6 petabytes per second per rack, next-generation optics increase that to 6.5 petabytes per second per rack, according to Singh. That massive densification translates to nearly a 50% reduction in the overall data center footprint, resulting in significant time and financial savings. The implication is that AI operators can pack more networking capacity into fewer racks, which matters as power, cooling, and physical space become constrained.

A Partner-Driven Go-to-Market Model

Arista is taking a partner-driven approach to integrated racks. Singh made clear that customers will not find a rack SKU on Arista’s price list. The company is not selling integrated racks. Instead, it is selling the switches that go into liquid-cooled racks, along with software and technology for integration with custom joint development designs.

Arista is partnering with value-added resellers and system integrators that customers can work with to deliver the fully integrated rack. The company is also working with different optics vendors as well as partners such as Foxconn, Quanta, and Hive. That model allows Arista to focus on its core networking and software strengths while relying on partners to handle rack-level integration, manufacturing, and deployment.

The approach may appeal to hyperscalers and large enterprises that want to customize their AI infrastructure rather than buy a fixed, proprietary system. By offering reference architectures and joint development options, Arista can support customers with different accelerator choices, power envelopes, cooling strategies, and rack designs. At the same time, the partner model lets Arista avoid the complexity of selling and supporting complete racks as a single SKU.

NetDI Provides the Diagnostics and Validation Layer

The foundation of all these designs is Arista’s Network Diagnostics Infrastructure, known as NetDI. It is a diagnostics and validation layer that runs between Arista’s EOS operating system or SONiC, FBOSS, or another network operating system. Arista’s newer 1.6T AI switches integrate NetDI specifically to provide consistent diagnostics across different operating environments.

NetDI can track several switch diagnostics functions, including cable and optic health, device telemetry, troubleshooting, and failure isolation. Those capabilities help cloud companies manage massive scale networks where a single faulty link or optic can affect a large AI workload. Arista said NetDI offers deep hardware-level validation, signal integrity analysis, secure boot attestation, and Single Event Upset resiliency. It also supports rich telemetry for the switch, physical optics, power shelves, and liquid cooling infrastructure.

As AI clusters grow denser and more complex, diagnostics and validation become more important. Operators need visibility into the health of cables, optics, power components, and cooling systems, not just the switch itself. Arista’s message is that Ethernet-based scale-up and scale-out designs can deliver the performance, flexibility, and operational visibility that AI infrastructure requires, while giving customers an alternative to proprietary networking fabrics.

Source: Network World News

The timing is deliberate. Apple unveiled the first iPod on October 23, 2001, at a moment when portable digital music was still fragmented and often awkward. The device did not invent the MP3 player, but it packaged the idea in a way that made digital music feel simple, elegant, and personal. Twenty-five years later, the iPod brand is discontinued, yet its influence remains visible in the design of Apple's hardware, the behavior of its music software, and the nostalgia many users still attach to the scroll wheel.

Headline and key facts

• Apple has marked the iPod's 25th anniversary with a hidden interactive iPod on a Page Not Found screen.
• The first iPod was introduced on October 23, 2001.
• The easter egg appears when a visitor reaches a nonexistent path on Apple's music website.
• The on-screen iPod is fully interactive: its buttons can be clicked, and the scroll wheel can be clicked and dragged.
• The device includes pre-installed songs from artists who were popular around the time of the original iPod launch.
• Visitors can select songs and play 90-second previews directly in the web interface.
• The iPod line has long been discontinued, but its cultural legacy remains strong.

A quiet tribute to a device that changed listening

The easter egg is easy to miss. It does not appear on the main Apple homepage, and it is not announced with a giant marketing campaign. Instead, it waits behind a broken path on Apple's music website, turning a routine error page into a small piece of interactive history. That placement feels appropriate for a product that began as a premium, almost luxury item and eventually became a familiar part of everyday life. The page invites exploration rather than demanding attention.

The interactive iPod is not a video or a static image. Visitors can click its buttons, drag the scroll wheel, and move through menus in a way that mimics the original device. Pre-installed songs are available to select, and each track can play a 90-second preview inside the web interface. The experience is limited compared with a real iPod, but it captures the essential rhythm of the hardware: the wheel, the click, the menu, and the sudden focus on a single song.

The original iPod arrived at a turning point

When Apple introduced the iPod in October 2001, the music industry was in the middle of a difficult transition. CDs were still dominant, but digital files were becoming easier to find, share, and store. Early MP3 players existed, yet many were clumsy, limited by small storage, poor software, or confusing controls. Apple's pitch was not simply that the iPod held music. The pitch was that it held a lot of music and made it easy to find.

The first iPod offered 5GB of storage, enough for roughly 1,000 songs, and it used a small hard drive rather than flash memory. It sold for $399, a price that made it a premium product rather than an impulse purchase. It worked with a Mac and relied on iTunes for syncing. The combination of hardware, software, and a simple scroll wheel gave the device a clear identity. Users could carry a large library in a pocket and navigate it with one hand.

The launch also reflected Apple's growing interest in consumer electronics. The company was known for Mac computers, but the iPod pushed it into a new category and eventually helped reshape its business. The device arrived before the iTunes Store, which launched in 2003. In the early years, iPod owners ripped CDs or found digital music elsewhere and then synced their libraries. The later arrival of legal downloads made the ecosystem more complete, but the hardware had already established the basic ritual: build a library, connect the device, and take it everywhere.

From scroll wheel to cultural symbol

The iPod line evolved quickly. Apple introduced smaller models, flash-memory players, colorful designs, and eventually the iPod touch, which brought a touchscreen and app support. The iPod mini and iPod nano appealed to users who wanted something lighter. The iPod shuffle removed the screen entirely and leaned on physical controls. The iPod classic carried the large-library idea forward for years. Each version changed the formula, but the core promise remained: a dedicated device for music that was simpler than a phone or a computer.

That promise made the iPod a cultural symbol. White earbuds became recognizable. The silhouette advertising campaigns turned dancers and musicians into moving shapes against bright backgrounds. The click wheel became a tactile interface that people could operate without looking. The device appeared in movies, television shows, and everyday conversations. For a generation of listeners, the iPod was the first place where a personal music collection felt portable, searchable, and entirely under their control.

The product also changed expectations for how digital media should work. Users expected fast syncing, clean menus, and reliable playback. They expected their devices to feel responsive. Those expectations carried into later Apple products, including the iPhone and the Apple Watch. The iPhone eventually absorbed most of the iPod's functions, and smartphones from many companies became the default music players for millions of people. The dedicated music player did not disappear because people stopped listening. It faded because the phone became good enough to do the job.

What the interactive model includes

The hidden page recreates the look of an original iPod, including its display area, controls, and scroll wheel. The wheel responds to click-and-drag input, which means visitors can simulate the circular motion of the original interface. The buttons allow navigation through the device's menus, and the screen updates as the user moves around. The experience is deliberately small in scope, but it is more than a decorative image.

The pre-installed music is drawn from artists who were popular around the time of the iPod's introduction. That choice gives the easter egg a period feel. It is not a random selection of modern tracks placed inside an old shell. The page tries to evoke the era of the original launch, when playlists were measured in burned CDs and ripping a library was a weekend project. The 90-second previews let visitors sample songs without leaving the page, which keeps the experience self-contained.

For users who remember the original hardware, the appeal is partly nostalgic. The wheel, the menu structure, and the small screen bring back the physical habits of using an iPod. For younger users, the page offers a glimpse of what portable music felt like before streaming, touchscreens, and always-connected devices became the default. The easter egg does not need to explain that history. It lets visitors feel a small part of it.

Why the iPod still matters

The iPod was discontinued years ago, and Apple stopped selling the last model in 2022. Even so, the device remains a reference point in discussions about product design, digital music, and the way technology becomes personal. The iPod was not the most powerful computer Apple ever made, and it was not the most expensive. Its importance came from focus. It did one job well, and it made that job feel enjoyable.

The new easter egg fits into a broader pattern of anniversary gestures that reward curious users. It turns a dead end into a discovery. It takes a product that exists mostly in memory and gives it a temporary digital form. The page does not pretend that the iPod is still a current product. Instead, it acknowledges the anniversary with a small, interactive tribute that can be explored in a few minutes.

Apple Music subscriptions start at $11.99 per month, and the service is also available through the Apple One bundle. The hidden iPod page is separate from those subscription offers, but it points to the same long relationship between Apple and music. The company no longer sells a dedicated iPod, yet the music business remains central to its ecosystem. The easter egg is a reminder of how that relationship began for many users: with a wheel, a screen, and a library of songs in a pocket.

Source: 9to5Mac News

After a week of use, the iPhone Duo feels less like a novelty and more like a preview of where the iPhone could go next. It is not a perfect first-generation product. It makes notable trade-offs, including Touch ID instead of Face ID and no dedicated telephoto camera. Even so, the hardware and software work together in a way that makes the foldable design feel useful rather than experimental.

Key facts at a glance

• 7.6-inch inner display with a 1.4:1 aspect ratio
• 5.4-inch outer display with the same 1.4:1 aspect ratio
• Grade 5 titanium frame with a polished finish
• Side-button Touch ID instead of Face ID
• A20 Pro chip, the same silicon used in the iPhone 18 Pro and Pro Max
• Dual-battery architecture with up to 24 hours of typical mixed use
• 48MP main and 48MP ultra wide rear cameras
• Under-display 4MP FaceTime camera on the inner screen
• IP68 dust and water resistance
• Starts at $1,999

Design and hinge

The iPhone Duo weighs 254 grams, five grams more than the iPhone 18 Pro Max. On paper, that sounds like a step in the wrong direction for comfort. In practice, weight distribution matters more than raw weight. The Duo is shorter and wider than a traditional large iPhone, so it sits differently in the hand. It is still heavy enough that extended one-handed use can tire the wrist, but it is more manageable than its weight suggests.

Thickness is the more obvious compromise. When closed, the iPhone Duo measures 11.3mm, about 29% thicker than the iPhone 18 Pro and Pro Max. When unfolded, it measures just 5.2mm, making it thinner than the iPhone Air. The folded footprint is 117.8mm tall and 84.1mm wide. That makes it 45.6mm shorter than the iPhone 18 Pro Max but only 6.1mm wider. The result is a device that carries well in a pocket despite its added bulk.

The hinge is central to the experience. It uses more than 100 individual components, a variable torque system, and an integrated magnet system that keeps the two halves closed. Opening the Duo is very much a two-handed operation. The resistance is strong enough that one-handed unfolding does not feel comfortable or secure. The hinge can also hold the device at multiple angles for seated and tent configurations, which is useful for watching video, taking FaceTime calls, or using StandBy without a separate stand.

The frame is made from Grade 5 titanium with a polished, mirror-like finish. The device is available in Night Sky and Star White. Night Sky shifts between dark blue and black depending on the light. The iPhone Duo also carries an IP68 rating for dust and water resistance. After the first day, the novelty of folding an iPhone fades, and the hinge stops feeling like something fragile. That is a compliment: the device begins to feel like a normal phone that happens to unfold.

Outer display

The outer display measures 5.4 inches diagonally and uses the same 1.4:1 aspect ratio as the inner screen. That sounds small, especially to anyone who remembers 5.4-inch mini iPhones. The comparison is misleading because the aspect ratio is different. Apple says the outer display has 90% of the screen area of the iPhone 18 Pro. The shape is wider and shorter, more like a passport than a tall smartphone.

For right-handed users, the outer screen is easy to use one-handed. All corners and edges are reachable, unlike taller iPhones that require stretching to reach the top-left corner. Typing is comfortable, though the keyboard takes up just under half of the display. Two-handed typing and one-handed swipe typing both work well. Left-handed users may find the right-aligned software controls less comfortable, particularly when reaching for Control Center in the top-right corner.

The outer display works well for quick interactions: replying to messages, checking email, reviewing a shopping list, or jotting notes. It feels like pulling a small physical notebook from a pocket. It is also usable for longer social media scrolling and reading, even if those tasks are better on the inner screen. The main weakness appears in Safari. The modern web is not optimized for a short, wide 5.4-inch display, and cookie banners or newsletter popups can dominate the view. Once a long-form article is open, reading is pleasant, but getting there can be frustrating.

Inner display and crease

The inner display is the heart of the iPhone Duo. It measures 7.6 inches and offers 50% more screen area than the iPhone 18 Pro Max and 80% more than the iPhone 18 Pro, according to Apple. It is the largest display ever used on an iPhone. Unfolding the device reveals a screen that feels like an iPad mini, and the wow factor does not wear off quickly.

The 1.4:1 aspect ratio gives the inner screen a wider, more rectangular shape than many competing foldables. That makes it well suited to video with minimal letterboxing, web browsing, reading, and social media. It also avoids the tall, square shape that can feel like two normal phones taped together. On the iPhone Duo, the foldable design feels like the main event rather than an afterthought.

Apple has worked extensively on the crease. The inner display uses more than 10 ultra-thin layers, a scratch-resistant coating, a nano-textured polymer layer, high-strength glass above and below the OLED panel, special adhesives that let layers slide, and a titanium plate for support. The nano-texture finish also diffuses reflected light, which reduces how easily the eye picks up the crease. The coating is replaceable as a repairability benefit, not because users are expected to wear it out quickly.

In real-world use, the crease is not invisible, but it is rarely distracting. It can be seen at unflattering angles or in bright sunlight when viewing dark interfaces, such as a black loading screen or video frame. During normal use, it is more something you can find when looking for it than something you notice while working. The crease can be felt while scrolling, especially in portrait mode, but it becomes easy to ignore. The nano-texture coating also reduces glare and gives the screen a paper-like feel, though fingerprints are a little harder to remove than on a glossy display.

Software and apps

The iPhone Duo runs iOS 27 rather than a hybrid operating system or iPadOS. That decision keeps the experience familiar, but iOS has been adapted for the Duo's two screens. Many navigation controls move from the bottom to the right side of the interface. In apps like Music, the navigation bar and top controls shift into a right-hand sidebar. On the Home Screen, the dock moves to the right side, with status bar icons above it in the upper-right corner. When the device is unfolded in portrait orientation, the dock and in-app navigation bars return to the bottom.

The interface adapts well after a few days. UI elements appear where expected regardless of how the device is held, and the transition between folded and unfolded states is polished. The Duo also supports Split View, a first for an iPhone. Two apps can run side by side, or two windows from the same app can be used together. A video can be pinned to the top of the screen while another app fills the space below. This makes multitasking far more capable than on any previous iPhone.

All preinstalled Apple apps have been updated for the Duo's displays, with controls that float alongside content and make better use of the wider screen. Third-party apps fall into three broad tiers. Apps built with older SDKs still run, but they may appear letterboxed or floating in the center. Apps built with the iOS 27 SDK resize to fill most of the inner display while leaving room for the status bar. Apps built with the iOS 27.1 SDK or later can be fully optimized for both displays.

The difference between optimized and non-optimized apps is significant. In a week of testing, many developers released updates for the Duo, including apps for social media, weather, podcasts, reading, and productivity. For some users, proper Duo support may become a deciding factor when choosing between similar apps. Still, the software experience is not fully there yet. Split View would benefit from iPadOS-style drag and drop, and Slide Over feels like an obvious missing feature. Safari also does not always receive desktop-class web layouts on the inner display, even though the screen is close in size to an iPad mini.

Touch ID and cameras

The biggest hardware trade-off is Touch ID instead of Face ID. The sensor is integrated into the side button, and it supports Rest to Open, so a finger can unlock the device without pressing the button. Touch ID works quickly and reliably, and it is convenient when the phone is lying flat on a table. Apple Watch unlock also improves the experience significantly. Even so, reaching for a sensor feels like a step backward after years of glance-based Face ID unlocking. The Duo also lacks attention-aware features, such as expanding notifications when the user looks at the screen.

The camera system is good but not the best Apple offers. The rear includes a 48MP main camera and a 48MP ultra wide camera. The main camera performs similarly to the iPhone Air and a notch below the iPhone 18 Pro, though the gap is not dramatic. There is no dedicated telephoto camera, which is a meaningful omission at this price. The outer display has a 12MP Center Stage camera, and the inner display has a 4MP under-display FaceTime camera. The under-display camera is adequate for video calls and quick selfies, but not for demanding photography.

The foldable design creates new camera workflows. Users can unfold the phone and use the outer display as a viewfinder while shooting with the higher-quality rear cameras. A feature called Duo Preview lets subjects see a live preview when someone else is taking their photo. Smart Take uses on-device AI to capture a photo when everyone is posed, removing the need to race against a timer.

Battery, speakers, MagSafe, and performance

The iPhone Duo uses a dual-battery architecture, with one battery in each half of the device. Apple manages charging, discharging, and balancing so the two cells operate like a single battery. Apple rates the device for up to 44 hours of video playback on the outer display, up to 31 hours on the inner display, and up to 24 hours of typical mixed use. In testing, the Duo lasted through a normal day with about 20% remaining, with roughly 60% of usage on the inner screen and 40% on the outer screen.

The speakers are less impressive. Apple optimized placement and tuning for Spatial Audio in folded, open, seated, and tent configurations, but the sound is thinner and quieter than on the iPhone 18 Pro. MagSafe support is included, and existing stands and chargers work. Accessories such as wallets and batteries may hang off the bottom because the Duo is shorter than a standard iPhone.

Performance is essentially identical to the iPhone 18 Pro because both use the A20 Pro chip. Everything feels fast and fluid, including Split View. Apple developed a custom thermal architecture with a vapor chamber and flexible graphite to spread heat across the two halves. The Duo can run a bit hotter than the iPhone 18 Pro, but not in a way that affects performance. One small omission is the Action button. The Duo has a Camera Control instead, and users who rely on the Action button for shortcuts will miss it.

Pricing and trade-offs

At $1,999, the iPhone Duo is an expensive first-generation product. It gives up Face ID and a telephoto camera compared with the iPhone 18 Pro, and some software features still need to catch up with the hardware. For most buyers, a traditional iPhone remains the more practical purchase today. The Duo is still a remarkable piece of engineering: the hinge feels excellent, the shorter footprint is easier to carry than expected, and the inner display turns the device into a small tablet when needed. The outer screen is far more capable than its size suggests, and the ability to move between a compact phone and a large-screen device without compromise is the defining achievement of the design.

Source: 9to5Mac News

Key facts at a glance

• The iPhone Duo is the first iPhone to support Apple Pencil, with support arriving through the iOS 27.2 software update.
• Only the Apple Pencil (USB-C) is compatible, and it works on both the outer and inner displays.
• Fast charging can reach 50 percent in about 20 minutes, but the iPhone Duo ships with only a USB-C cable, not a power adapter.
• Recommended accessories include a MagSafe car mount, protective cases, HomeKit cameras for StandBy mode, and AirPods Pro 3 as a bonus.

Apple Pencil adds handwriting and sketching to the foldable experience

The iPhone Duo is the first iPhone to offer support for Apple Pencil, a change that gives the foldable a capability previously reserved for the iPad. Users will be able to jot handwritten notes, sketch ideas, mark up documents, and draw directly on the device. Apple says the feature will arrive next week through the iOS 27.2 software update, so early buyers should expect to install the update before the Pencil becomes useful.

Compatibility is unusually specific. Apple says the iPhone Duo requires an Apple Pencil (USB-C), and no other Pencil models are compatible. That detail matters for anyone who already owns an older Apple Pencil or is considering a used model. The USB-C version is the only safe choice. The current price is $69, down from $79, which makes the stylus a more attractive add-on for note-takers and artists.

Apple Pencil support works on both the outer display and the larger inner display. The inner screen is likely to be the more natural canvas for sketches and longer notes, while the outer screen could be handy for quick reminders or signatures. Apple also says the design of its first foldable was inspired by the proportions of a standard A-series paper notebook. That design choice reinforces the idea that the iPhone Duo can function as a digital notebook, especially when paired with a stylus.

Fast charging needs a capable power adapter

Fast charging is one of the iPhone Duo’s more practical features. Apple says the device can charge up to 50 percent in just 20 minutes, which is useful for anyone who needs a quick top-up before heading out. But that speed depends on the power adapter. The iPhone Duo does not come with a power adapter in the box. It ships with a USB-C cable only, continuing a packaging approach Apple has used across several product lines.

Many households already have 20W chargers lying around, and those will still charge the phone. However, they will not deliver the fastest possible speeds. Apple’s 40W Dynamic Power Adapter with 60W Max supports iPhone Duo fast charging, and it is currently priced at $34, down from $39. A strong third-party alternative is Anker’s 100W USB-C Charger with a 5FT cable, which is listed at $29. The Anker option offers more headroom and could also be useful for charging a laptop or tablet, making it a flexible travel accessory.

Buyers should pay attention to the difference between a basic charger and one that supports the necessary fast-charging standards. The iPhone Duo’s charging speed is a headline feature, but it is only as good as the adapter connected to it. For people who want the quickest results, the Apple 40W adapter or the Anker 100W model are the two most direct recommendations.

A MagSafe car mount keeps the large screen secure on the road

For drivers, a MagSafe car mount is one of the most useful accessories for any iPhone, and the iPhone Duo should benefit from the same approach. The ESR MagSafe Car Mount is a long-standing favorite. It uses a strong magnetic attachment to hold the phone securely, and it clips to an A/C vent so the screen remains visible and easy to reach. The current price is $19, down from $29.

A secure mount matters even more with a foldable. The iPhone Duo’s larger inner display could make navigation, media controls, and hands-free calling easier to see at a glance. When paired with CarPlay, the phone can show full step-by-step directions while the in-vehicle screen displays the navigation map. That split-view arrangement is already useful on a standard iPhone, but the iPhone Duo’s larger screen could make it more comfortable and more informative.

The vent-clip design also avoids blocking the windshield, which is important for safety and for drivers who live in areas with strict mounting rules. A magnetic mount is not a substitute for safe driving habits, but it does keep the phone in a predictable place instead of sliding around the cabin. For anyone who uses navigation daily, the ESR mount is a low-cost upgrade.

Cases balance protection, price, and foldable-specific needs

Foldable phones need protection that accounts for hinges, moving parts, and two separate displays. The iPhone Duo already has a range of third-party cases alongside Apple’s first-party options. The choices range from budget-friendly models with built-in stands to premium folio designs. Buyers should consider how often they will use the outer screen, whether they need MagSafe compatibility, and whether a kickstand is worth the extra bulk.

The SPIDERCASE Magnetic Case is priced at $29 and includes a built-in stand and screen protector. That combination makes it a strong value option for someone who wants basic protection and hands-free viewing without spending a lot. The Spigen Ultra Hybrid Pro is priced at $39 and offers a clear case that is MagSafe-compatible. A clear design can show off the iPhone Duo’s finish while still providing daily protection, and MagSafe compatibility keeps accessory options open.

Apple’s own offerings sit at the higher end. The Apple iPhone Duo Folio with Kickstand is priced at $129, while the Apple iPhone Duo Case is $79. The folio adds a cover and kickstand, which could appeal to buyers who want a more notebook-like experience. The standard Apple case is a simpler option for people who prefer first-party fit and finish. With both third-party and Apple cases available at launch, buyers can choose between saving money and paying for Apple’s design language.

HomeKit cameras make the most of StandBy mode

The iPhone Duo will include a powerful upgrade to the current iPhone StandBy feature. When the device is in standing mode, it can display a variety of StandBy views even when it is not charging. That is a meaningful change because StandBy has traditionally been associated with a charging phone on a stand. The iPhone Duo’s foldable design makes the standing orientation more natural, and the software is designed to take advantage of it.

Those StandBy options include Home Camera and Home Modular views, both of which provide access to smart home accessories. The Home Camera view should be especially useful for keeping track of security cameras throughout a home at a glance. Instead of opening an app and navigating through camera feeds, a user could leave the iPhone Duo in standing mode and see selected cameras on the display. The Home Modular view could surface other smart-home controls and status information in a similar always-visible format.

For HomeKit-compatible cameras, the Aqara G100 2K security camera is a budget-friendly option at $39. It offers a low-cost way to add a camera feed to the StandBy setup. The eufy 4K security camera is priced at $69 and adds 4K support without becoming too expensive. Both cameras are recommended for users who want to monitor a room, entryway, or outdoor area from the iPhone Duo’s standing display. As with any home camera, placement and privacy settings should be considered carefully.

AirPods Pro 3 are a bonus everyday accessory

AirPods Pro 3 are not specific to the iPhone Duo, but they are a natural companion for anyone building out an Apple accessory kit. If a buyer does not yet own AirPods, or is holding on to an older model, the latest Pro earbuds are a strong choice. They are a personal favorite for all-day use, especially because the noise cancellation helps with focus at work. The latest features add value for calls, media, and daily listening.

AirPods Pro 3 are currently priced at $199, down from $249. That discount makes them more appealing as an add-on purchase alongside the iPhone Duo. The earbuds are not required to use the foldable, but they complement the device well for commuting, travel, and desk work. The current discount also makes them easier to justify for buyers who want noise cancellation and a seamless audio experience across Apple devices.

Source: 9to5Mac News

Zcash has long offered shielded transactions that hide sender, receiver, and amount using zero-knowledge proofs. To make selective disclosure possible, the network supports viewing keys: cryptographic material that lets a holder see transaction details without gaining the ability to spend funds. Spending keys remain the ultimate control mechanism. That separation is central to the new proposal. If a refund address is posted publicly, the concern is that it could be used as a starting point for linking swap activity to a broader payment history. A per-swap viewing key would keep that disclosure narrow, so a single refund event does not become a master key for a user’s entire financial trail.

NEAR Intents is designed around a simple idea: a user states what they want to trade, and competing solvers execute the trade. The model can improve pricing and reduce friction across chains, but it also creates operational touchpoints. If a swap fails, the user may need a refund address to recover funds. That address can be visible on NEAR’s public chain. Over time, analysts can cluster addresses, trace repeated interactions, and infer relationships. In a world with powerful quantum computers, data that looks harmless today could be re-examined later. Zakura’s proposal addresses that harvest-now, analyze-later risk by making each swap’s disclosure scope independent.

Key facts at a glance

• Zakura wants a separate viewing key for every ZEC swap made through NEAR Intents.
• The change is meant to limit the payment history that a future quantum computer could uncover from a publicly posted refund address.
• Spending keys would stay with users.
• Private information retrieval would help wallets locate refunds without revealing which addresses they are checking.
• Compatible wallets could adopt the open-source feature within weeks.
• The work is timed ahead of a January target for quantum-resistant payments from Zcash developers.

Private information retrieval, or PIR, is the second part of the design. In a normal refund lookup, a wallet might ask a server whether a particular address has a pending refund. That query can reveal the address being checked, even if the transaction itself is private. PIR allows a wallet to retrieve information from a database without telling the server which item it requested. In this context, it could let a wallet check refund status while hiding the set of addresses under consideration. The combination matters: per-swap viewing keys limit what can be disclosed, while PIR limits what is leaked during the lookup process. Together, they aim to prevent the refund workflow from becoming a surveillance channel.

Keeping spending keys with users is a deliberate design choice. The proposal does not turn Zakura, NEAR Intents, or any solver into a custodian. A viewing key can reveal transaction details for a defined purpose, but it cannot move funds. That distinction is important for users who want privacy without surrendering control. It also gives wallet developers a clear boundary: they can implement refund discovery and disclosure features without holding the keys that authorize spending. If the feature is open source, compatible wallets could integrate it without waiting for a centralized rollout, which may explain why the team believes adoption could happen within weeks rather than months.

The quantum threat to public-key cryptography is still prospective, but it is serious enough that governments, standards bodies, and blockchain developers are preparing for it. A sufficiently powerful quantum computer running Shor’s algorithm could break elliptic-curve cryptography, which underpins many digital signatures and key exchanges. Zcash, like many cryptocurrencies, relies on elliptic-curve constructions in parts of its protocol. Quantum computers could also weaken symmetric cryptography, though the impact depends on key sizes and algorithms. The more immediate concern is data collection: an adversary can record public blockchain data today and wait for future decryption or analysis capabilities. Privacy systems that depend on long-term secrecy must therefore consider not only current attackers but also future ones.

Zcash developers have set a January target for quantum-resistant payments, according to the proposal’s context. That target suggests the ecosystem is already moving toward post-quantum signature schemes and related upgrades. Zakura’s fix is narrower. It does not make every Zcash transaction quantum-proof, and it does not replace the need for protocol-level changes. Instead, it tries to reduce the privacy damage that could occur at the edges of the system, where swaps, refunds, and wallet queries meet public chains. In that sense, the proposal is a complement to broader quantum-resistance work rather than a substitute for it.

Wallet adoption will determine whether the proposal has practical effect. A privacy feature that exists only in a specification does not protect users. The team says compatible software wallets could adopt the open-source feature within weeks. That timeline is plausible if the change can be implemented as a wallet-level upgrade rather than a consensus change. Wallets would need to generate per-swap viewing keys, manage refund discovery through PIR, and present disclosure controls in a way users can understand. The hard part may be user experience: privacy features often fail when they add friction. A design that hides complexity while preserving user control has a better chance of being used by default.

Several challenges remain. PIR systems can be computationally expensive, and performance may vary depending on the database size and query pattern. Wallets must also guard against metadata leaks outside the refund lookup itself, including timing, IP addresses, and repeated query behavior. Per-swap viewing keys add key management overhead. If users lose a viewing key, they may lose the ability to audit a refund; if they share it too broadly, the privacy benefit shrinks. The proposal’s success will depend on clear defaults, robust recovery options, and audits. It will also depend on solvers and infrastructure providers respecting the separation between viewing and spending authority.

The broader significance is that privacy is becoming a cross-chain problem. A shielded transaction can be private on one network, but a swap or refund on another network can create a public footprint. Attackers do not need to break Zcash’s cryptography if they can correlate addresses, timing, and refund behavior across chains. That is why per-swap keys and private refund lookups matter. They treat the public chain as a place where metadata can accumulate, and they try to prevent a single operational address from becoming a permanent identifier. As cross-chain swaps grow, these edge cases will matter more, not less.

Privacy advocates are likely to welcome the proposal because it strengthens selective disclosure without weakening spending security. Regulators and compliance teams may also see value in a design that allows refunds to be audited when necessary while limiting unnecessary exposure. The key question is whether the feature can be standardized across wallets and swap providers. If it remains optional and inconsistently implemented, users may still fall back to reusable refund addresses. If it becomes a default in compatible wallets, it could raise the baseline privacy of ZEC swaps without requiring users to understand viewing keys or PIR.

The proposal arrives as Zcash developers pursue a January target for quantum-resistant payments, but the two efforts operate on different layers. The January work addresses the cryptographic foundations of payments. The Zakura proposal addresses how privacy is preserved when ZEC moves through NEAR Intents and refunds are posted publicly. For users, the practical result could be a swap flow that keeps spending keys local, reveals only what a specific refund requires, and avoids turning a public refund address into a map of past transactions.

Source: Coindesk News

The patch addresses issues in Ripple's proposed uncollateralized lending design, a feature that has been sitting in the amendment queue without the consensus needed to switch on. While the repair work moves toward activation, the lending amendment itself holds only 19 validator votes and the related vault amendment holds 21 — both short of the 29-vote threshold, leaving lending 10 votes shy of going live.

A fix that arrives before the feature does

On the XRP Ledger, code changes and feature activations travel on separate tracks. Amendments are proposals that validators vote on over successive ledger close times; once an amendment sustains the required supermajority for a set period, it becomes part of the network's rules. Bug-fix packages, by contrast, are typically treated as maintenance and can be approved on their own schedule.

That separation is why the Oct. 23 activation is possible at all. Validators are not voting to turn lending on. They are voting to correct defects in code that would govern lending if and when the relevant amendments clear their own hurdles. The practical effect is that the ledger's rulebook gets cleaned up ahead of time, so the feature does not launch with known flaws baked in.

It is a sequence that cuts against the usual pattern in crypto, where features often ship first and patches follow under pressure. Here, the repair work front-runs the rollout, leaving the ledger with a lending engine that is better documented and safer on paper than the voting tally currently allows it to be in practice.

The lending amendment remains stuck at 19 votes

The gap between the two tracks is the central wrinkle in the story. Ten more validator votes are needed before the lending amendment can take effect, and eight more are needed for the vault amendment. Until then, the lending system — including the vault structures that would hold collateral and manage credit lines — stays dormant regardless of how many bugs are fixed.

Uncollateralized lending is a sensitive proposal for a ledger best known for fast, low-cost settlement. Extending credit without collateral requires the protocol to encode assumptions about identity, repayment behavior, and default handling that most decentralized networks have historically avoided. Validators tend to move carefully on amendments that could introduce bad debt into the system, and the vote count reflects that caution.

The vault amendment is the less contentious of the two, sitting closer to the threshold. Vaults are a more general primitive: they can hold assets on behalf of users and serve as building blocks for a range of on-chain financial products beyond lending. That broader utility may explain why it has attracted more support than the lending amendment itself.

Closing a token-permission loophole

Among the defects being corrected is a permissions flaw with implications well beyond lending. The patch closes a loophole that could have allowed certain accounts to create tokens despite holding permission only to destroy them.

The XRP Ledger supports granular account permissions, letting an account delegate specific capabilities to other accounts — the ability to issue a token, for instance, or the ability to claw it back. The bug meant the permission model could be circumvented in one direction: an account restricted to destroying tokens could, under certain conditions, mint them instead. That is the kind of asymmetry that turns a tightly scoped delegation into an uncontrolled issuance channel.

The fix matters because permission systems are only as strong as their weakest branch. A token issuer that grants a limited role to a treasury operator, a market maker, or an automated contract relies on the protocol enforcing the boundary exactly as written. Any path that lets a destroy-only account produce new supply undermines the entire premise of delegated control.

Vault rounding and loan-default rules get revised

The package also revisits two areas of lending logic that tend to cause trouble in production: vault rounding and loan-default handling.

Rounding is an unglamorous but consequential detail in any share-based vault. When deposits and withdrawals are converted into vault shares, rounding decisions determine whether value leaks out of the pool over time — a phenomenon often described as dust accumulation or share-price drift. Repeated rounding in the protocol's favor can quietly dilute holders; rounding the other way can let users extract value through repeated small transactions. The revised rules aim to make those calculations consistent and resistant to exploitation.

Loan-default rules are equally foundational. A lending protocol needs a clear, deterministic answer to what happens when a borrower fails to repay: when the loan is declared delinquent, how collateral or vault assets are handled, and what remedies remain available to the parties involved. Ambiguity in that logic is dangerous because it invites disputes exactly when positions are already under stress. Tightening the rules before activation reduces the chance that the first real default becomes a test case for undefined behavior.

Why amendments move slowly by design

The 29-vote threshold is not arbitrary. It is intended to ensure that a rule change has broad backing across the validators who independently maintain the network's ledger history. Amendments that clear the bar are baked into the protocol; amendments that do not simply expire without altering anything. That design protects against narrow coalitions rewriting the rules, but it also means genuinely useful features can idle for long stretches while support accretes.

For the lending proposal, the delay has an upside. Every week the amendments sit below threshold is another week of review, testing, and bug-fixing on code that has not yet been exposed to real capital. The Oct. 23 package is a direct product of that window. Had lending activated earlier with the same defects, the token-permission loophole and the vault rounding issues would have been live problems rather than preemptive fixes.

What the vote count needs to change

Reaching activation requires ten additional validators to back the lending amendment and eight to back the vault amendment. Neither number is out of reach, but neither is guaranteed. Validator operators weigh factors that go beyond the technical merits: the maturity of the code, the behavior of similar systems on other networks, the quality of the default and liquidation logic, and the broader appetite for credit products on a ledger whose primary use case has been payments and tokenized assets.

Until those votes materialize, the ledger's lending machinery will remain a well-maintained but unused subsystem, patched ahead of schedule and waiting on consensus. The bug-fix package is expected to take effect around 09:34 UTC on Oct. 23, closing the permission loophole, revising vault rounding and loan-default rules, and leaving the feature it repairs still ten votes short of ever being switched on.

Source: Coindesk News

The security firm found that about $900 million of the quarterly losses came from compromised suppliers or developer tools, uninstalled security fixes, or components excluded from security reviews. It also found that 77% of third-quarter losses involved compromised suppliers, delayed fixes, or systems outside the latest security reviews.

Key facts

• Hackers stole $1.18 billion from crypto projects in the third quarter, up 55% from the previous quarter.
• About $900 million in losses stemmed from compromised suppliers or developer tools, uninstalled security fixes, or components excluded from security reviews.
• 77% of third-quarter losses involved compromised suppliers, delayed fixes, or systems outside the latest security reviews.
• The security firm warned that audit badges can give investors a false sense of security because reviews often omit wallets, signing systems, and third-party software that attackers target.

The findings land at a difficult moment for crypto security. Projects have spent years promoting third-party audits as a sign of maturity, and investors have learned to look for audit reports before committing capital. But the quarterly data suggests that a clean report can coexist with unpatched software, unchecked vendors, and unprotected signing infrastructure.

Audit badges and the scope gap

Security audits are usually scoped to code. A project may commission a review of a smart contract, a token contract, a bridge, or a lending protocol, and the resulting report can be detailed and professionally produced. Yet the review may not cover the wallets that hold protocol funds, the signing systems that authorize transactions, the web applications that interact with users, or the third-party services that support the stack.

That scope gap matters because attackers follow the path of least resistance. A flaw in a reviewed contract may be difficult to exploit; a compromised developer tool or an unpatched server may be much easier. When investors see an audit badge, they may assume the whole system has been tested. In practice, the badge may cover only one layer of a much larger attack surface.

Supply chain attacks move up the list

The concentration of losses around compromised suppliers and developer tools reflects a broader shift in crypto attacks. Rather than trying to break a well-reviewed contract directly, attackers can target the people, software, and services that build and maintain it. A compromised package, a malicious update, a phishing attack against a developer, or a breach at a hosting provider can give an attacker a route into the project without ever touching the audited code.

Supply chain risk is hard to eliminate because modern crypto projects depend on many external components. Teams rely on open-source libraries, cloud services, code repositories, continuous integration pipelines, domain registrars, wallet software, and communications tools. Each dependency can become an entry point. Attackers know that a single weak vendor can provide access to multiple projects at once, which makes supplier security a systemic issue rather than a simple procurement problem.

Delayed fixes leave known doors open

The quarterly review also highlighted uninstalled security fixes. A vulnerability does not need to be unknown to be dangerous. If a patch is available but not deployed, the window for exploitation stays open. Crypto teams often move quickly to launch features but may move more slowly on patch management, especially when upgrades require coordination, testing, or governance approval.

Delayed fixes can be particularly damaging in decentralized systems where upgrades are complex. A smart contract may be immutable, but the surrounding infrastructure is not. Servers need updates, dependencies need patching, and access controls need review. When those tasks slip, an audit that was accurate at the time of review can become outdated. The report remains on the website, but the system it described has changed.

What audits often leave out

The security firm's warning focused on wallets, signing systems, and third-party software as common omissions. These are critical components. A wallet holds assets; a signing system authorizes transfers; third-party software connects the project to users and markets. If those systems are not in scope, an audit may provide limited assurance about the safety of funds.

Investors may not see the boundaries of an audit. Reports can be technical, lengthy, and difficult to compare. A project may display a logo or a short badge rather than a clear statement of what was reviewed, when it was reviewed, and what remained outside scope. That simplification can create a false sense of security. It can also reward projects for appearing secure rather than for maintaining security over time.

Why the numbers matter for investors

The $1.18 billion total is large, but the composition of losses may be more important than the headline figure. If most losses come from overlooked suppliers, uninstalled fixes, and systems outside the latest reviews, then the biggest security gains may come from operational discipline rather than another code review. That means vendor management, patch tracking, access control, key management, monitoring, and incident response deserve as much attention as smart contract testing.

For investors, the practical lesson is to treat an audit as one data point rather than a guarantee. Useful questions include: What exactly was reviewed? Was the review limited to contracts or did it include wallets, signing systems, and third-party dependencies? How old is the review? Have material changes been made since? Were findings fixed and verified? Does the project have a process for monitoring threats after the review ends?

The marketing problem in crypto security

Audits are also a marketing tool. Projects use them to build trust with users, exchanges, and investors. That creates an incentive to present audits prominently and to gloss over limitations. The result is a market where a badge can signal diligence even when the underlying security program is narrow. The quarterly loss data challenges that signal.

Security is not a one-time event. It is a continuous process that includes code review, threat monitoring, vulnerability management, supplier oversight, and incident response. A project can pass an audit and still fail operationally. A project can also maintain strong security without displaying a perfect badge. The distinction matters because attackers do not respect audit scope; they look for any gap they can reach.

Building a more realistic security posture

Projects can reduce the risk of overlooked systems by expanding the scope of reviews and repeating them when the stack changes. They can require evidence that critical findings have been fixed, not just acknowledged. They can assess vendors and developer tools with the same seriousness as smart contracts. They can separate signing authority, limit access, and monitor for unusual activity. They can test incident response plans before an attack rather than during one.

Investors can also push for better disclosure. A clear scope statement, a remediation record, and a description of post-audit monitoring would be more useful than a logo. Those disclosures would not eliminate risk, but they would help distinguish a project that has reviewed a narrow piece of code from one that has a broader security program.

The quarterly figures show a security market that is still learning the difference between proof of review and proof of resilience. With $1.18 billion lost in three months and 77% of losses tied to compromised suppliers, delayed fixes, or systems outside the latest reviews, the evidence suggests that the most dangerous gaps are often the ones that never appear in the audit report.

Source: Coindesk News

The result has been greeted as a genuine discovery. It does not, by itself, settle every question about the NIST elliptic curves or about the NSA's role in shaping them. It does make one long-running suspicion considerably weaker: that the numbers were chosen to conceal a backdoor. For the millions of systems that use NIST curves, the practical takeaway is that no immediate action is required. For Bitcoin, the answer is simpler still: the leading cryptocurrency was never on the list.

Key facts at a glance

• An Anthropic AI model traced five unexplained seeds for NIST elliptic curves to an ordinary random-number generator.
• The seeds appear to have been generated with timestamps from 1997 and 1998.
• The work took less than 30 minutes.
• The finding makes a hidden NSA backdoor less likely.
• Users of the NIST curves do not need to take action because of the discovery.
• Questions remain about how some curve elements were selected.
• Bitcoin was not affected because it uses secp256k1, not a NIST curve.
• Sufficiently powerful quantum computers could still pose a separate threat to Bitcoin and other public-key systems.

Why five seeds mattered for so long

Elliptic-curve cryptography relies on parameters that define the shape of a curve and the arithmetic performed on it. The security of the system depends on those parameters being chosen in a way that does not introduce hidden structure. A seed is a starting value from which such parameters can be derived. If the seed is generated by a flawed or malicious process, the resulting curve could in theory contain a weakness that its designers know about and outsiders do not.

NIST published a set of elliptic curves in the late 1990s after a process that involved the NSA. The curves became widely used in standards and commercial products. The agency supplied seeds used to derive some parameters. For years, cryptographers noted that the provenance of those seeds was not fully documented. Without knowing how the seeds were produced, observers could not rule out that they had been selected to make the curves easier to attack. That uncertainty turned five numbers into a durable source of suspicion.

The backdoor suspicion and what changed

The suspicion did not come from nowhere. It was fueled by the later disclosure of weaknesses in Dual_EC_DRBG, a random-number generator standardized by NIST that contained a suspected NSA backdoor. That episode trained a generation of security researchers to treat unexplained constants in standards with suspicion. The NIST curves were not Dual_EC_DRBG, and the two issues are distinct, but the broader anxiety spilled over. If the seeds behind the curves had been manufactured, critics reasoned, the curves themselves might carry a hidden weakness.

The new finding does not prove that every parameter is benign. It does show that the five seeds match a mundane process: a random-number generator fed with timestamps from 1997 and 1998. That is exactly the kind of unglamorous origin story that is hard to square with a deliberate attempt to plant a backdoor. A hidden trap would more likely require carefully chosen numbers, not values that can be reproduced from dates. The discovery also fits the historical context: in the late 1990s, timestamps were a common source of entropy in software and hardware random-number generators, even if they were not ideal by modern standards.

What the AI model actually did

The model searched for a simple explanation. Instead of treating the seeds as opaque cryptographic artifacts, it considered the possibility that they had been produced by a conventional random-number generator seeded with a date or time. It then tested candidate timestamps and generator designs until it found a match. The entire process took less than half an hour. That speed is part of why the result landed as a surprise: the puzzle had survived not because it required enormous compute, but because no one had framed it in quite the right way.

The episode is a reminder that AI systems can be useful in cryptanalysis and standards archaeology when they are pointed at well-defined pattern-matching problems. It is not a claim that AI has broken elliptic-curve cryptography or that it can find backdoors in arbitrary systems. The model did not attack the curves themselves. It reconstructed a likely generation process for five seeds. That narrower achievement is still significant because it replaces speculation with a testable explanation.

What it means for NIST curve users

For organizations using NIST curves, the finding is reassuring but not a reason to become complacent. The curves remain widely deployed in TLS certificates, secure messaging, hardware security modules, government systems, and enterprise software. The discovery makes a deliberate backdoor in the five seeds less likely. It does not mean that every implementation is secure, that every parameter choice is beyond doubt, or that all elliptic-curve systems are free of risk. Implementation bugs, side-channel attacks, weak key generation, and protocol design errors remain separate concerns.

Security teams should interpret the result as a reduction in one specific area of uncertainty, not as a clean bill of health for every system that touches a NIST curve. The original questions about how some elements were selected have not disappeared entirely. Cryptographers may still want more documentation and more independent verification. But the most dramatic explanation, a deliberately planted NSA backdoor in the seeds, now has a much higher burden of proof.

Why Bitcoin was never exposed

Bitcoin's relationship to this story is often misunderstood. The cryptocurrency does not use the NIST curves at the center of the mystery. Its signatures rely on secp256k1, a curve defined by the Standards for Efficient Cryptography Group. That curve is simpler in structure, and its parameters are widely documented. It was not supplied through the same NIST process, and it does not carry the same historical baggage. The five seeds at issue in the new finding are not Bitcoin seeds, and the discovery does not change Bitcoin's signature scheme.

That does not make Bitcoin invulnerable. It means that this particular backdoor scare was never a Bitcoin problem. The distinction matters because headlines about cryptography and spy agencies can easily blur into generic claims that a major cryptocurrency was affected. In this case, the technical facts are clear: Bitcoin's curve is different, and the mystery seeds belong to a different family of standards.

The quantum problem is still separate

The NIST curve mystery and the quantum threat to Bitcoin are often discussed in the same breath, but they are distinct issues. The new finding says nothing about the difficulty of breaking elliptic-curve cryptography with a sufficiently powerful quantum computer. Bitcoin's secp256k1 signatures, like many public-key systems, could eventually be vulnerable to a quantum computer capable of running Shor's algorithm at the necessary scale. That remains a long-term engineering and research challenge, not a present-day break.

The quantum question is also not solved by better documentation of 1990s seeds. It requires cryptographic migration, new signature schemes, and careful upgrades across wallets, exchanges, and protocols. The AI-assisted discovery is important because it clarifies a historical mystery and reduces one source of distrust. It does not change the timeline for quantum risk, and it does not provide a shortcut around the work that post-quantum migration will require.

What remains unknown

Even with a convincing explanation for the five seeds, some questions remain. The finding does not automatically explain every parameter in every NIST curve. It does not document the full internal process by which the NSA and NIST selected the curves. It does not prove that no other standards from the era contain similarly unexplained values. And it does not resolve the broader debate over how much influence intelligence agencies should have in public cryptographic standards.

The most useful lesson may be procedural. Standards bodies should document seed generation, preserve randomness sources, and publish enough information for independent researchers to reproduce parameter choices. When such records are missing, suspicion fills the gap. In this case, an AI model filled part of that gap by finding a mundane explanation. That is a better outcome than a hidden backdoor, but it is also a reminder that transparency in cryptography is not optional. The discovery closes a chapter on five numbers, while leaving the larger work of verification and trust in standards very much open.

Source: Coindesk News

The result is an engineering advance, not a break. It does not mean a quantum computer can now crack bitcoin wallets, and it does not mean a future attack has become 40 times cheaper. What it does show is that one stubborn hardware bottleneck — producing high-fidelity magic states at useful rates — is moving forward. For bitcoin, especially early Satoshi-era coins whose public keys are exposed on the blockchain, that is a signal worth tracking.

Key Facts

• Google said Willow reduced errors by about 40-fold when preparing magic states, a required resource for advanced quantum calculations.
• The experiment achieved about 99.99 percent fidelity among states that passed inspection, but only about 8 percent of preparation attempts succeeded.
• Substantial scaling challenges remain, including the need to produce reliable magic states far more consistently and at much higher rates.
• The findings do not make a bitcoin attack 40 times cheaper, but they advance an engineering method that could help future machines sustain roughly 500,000 reliable magic states per second under one modeled attack.

Why Magic States Matter

Quantum computers are often described by qubit counts, but qubits alone are not enough. Most quantum operations are relatively easy to implement in ways that are tolerant of some noise. The difficult operations — those needed to make a quantum computer universal — require additional resources. Magic states are one of the best-known ways to supply those resources. They are consumed during computation, much like fuel, and their quality determines whether a large quantum algorithm can run reliably.

Preparing magic states is hard because the process must produce quantum states that are pure enough and consistent enough for error correction to work. If too many prepared states are flawed, the computer must either discard them or spend enormous resources distilling better ones from lower-quality inputs. A 40-fold reduction in preparation errors is therefore meaningful. It suggests that one part of the error-correction stack can be improved, even if the full system remains far from practical cryptographic relevance.

The nuance is in the experiment's success rate. The 99.99 percent fidelity figure applies to states that passed inspection. Only about 8 percent of attempts succeeded. That means the result is a proof of improved quality under post-selection, not a demonstration of a magic-state factory that can run continuously at scale. For a cryptographic attack, post-selection is not enough. A quantum computer would need to produce enormous numbers of reliable magic states on demand, with failure rates low enough that the overall computation does not collapse.

What It Means for Bitcoin

Bitcoin's security relies on elliptic-curve cryptography, specifically the secp256k1 curve. A bitcoin private key is used to sign transactions, and the corresponding public key allows the network to verify those signatures. In most modern bitcoin addresses, the public key is not revealed until a coin is spent. That adds a layer of protection: an attacker who wants to steal coins must usually derive the private key from a public key that becomes visible only when a transaction is broadcast.

Satoshi-era bitcoin is different. Many early outputs used pay-to-public-key scripts, which expose the public key directly on the blockchain. Those coins have been dormant for more than a decade, and their public keys are visible to anyone. If a sufficiently powerful quantum computer could run Shor's algorithm at the required scale, it could derive the private key from the public key. That would put those outputs at risk, along with any other coins whose public keys are exposed through address reuse, spending patterns, or older script types.

This is why the quantum debate in bitcoin has focused on exposed public keys rather than on all coins equally. A quantum attack would not need to break every wallet at once. It could target the subset of outputs whose public keys are already known. The scale of that subset is debated, but early Satoshi-era coins are a symbol of the problem because they are old, valuable, and in some cases tied to public keys that have been visible for years.

The Willow result does not change that risk overnight. It also does not provide a timeline for when a cryptographically relevant quantum computer might exist. What it does is strengthen the case that error correction is progressing. The gap between a laboratory demonstration and a machine capable of attacking bitcoin remains enormous, but the gap is being narrowed at the component level.

The 500,000 Magic States Per Second Benchmark

One modeled attack described in quantum computing research would require roughly 500,000 reliable magic states per second. That figure is not a universal constant, and it depends on assumptions about the quantum computer's architecture, the error-correction code, the algorithm, and the target. But it illustrates the scale of the challenge. A machine that can prepare a few high-quality magic states in a controlled experiment is not close to sustaining half a million per second with the reliability needed for a long computation.

The Willow experiment's 8 percent success rate is especially important here. If only about one in twelve preparation attempts succeeds, the system must either run many attempts in parallel or improve the preparation method. Parallelism adds hardware, control complexity, and heat. Improving the method requires better materials, better calibration, and better error correction. The 40-fold error reduction is a step toward the latter, but it does not solve the throughput problem by itself.

That distinction is why researchers caution against reading the result as a 40-fold improvement in attack feasibility. A quantum attack is not a single operation. It is a massive, error-corrected computation that consumes magic states continuously. A lower error rate in one preparation step can reduce overhead, but the attack still requires a large logical quantum computer with thousands or millions of physical qubits, depending on the design. The latest result is one data point in that long engineering road.

Error Correction Is the Real Battleground

For years, quantum computing milestones were measured by qubit counts. More recently, the focus has shifted to error correction. Physical qubits are noisy, and useful computations require bundling many physical qubits into logical qubits that can detect and correct errors. Google's Willow chip has been central to that shift. The latest magic-state result fits into a broader effort to make error correction more efficient and more predictable.

Magic states are part of the non-Clifford resource story. Clifford operations can be implemented fault-tolerantly with relative ease in many architectures, but they are not universal on their own. To run arbitrary quantum algorithms, a computer needs non-Clifford operations, and magic states are a leading way to supply them. Improving magic-state preparation therefore improves the practical viability of universal fault-tolerant quantum computing. That is why the result is relevant beyond bitcoin, even though bitcoin is the most prominent financial system whose cryptography could be affected.

The experiment also highlights how quantum computing progress often looks incremental. A 40-fold error reduction is significant, but it is not a headline-grabbing demonstration of a new machine. It is a careful engineering result about one component. Those component-level improvements are how large systems eventually become possible. For bitcoin, the lesson is to watch the trend rather than overreact to a single result.

What Bitcoin Developers Are Watching

The bitcoin community has long discussed post-quantum cryptography. Elliptic-curve signatures would eventually need to be replaced or supplemented by quantum-resistant schemes if the threat becomes credible. That is easier said than done. Bitcoin's consensus rules are difficult to change, and a migration would require wallets, exchanges, custodians, and users to move coins to new address types. Coins whose public keys are already exposed would be especially difficult to protect because their owners may be inactive or unknown.

There are also questions about how to handle dormant coins. If a quantum-resistant upgrade were activated, should old outputs be frozen, migrated, or left vulnerable? Those are policy questions as much as technical ones. The existence of Satoshi-era coins makes the debate more urgent because those outputs are a visible reminder that bitcoin's early design did not anticipate quantum computing. Modern address types reduce the exposure, but they do not eliminate it for coins that have already revealed their public keys.

For now, the practical advice from cryptography researchers is not to panic but to plan. The latest Willow result does not mean bitcoin is about to be broken. It does mean that the error-correction layer of quantum computing is advancing, and that long-term security planning should not assume the threat is static.

The Timeline Remains Uncertain

The most important sentence in the Willow result may be the one that limits it: the findings do not make a bitcoin attack 40 times cheaper. They advance an engineering method that could help future quantum computers sustain the roughly 500,000 reliable magic states per second required under one modeled attack. That is a meaningful distinction. It separates a real technical achievement from an overstated threat narrative.

For Satoshi-era bitcoin, the risk remains theoretical but worth monitoring. The public keys are exposed, the cryptography is known, and the quantum research community is steadily improving the components needed for fault-tolerant computation. The timeline is uncertain, but the direction of travel is clear. The question is not whether error correction will improve. It is how quickly it will improve, and whether bitcoin's governance can coordinate a post-quantum migration before a cryptographically relevant machine arrives.

Until then, the Willow result belongs in the category of long-range risk research: important, technically substantive, and not a reason to expect bitcoin wallets to fall tomorrow. The error rate fell by 40 times in one preparation step. The attack still requires a quantum computer that does not yet exist at the necessary scale. That gap is the story.

Source: Coindesk News

Die wichtigsten Fakten

• Adut Akech ist das Gesicht des neuen Valentino-Dufts „Born in Roma Intense“.
• Die Kampagne zeigt das 23-jährige Topmodel an der Seite des gleichaltrigen Anwar Hadid.
• Es ist die vierte Duftkampagne für Valentino Beauty.
• Der Flakon ist mit markanten Rockstud-Nieten verziert.
• Akech präsentiert den Duft in einer von Pierpaolo Piccioli entworfenen Traumrobe in leuchtendem „Valentino Pink PP“.
• Die Parfümeurinnen Amandine Marie und Honorine Blanc schufen „Born in Roma Intense Donna and Uomo“.
• Zu den Duftnoten zählen Vanille, ein Jasmin-Trio und Benzoe-Harz.
• Akech beschreibt ihre Beziehung zu Valentino als familiär und betont Respekt sowie Moralvorstellungen der Marke.
• Sie trägt tagsüber leichte und frische Düfte, nachts eher sinnliche und lang anhaltende Kreationen.
• Für Akech ist Parfum ein zentrales Element ihrer Schönheitsroutine.

Haute-Couture-Spirit und eine italienische Nacht

Die Inszenierung des neuen Duftes verbindet Mode, Sinnlichkeit und römisches Lebensgefühl. Akech zeigt sich in einer Robe, die von Pierpaolo Piccioli entworfen wurde und in einem intensiven Pink gehalten ist. Der Flakon greift mit den charakteristischen Rockstud-Nieten die markante Designsprache von Valentino auf. Damit wird der Duft nicht nur als olfaktorisches Produkt präsentiert, sondern als Teil einer umfassenden Haute-Couture-Welt. Die Kampagne setzt auf eine melancholische wie leidenschaftliche Stimmung, die an eine „Amour fou“ erinnert. Rom erscheint dabei als Kulisse für geheime Partys, nächtliche Begegnungen und eine luxuriöse Leichtigkeit, die in der Duftkomposition eingefangen werden soll.

Adut Akech: Ein internationales Topmodel mit klarer Haltung

Adut Akech zählt zu den prägenden Gesichtern der internationalen Modewelt. Die 1999 geborene Südsudanesin wuchs zunächst in Kenia auf und kam als Kind mit ihrer Familie nach Australien. Ihre Karriere begann sie als Teenager, ehe sie schnell zu einer gefragten Größe auf den Laufstegen und in Kampagnen internationaler Luxusmarken wurde. Heute arbeitet sie mit großen Modehäusern und Fotografen zusammen und ist regelmäßig auf den Titelseiten von Modemagazinen zu sehen. Ihre persönliche Geschichte, ihre Herkunft und ihr Engagement für Diversität haben sie zu einer Stimme gemacht, die über das Modeln hinausreicht. In der Zusammenarbeit mit Valentino verbindet sich diese Biografie mit dem Anspruch des römischen Modehauses, klassische Eleganz und moderne Offenheit zusammenzubringen.

Warum Valentino für Akech mehr als ein Auftrag ist

Im Gespräch über die Marke wird deutlich, dass Akech ihre Rolle nicht als reine Kampagnenarbeit versteht. Valentino spreche sie wegen seiner Grundüberzeugungen und Moralvorstellungen an, sagt sie. Sie respektiere die Marke, weil sie Menschen aus allen Gesellschaftsschichten respektiere. Diese Haltung sei für sie ein wesentlicher Grund, warum die Zusammenarbeit so stimmig wirke. Ihre Beziehung zu Valentino beschreibt sie als familiär. Das Modehaus habe sie von Anfang an akzeptiert. Sie liebe Pierpaolo und das Team. So wie sie von den Verantwortlichen geliebt und respektiert werde, so liebe und respektiere sie alle Beteiligten. Diese Verbindung sei über die Jahre gewachsen und habe zu vielen gemeinsamen Erfahrungen geführt.

Starke Duos und persönliche Einflüsse

Die Kampagne zu „Born in Roma“ wird von starken Duos getragen. Akech nennt auf die Frage nach den einflussreichsten Partnern in ihrem Leben mehrere Menschen. Ihre Mutter sei eine besonders prägende Frau, und gemeinsam seien sie ein starkes Duo. Auch ihre ältere Schwester beschreibt sie als Vertraute und Komplizin. Dazu kommen Freunde und ihr jetziger Partner. Akech betont, dass sie von beeindruckenden und mächtigen Menschen umgeben sei, Männern wie Frauen, die sie täglich inspirieren. Diese persönlichen Beziehungen prägen ihr Selbstverständnis und spiegeln zugleich das Motiv der Duftkampagne wider, in der Nähe, Leidenschaft und Verbundenheit eine zentrale Rolle spielen.

Das Alter Ego: eine furchtlose Frau

„Born in Roma Intense“ wird als zweites Ich des Originalduftes beschrieben. Auf die Frage, wonach sie in ihrem Alter Ego suche, antwortet Akech mit einem klaren Bild: Sie suche eine furchtlose Frau für jeden Tag ihres Lebens. Eine Frau, die sich selbst kenne und so sehr an sich glaube, dass niemand etwas anderes behaupten könne. Das gelte für alles, was sie sich vornehme. Akech beschreibt sich selbst als facettenreiche Persönlichkeit. Sie glaubt, dass jeder Mensch verschiedene Seiten in sich trage. In ihren unterschiedlichen Ichs sehe sie sich als jemanden, der sich um die Menschen kümmert, die er liebt. Zugleich könne sie, wenn eine Beziehung zu Ende gehen müsse, so tun, als hätte sie die andere Person nie gekannt. Diese Mischung aus Fürsorge und innerer Stärke passt zu dem Bild einer modernen Valentino-Frau, die Verletzlichkeit und Selbstbewusstsein miteinander verbindet.

Tag und Nacht: Wie Akech ihre Düfte wählt

Bei der Frage, wie sie entscheidet, welchen Duft sie tagsüber und welchen sie nachts trägt, nennt Akech Stimmung und Outfit als wichtigste Kriterien. Tagsüber halte sie es meist leicht und frisch. Nachts dürfe es etwas intensiver sein. Die Nacht beschreibt sie als intimen Moment, weshalb sie dann sinnlichere Düfte wähle, die eine lang anhaltende Dufterinnerung hinterlassen. Diese Unterscheidung zwischen hellen Tagesnoten und tieferen Abendnoten ist typisch für moderne Parfumroutinen. Sie zeigt, dass Düfte für Akech nicht bloß Accessoire sind, sondern Teil einer bewussten Inszenierung von Stimmung und Persönlichkeit.

Weniger ist mehr: Beauty-Routine und Parfum

Akech verrät auch, dass ihr absolutes Lieblingsschönheitsprinzip „Weniger ist mehr“ lautet. Dennoch nimmt Parfum in ihrer Routine einen wichtigen Platz ein. Ohne Parfum fühle sie sich untertags oder beim Ausgehen nicht vollständig, sagt sie. Damit wird der Duft zu einem elementaren Bestandteil ihrer täglichen Vorbereitung. Er ergänzt den minimalistischen Ansatz, den sie in der Pflege verfolgt, und setzt einen klaren, persönlichen Akzent. Für Akech ist Parfum offenbar weniger ein dekoratives Extra als vielmehr ein Gefühl von Vollständigkeit und Präsenz.

Born in Roma Intense: magnetisch, cremig und modern

Was Akech am neuen Duft besonders gefällt, ist seine Intensität. Er halte sich sehr lange auf der Haut und auf der Kleidung. Sie bekomme viele Komplimente, wenn sie ihn trage. Auch das Design des Flakons mit dem Farbverlauf und den ikonischen Nieten hebt sie hervor. Für sie ist der Flakon ein schönes Element auf dem Parfumständer oder dem Badezimmertisch. Auf die Frage, was der Duft konkret auslöse, antwortet sie, er rufe das Gefühl hervor, die kühnste und schönste Frau im Raum zu sein. Sie möchte, dass sich die Menschen umdrehen, wenn sie ihren Duft riechen, und dass jeder, der ihn wahrnimmt, neugierig wird. Gegenüber jemandem, der den Duft noch nicht riechen konnte, würde sie ihn als magnetisch beschreiben. Die Vanille, das Jasmin-Trio und das Benzoe-Harz verliehen ihm eine leuchtende, cremige und moderne Blumigkeit. Akech ist überzeugt, dass alle Frauen einen Flakon davon besitzen sollten.

Anlässe für den Duft

Für den neuen Valentino-Duft nennt Akech zwei besonders passende Gelegenheiten: eine Date-Nacht mit ihrem Partner und ein großes Event mit rotem Teppich. Beide Anlässe verbinden Intimität und Öffentlichkeit, Nähe und Auftritt. Genau diese Spannung trägt auch die Kampagne, in der Rom als Ort zwischen geheimer Leidenschaft und glamouröser Bühne erscheint. Der Duft wird damit zu einem Begleiter für Momente, in denen Präsenz, Sinnlichkeit und ein selbstbewusster Auftritt gefragt sind. Akech selbst macht deutlich, dass sie Parfum als unverzichtbaren Teil ihres Auftretens versteht, ob im Alltag oder bei besonderen Anlässen.

Source: oe24 News

La escena transcurrió en un estudio de televisión iluminado de forma impecable, con dos cámaras fijas apuntando a la entrevistadora y a su invitado. Todo parecía transcurrir dentro de los cauces habituales de una entrevista promocional: preguntas amables, respuestas ensayadas y alguna que otra anécdota ligera. Sin embargo, bastó una sola frase para que el ritmo de la conversación se rompiera en pedazos y el silencio se apoderara del set durante varios segundos que, en televisión en vivo, se sienten como una eternidad.

El youtuber y empresario mexicano Luis Arturo Villar Sudek, conocido mundialmente como Luisito Comunica, era el invitado. La conductora le pidió compartir un dato personal que nunca hubiera revelado públicamente, una de esas preguntas diseñadas para arrancar una confesión curiosa o una risa cómplice. El creador de contenido, con un gesto de seriedad absoluta, respondió: «Yo nací mujer, por ejemplo». La cámara siguió captando los rostros de ambos. Nadie dijo nada.

Fueron segundos de incomodidad pura, del tipo que el espectador percibe de inmediato en el cuerpo. La conductora lo miró fijamente, sin encontrar la salida. Él mantuvo la expresión seria, consciente del efecto de su ocurrencia. Finalmente, Luisito aclaró que se trataba de una broma y expresó que esperaba que la presentadora se lo tomara con humor. La respuesta llegó con calma, pero con una firmeza que cambió por completo el tono del intercambio: no se había reído porque consideraba que el tema era sensible.

Quién es el creador de contenido detrás de la broma

Para entender por qué una frase dicha al azar en un estudio de televisión se convierte en noticia internacional, conviene repasar quién es el personaje. Luisito Comunica nació en la Ciudad de México y construyó, desde la década pasada, uno de los imperios digitales más grandes del idioma español. Su canal de YouTube acumula decenas de millones de suscriptores y sus videos de viajes, retos y exploraciones urbanas se han convertido en referencia obligada para toda una generación de creadores latinoamericanos.

Su estilo se define por un humor espontáneo, un ritmo de edición acelerado y una cercanía con la audiencia que rompe la cuarta pared. Esa misma fórmula lo llevó a convertirse en empresario: incursionó en la gastronomía, en el sector de los alimentos empaquetados, en el comercio electrónico y en la producción audiovisual. Con el paso de los años, su figura dejó de pertenecer únicamente al universo de internet para instalarse también en los medios tradicionales, en la publicidad y en la conversación pública.

Ese tránsito explica buena parte del episodio. Un creador digital acostumbrado a la complicidad de su comunidad, a los comentarios instantáneos y a la retroalimentación inmediata, se enfrenta de pronto a un formato donde las reglas son distintas: tiempos medidos, un público heterogéneo y una audiencia que no necesariamente comparte su código de humor. Lo que en un video de YouTube funcionaría como un chiste reconocible, en televisión abierta puede convertirse en un momento de tensión.

El contraste entre dos formas de comunicar

La televisión en vivo impone condiciones particulares. No existe la posibilidad de repetir la toma, de editar el silencio o de matizar la frase con un corte rápido. Cada palabra se pronuncia ante una audiencia masiva y diversa, integrada por espectadores que no eligieron necesariamente ese contenido y que pueden interpretarlo desde marcos culturales muy distintos. En ese contexto, el margen de error se reduce a cero.

El humor de internet, en cambio, se sostiene sobre un pacto previo con la audiencia. Quien ve un video de un creador digital conoce su tono, su historial y sus límites. Sabe cuándo habla en serio y cuándo está actuando. En televisión ese pacto no existe: el espectador ocasional no tiene ese contexto y puede tomar la declaración de manera literal. La broma, entonces, queda suspendida en un territorio ambiguo donde el significado depende completamente de quien la escucha.

Esa brecha quedó expuesta con claridad durante la entrevista. La conductora actuó como lo haría cualquier profesional formado en el rigor periodístico: ante una declaración que involucra la identidad de género de una persona, optó por no reír y por señalar que el tema merece respeto. Su respuesta no fue un reproche airado, sino una observación serena que reencuadró la conversación. Y ese gesto, pequeño en apariencia, terminó siendo el verdadero centro de la polémica.

Las reacciones divididas en redes sociales

El fragmento comenzó a circular casi de inmediato en plataformas sociales, donde se multiplicaron las capturas de pantalla, los clips recortados y los comentarios. Las reacciones se dividieron en dos bloques claramente diferenciados. Por un lado, hubo usuarios que aplaudieron la postura de la presentadora y valoraron que no entrara al juego de una broma que consideraban fuera de lugar. Por el otro, aparecieron quienes defendieron al creador de contenido y calificaron la situación como una simple diferencia de percepción del humor.

Las críticas hacia el youtuber se centraron en la elección del tema. Diversos internautas señalaron que la identidad de género no es material para una broma improvisada, especialmente en un medio de alcance masivo y ante una audiencia que puede incluir personas directamente afectadas por la discriminación. Argumentaron que existen infinidad de anécdotas personales con las que se puede generar complicidad sin recurrir a un asunto delicado.

Del otro lado, sus seguidores insistieron en que se trataba de una broma evidente, aclarada segundos después por el propio protagonista, y que la reacción de la conductora fue excesivamente rígida. Algunos incluso interpretaron el silencio como una falta de adaptación al registro del invitado. Lo cierto es que el episodio ilustra un fenómeno recurrente en la era digital: la colisión entre el humor de las comunidades en línea y las normas de la conversación pública tradicional.

El debate sobre los límites del humor

Más allá del episodio puntual, el caso reabrió una discusión que aparece con regularidad en la opinión pública: ¿dónde están los límites del humor? La pregunta no tiene una respuesta única, porque depende de factores como el contexto, la intención, el poder del emisor y la vulnerabilidad del grupo al que se alude. Una broma entre amigos no opera con la misma lógica que una broma lanzada al aire ante millones de personas.

Los estudios sobre comunicación y humor coinciden en que el chiste construye complicidad entre quien lo emite y quien lo recibe, pero también puede excluir a un tercero. Cuando el objeto de la broma es una característica identitaria históricamente estigmatizada, el efecto cambia: lo que para unos es ingenio, para otros puede ser un recordatorio de burlas sufridas. Esa asimetría explica por qué el mismo comentario puede generar risas en un aula y desconcierto en un set de televisión.

En el caso de las figuras públicas con audiencias enormes, la responsabilidad se amplifica. Un creador de contenido no solo comunica: también modela conductas para millones de personas, muchas de ellas jóvenes. Lo que dice en un medio tradicional adquiere un peso adicional porque circula fuera de su ecosistema habitual y llega a públicos que no lo conocen.

La identidad de género como tema de conversación pública

El segundo eje del debate tiene que ver con la sensibilidad en torno a la identidad de género. En los últimos años, la conversación pública ha incorporado con fuerza los derechos de las personas trans y no binarias, así como el reconocimiento de que la identidad no siempre coincide con el sexo asignado al nacer. Ese avance ha venido acompañado de un mayor cuidado en el lenguaje y de una revisión de los chistes que durante décadas se usaron como recurso fácil.

Para muchas personas, bromea con haber nacido mujer no es una travesura inocua, sino una forma de trivializar una experiencia vital que enfrenta discriminación estructural en ámbitos laborales, educativos y familiares. Para otras, se trata simplemente de un absurdo sin consecuencias. La distancia entre ambas lecturas es enorme y no se resuelve con una aclaración al aire.

La reacción de la conductora, al señalar que el tema era sensible, funcionó como una suerte de recordatorio público de esa distancia. No calificó la broma de ofensiva ni exigió disculpas: simplemente explicó por qué su rostro no había acompañado la risa. En televisión en vivo, esa clase de gesto adquiere una dimensión pedagógica que trasciende la anécdota.

El alcance mediático del episodio

El clip se replicó en portales de noticias de espectáculos, en cuentas de memes y en programas de análisis de medios. La conversación se extendió durante días y mutó hacia temas más amplios: la relación entre creadores digitales y prensa tradicional, la preparación de los invitados para entrevistas en vivo y la forma en que las audiencias interpretan el sarcasmo cuando no conocen al emisor.

También surgió un interés renovado por la trayectoria del youtuber, cuya carrera se ha caracterizado por un crecimiento sostenido y por una capacidad notable de adaptación a distintos formatos. Ha incursionado en documentales de viaje, en series para plataformas de streaming, en libros y en proyectos empresariales, siempre con una marca personal reconocible. Esa versatilidad, sin embargo, no lo exime de las reglas no escritas que rigen la conversación en medios masivos.

El episodio deja además una constatación sobre el funcionamiento de la esfera pública contemporánea: un instante de diez segundos puede convertirse en el tema central de una semana, generar miles de publicaciones y abrir discusiones que exceden por mucho el hecho original. La velocidad de circulación es tal que la aclaración del protagonista llega casi al mismo tiempo que la broma, y las audiencias reciben ambos elementos mezclados, sin orden cronológico claro.

En ese escenario, la reacción de la conductora se convirtió en el punto de referencia para buena parte de la conversación. Su decisión de no reír y de explicar sus razones fue leída por unos como una lección de prudencia y por otros como una muestra de rigidez. Lo que resulta innegable es que el silencio de esos segundos dijo más que muchas declaraciones posteriores, y que la broma improvisada dejó de pertenecer a quien la pronunció para instalarse, definitivamente, en el terreno de la interpretación colectiva.

Source: xeu.mx News

At the center of the allegation is distillation, a technique in which a less capable AI model is trained on the outputs of a more advanced system. If successful, distillation can allow a rival to replicate some capabilities of a leading model at lower cost and in less time than building a frontier system from scratch. Anthropic said the alleged campaign was conducted by operators affiliated with Alibaba and Alibaba Qwen, the company's AI lab. Alibaba did not immediately respond to a request for comment.

What Anthropic alleges

According to the letter, the alleged operation relied on a large network of fake accounts rather than a single compromised credential or a conventional cyber intrusion. The accounts were used to query Claude repeatedly over roughly six weeks, producing tens of millions of exchanges. That volume matters because distillation depends on scale: the more outputs an attacker can collect, the more training signal is available to shape a competing model.

Anthropic framed the activity as model extraction, a category of abuse that sits somewhere between intellectual property theft and a security breach. The alleged campaign did not involve stealing model weights directly. Instead, it sought to capture behavior - the patterns, reasoning traces, and response styles that make a model useful - through the application programming interface that ordinary customers use.

The company said the alleged campaign could help accelerate China's ability to approach the capabilities of its advanced Mythos Preview model. The allegation lands in an already tense technology environment. AI development has become closely tied to US-China competition, and US officials have stepped up scrutiny of advanced AI systems over concerns that they could be used by military or intelligence users in countries of concern.

The timing also reflects a broader shift in the enterprise AI market. Businesses are adopting generative AI tools across functions including software development, customer support, research, and back-office operations. That adoption puts pressure on vendors to show they can detect misuse while keeping services available to corporate customers. A large-scale extraction campaign, if confirmed, would raise questions about whether those two goals can be balanced.

Prior campaigns and an escalating pattern

Anthropic said the Alibaba allegation is not an isolated data point. In February, the company said it had identified similar campaigns by DeepSeek, Moonshot AI, and MiniMax to extract capabilities from Claude. The alleged activity ranged from more than 150,000 exchanges by DeepSeek to more than 13 million by MiniMax. Viewed together, the incidents suggest that model extraction is becoming more industrial in scale rather than remaining a niche risk.

Analysts said the pattern points to a wider problem for the AI industry. A frontier model is expensive to train, but its outputs can be gathered through ordinary product interfaces. Attackers can spread queries across many accounts to stay below rate limits, rotate access patterns, and make automated traffic look like normal usage. Defenders must then separate legitimate customer activity from coordinated extraction without degrading service for paying enterprises.

A new supply chain risk

If Anthropic's claims are true, the alleged campaign could allow Alibaba to build a comparable model in a short period of time and offer it at a much lower cost, said Anand Joshi, an AI analyst at TechInsights. That prospect turns model extraction into an economic threat as well as a security one. A competitor would not need to repeat the full research and compute burden of training a frontier system if it could learn enough from a rival's outputs.

Analysts said the alleged campaign also points to a broader pattern beyond the two companies. Viewed alongside previous incidents cited by Anthropic, they said, model extraction appears to be escalating rather than remaining an isolated risk.

Sanchit Vir Gogia, chief analyst at Greyhound Research, said the enterprise supply chain no longer ends at software, APIs, and cloud regions. It now includes rented intelligence, and rented intelligence can be copied and redeployed well outside the safety controls it was born with. Gogia said distillation should be a board-level concern because a weaker model trained on a stronger one can inherit its capabilities without the governance and controls around the original system.

For enterprises, the allegations point to a potentially more serious risk than conventional intellectual property theft: reverse engineering at scale. If proven, they would suggest that AI models can be copied systematically, turning model extraction into a new AI supply-chain risk.

Pareekh Jain, CEO of Pareekh Consulting, said a rival that can clone the exact brain of the AI a company relies on could find its blind spots, hack automated systems, or cause the AI vendor to panic and shut down services that a business needs to run every day. That scenario moves the risk from legal departments into operational resilience. An enterprise that depends on a single model provider may face disruption not only from outages or price changes but also from a vendor's response to abuse.

Mitigating the risks

The allegation raises questions about the controls AI vendors have in place and how customers can protect themselves. Jain said vendors should provide verified accounts, smart rate limits, abuse detection, usage monitoring, contractual bans on distillation, incident disclosure, and audit rights. Enterprises should ask how the vendor detects and blocks large-scale model extraction and can demand contracts that guarantee backup plans and financial refunds if the AI service gets attacked or suddenly shut down.

Joshi said enterprise customers should also press vendors for greater transparency around model development and safeguards. Enterprise buyers should ask what training data was used, how it was trained, what guardrails exist, and how they can audit it. Model publishers will have to come up with watermarking technology in models as well as model responses, so if model skills are stolen, they should be able to find the thief.

The alleged Alibaba campaign also highlights the difficulty of proving extraction after the fact. Outputs from a model can be paraphrased, filtered, and mixed with other data before they are used in training. That makes attribution technically challenging and raises the bar for vendors that want to pursue enforcement. Detection may depend on correlating account behavior, query patterns, payment signals, and infrastructure fingerprints long before a competing model appears.

For now, the claims remain allegations. Alibaba did not immediately respond to a request for comment, and the letter does not establish a public finding of wrongdoing. But the dispute has already put a spotlight on the controls that surround commercial AI services and on the willingness of enterprises to depend on models that can be probed at scale. The biggest question may be less about one campaign than about whether the industry can make model extraction expensive, detectable, and contractually risky enough to deter it.

Source: InfoWorld News

The pitch is straightforward. AI agents can remain active for hours or days, but they often spend much of that time doing nothing. Traditional cloud infrastructure bills for compute while the agent is waiting. Enterprises can shut down the underlying resources to save money, but restarting an agent usually means rebuilding the execution environment and restoring context, which adds latency and engineering work. Google says Agent Compute Engine addresses that trade-off by preserving agent state in hardware-isolated microVMs that can be suspended and resumed on demand.

Key facts

• Google launched Agent Compute Engine to reduce idle infrastructure costs for AI agents.
• The service suspends inactive agents and restores them on demand while preserving execution state.
• It uses hardware-isolated microVMs, which Google says boot faster and require fewer resources to initialize than conventional virtual machines.
• MicroVMs run their own kernels, giving stronger isolation than containers that share the host operating system kernel.
• Google argues the stronger isolation boundary helps limit the impact of compromised agents executing AI-generated code or handling untrusted inputs.
• Analysts see savings for long-running coding and research agents, but also warn about vendor lock-in, operational complexity, storage costs, checkpointing risks, and uncertain savings for continuously active workloads.

The choice of microVMs is central to Google’s argument. Conventional virtual machines typically take longer to boot and require more resources to initialize, which can make frequent suspend-and-resume cycles inefficient. Containers are lighter, but they share the host operating system’s kernel, creating a weaker isolation boundary. For agents that execute AI-generated code or process untrusted inputs, that difference matters. A compromised agent running in a container could potentially affect other workloads on the same host. A microVM, by contrast, runs its own kernel, giving each agent environment a stronger separation from neighbors.

That isolation argument is likely to resonate with CIOs and security teams as agents move from experimental pilots into production fleets. The more autonomy an agent has, the more likely it is to call external tools, query databases, send emails, initiate payments, or modify code. Each of those actions expands the attack surface. Google says the microVM design limits the potential impact of a compromised agent, which could make the service more attractive for workloads that combine stateful execution with sensitive tools and data.

Why idle time matters

Idle time is an unusual cost problem because it is easy to overlook at small scale. A single agent that waits a few minutes between tasks may not consume enough compute to justify close monitoring. But enterprises are increasingly deploying agents in fleets. A customer support agent may wait for a human approval. A coding agent may wait for a test suite to finish. A research agent may pause between web searches or document analysis steps. An operations agent may wait for an alert or a scheduled event. When thousands of such agents are running, the cumulative cost of idle compute can become significant.

Traditional cloud compute was not designed around that pattern. It is optimized for workloads that are either active or shut down, not for stateful processes that activate thousands of times and otherwise sit idle. Shutting down a conventional instance saves money, but the restart process can be slow and brittle. The agent may need its filesystem, memory, dependency versions, and conversational context restored before it can continue. Rebuilding that environment from scratch can add seconds or minutes of latency, and it can introduce errors if the reconstructed state does not exactly match the original.

Google’s approach is intended to remove that reconstruction step. By preserving the agent’s execution state in a suspended microVM, the service aims to let the agent resume from where it left off. For developers, that could mean less custom code for saving and restoring state. Instead of building a save-and-resume layer into every agent, teams could rely on the platform to hold the environment while the agent is inactive.

Where savings are likely to appear

The clearest beneficiaries are likely to be long-running agents that hold state for hours or days. Coding agents that maintain a workspace across multiple steps, research agents that accumulate findings over time, and workflow agents that wait for approvals or external events are natural candidates. In those cases, the cost of keeping compute resources alive during idle periods can add up, while the cost of rebuilding context after a shutdown can be high.

Industry analysts broadly agree that the economics can work for those workloads. They also caution that savings will not be uniform. Agents that are continuously active or compute-intensive will see smaller gains because they spend little time idle. The service may do little for workloads that are already optimized around short-lived tasks or that do not need to preserve complex state. CIOs evaluating the service should first measure how many agents hold state across more than a few minutes of idle time. That number is a better guide to potential savings than the total number of agents in a portfolio.

The developer productivity argument is separate from the infrastructure bill. Maintaining long-running tasks often requires engineering work that exists only to survive pauses. Teams may build checkpointing logic, state stores, retry mechanisms, and custom orchestration. If a platform can preserve execution state and resume an agent on demand, some of that code may no longer be necessary. That could reduce operational overhead and make it easier to build agents that run for extended periods without constant supervision.

Trade-offs and risks

The service also comes with trade-offs. The most frequently cited is vendor lock-in. Agent Compute Engine is closely integrated with Google Cloud, which could make it difficult for enterprises to move agent workloads to another cloud provider or run them across multiple platforms without reworking parts of their infrastructure. Enterprises could also become dependent on Google’s approach to managing agent state, identity, and execution environments. That dependency matters because agent platforms are still evolving, and standards for state portability and cross-cloud orchestration remain immature.

Cost predictability is another concern. Suspending idle agents can reduce compute charges, but it may increase storage costs. Each suspended microVM needs disk space to retain its working state. At small scale, that storage may be negligible. At scale, storing thousands of persistent agent environments could add significantly to infrastructure bills. Finance teams may find it harder to predict net savings when compute costs fall but storage and management costs rise. The service’s maturity is also a factor. Because it is not yet in preview, enterprises would be betting on Google’s roadmap rather than proven production performance.

Checkpointing introduces its own technical challenges. The idea of suspending an agent and resuming it later is not entirely new. Competing cloud platforms have introduced mechanisms that stop billing CPU while an agent waits or suspend hosted agents between turns while keeping the filesystem. Google is claiming a step beyond those approaches by preserving memory and execution state so an agent can resume from where it left off rather than reconstructing its context and restarting execution.

That capability is powerful, but it can be difficult to implement safely when agents interact with external systems. If an agent is paused between an API call and its confirmation, it may wake up unsure whether a payment went through or a code commit landed. The agent may retry the action, creating duplicate side effects. Every external action must be safe to repeat, a property known as idempotency. Many agent codebases today are not written that way, which means developers may need to redesign tool calls, add confirmation records, or implement reconciliation logic before they can rely on pause-and-resume behavior.

Drift, debugging, and governance

Long-lived agent environments can also drift from their original configurations. Over weeks, files change, dependencies are updated, and software versions diverge. An environment that was once reproducible may become difficult to audit or rebuild. Bugs that appear after a resume may be harder to reproduce because the surrounding state has evolved. Developers may need expiry policies that force old workspaces to be rebuilt from fresh base images, along with stronger version control and environment documentation.

Debugging suspended environments adds another layer of complexity. Developers may need to inspect a paused virtual machine to diagnose a problem that emerged after an agent resumed. Tooling for debugging paused environments is still relatively immature, which could slow incident response and make root-cause analysis more difficult. Security teams will also want to understand how identity, secrets, and permissions are preserved across suspend-and-resume cycles. An agent that wakes up with stale credentials or an outdated policy could create compliance risk.

For enterprises, the decision will come down to workload economics and operational readiness. Agent Compute Engine targets a real cost problem, and its microVM isolation model offers security advantages over shared-kernel containers. But the savings depend on idle patterns, storage overhead, and the cost of adapting agent code to checkpointing. The service also ties workloads more closely to one cloud platform. CIOs will need to weigh those trade-offs against the value of preserving state, reducing restart latency, and simplifying long-running agent operations.

The broader market signal is that cloud providers are now competing on the operational details of agent infrastructure, not just model access. As agents become more numerous and more autonomous, idle-time management, state preservation, isolation, and debugging will become part of the platform decision. Google’s new service pushes those issues into the foreground, while the tooling for debugging paused environments remains relatively immature and the cost model still depends heavily on how each enterprise designs and runs its agent fleets.

Source: InfoWorld News

Die niederländische Make-up-Künstlerin Nikkie de Jager, Millionen Menschen besser bekannt als Nikkie Tutorials, hat im Januar 2020 öffentlich gemacht, dass sie transgender ist. In einem emotionalen Video mit dem Titel „I am coming out“ sprach die damals 25-Jährige über ihre Identität, ihre Kindheit, ihre Transition und den Grund, warum sie sich ausgerechnet zu diesem Zeitpunkt an die Öffentlichkeit wandte. Die Influencerin hatte sich mit Make-up-Tutorials eine der erfolgreichsten Beauty-Präsenzen im Netz aufgebaut und zählte zu diesem Zeitpunkt 12,7 Millionen Abonnenten. Ihr Coming-out war kein PR-Manöver, sondern die Reaktion auf eine Erpressung: Mehrere Personen drohten damit, ihre Geschichte gegen ihren Willen zu veröffentlichen. Nikkie de Jager entschied sich, selbst die Kontrolle über ihre Erzählung zu übernehmen.

Die wichtigsten Fakten im Überblick

• Nikkie de Jager ist als Nikkie Tutorials bekannt und arbeitet als Make-up-Künstlerin und Influencerin.
• Zum Zeitpunkt ihres Coming-outs war sie 25 Jahre alt und hatte 12,7 Millionen Abonnenten.
• In dem Video „I am coming out“ erklärte sie, im falschen Körper geboren worden zu sein und transgender zu sein.
• Sie habe bereits früh gemerkt, dass sie ein Mädchen ist; ihre Mutter habe sie unterstützt.
• In der Schule wurde sie nach eigenen Angaben von manchen Kindern gemobbt.
• Mit 14 Jahren begann sie eine Hormontherapie, mit 19 folgte eine geschlechtsangleichende Operation.
• Zu diesem Zeitpunkt produzierte sie bereits Videos und baute ihre Karriere als Influencerin auf.
• Sie hatte bereits mit Kim Kardashian und Lady Gaga zusammengearbeitet.
• Mehrere Menschen erpressten sie und drohten, ihre Geschichte zu veröffentlichen.
• Ihr Verlobter Dylan wusste nicht von Anfang an Bescheid, unterstützt sie aber.
• Zahlreiche Prominente und YouTuber reagierten auf ihr Coming-out, darunter Ariana Grande und Marc Jacobs.

Ein emotionales Video und eine klare Botschaft

Das Video, in dem sich Nikkie Tutorials als transgender outete, dauert knapp 18 Minuten und trägt einen ebenso schlichten wie eindeutigen Titel: „I am coming out“. Darin sagt sie, es sei ein neues Jahr, 2020, und sie wolle dieses Jahr mit der Wahrheit beginnen. Es sei an der Zeit, loszulassen und endlich wirklich frei zu sein. Sie wurde nach eigenen Worten im falschen Körper geboren, was bedeute, dass sie transgender sei. Diese Sätze richtete sie direkt an ihre Millionen Abonnenten, ohne Umwege und ohne juristische Vorbereitung. Der Ton des Videos war persönlich, verletzlich und zugleich entschlossen. Nikkie de Jager sprach nicht über eine abstrakte Debatte, sondern über ihr eigenes Leben. Sie erklärte, warum sie sich lange nicht wohlgefühlt habe, warum sie bestimmte Erwartungen an sich nicht erfüllen konnte und warum sie nun den Schritt in die Öffentlichkeit wage. Für viele Zuschauer war es das erste Mal, dass sie die Make-up-Künstlerin nicht nur als Expertin für Schönheit erlebten, sondern als Menschen, der einen langen und schwierigen Weg hinter sich hatte.

Kindheit, Familie und das Gefühl, im falschen Körper zu leben

Nikkie de Jager wurde mit dem bei der Geburt als männlich zugewiesenen Geschlecht geboren. Schon sehr früh habe sie jedoch gemerkt, dass sie ein Mädchen ist. In dem Video beschreibt sie, dass sie nicht verstehen konnte, warum sie kurze Haare tragen musste, Hosen und T-Shirts statt Kleider. Sie spielte mit Puppen, lackierte sich die Nägel und experimentierte mit falschen Haaren. Alles an ihr sei mädchenhaft gewesen, sagte sie. Ihre Mutter habe früh bemerkt, dass ihre Tochter sich in ihrem Körper nicht wohlfühlte. Rückblickend erklärte de Jager, ihre Mutter habe sofort gewusst, dass sie entweder schwul sein würde oder eben etwas anderes. Und es sei tatsächlich etwas anderes gewesen. Dass ihre Mutter sie nie verurteilt habe, habe ihr sehr geholfen. Sie betonte, genau das brauche ein Kind: Eltern, die verstehen und zuhören. Diese Unterstützung war offenbar ein wichtiger Grund dafür, dass sie ihren Weg vergleichsweise früh gehen konnte. Mit sechs Jahren ließ sie sich die Haare länger wachsen. Die Lehrer in ihrer Schule unterstützten sie, dennoch wurde sie von manchen Kindern gemobbt. Diese Erfahrung zeigt, wie stark gesellschaftliche Erwartungen und Rollenbilder bereits im Kindesalter wirken können.

Hormontherapie, Operation und eine bemerkenswerte Karriere

Mit 14 Jahren begann Nikkie de Jager eine Hormontherapie, die ihr schnelles Wachstum stoppte. Mit 19 Jahren folgte eine geschlechtsangleichende Operation. Zu diesem Zeitpunkt produzierte sie bereits Videos und hatte begonnen, sich im Netz eine Fangemeinde aufzubauen. Die medizinischen Schritte waren für sie offenbar Teil eines längeren Prozesses, den sie nicht öffentlich dokumentierte, sondern zunächst für sich behielt. Parallel entwickelte sie sich zu einer der bekanntesten Make-up-Künstlerinnen im Netz. Ihre Tutorials zeichneten sich durch hohe Kunstfertigkeit, Humor und eine große Nähe zu ihren Zuschauern aus. Sie zeigte nicht nur, wie man Produkte aufträgt, sondern schuf aufwendige Looks, erklärte Techniken und nahm ihr Publikum mit in ihren Alltag. Im Laufe der Jahre arbeitete sie mit internationalen Stars zusammen, darunter Kim Kardashian und Lady Gaga. Diese Kooperationen machten deutlich, wie weit ihre Reichweite über die Beauty-Nische hinausreichte. Dennoch blieb ein Teil ihrer Biografie jahrelang privat. Das Coming-out im Januar 2020 veränderte diese Trennung zwischen öffentlicher Karriere und persönlicher Geschichte.

Warum das Coming-out erst jetzt kam

Die erfolgreiche Unternehmerin habe immer gewusst, dass sie ihre Geschichte irgendwann erzählen würde, berichtete sie in dem Video. Dass der Zeitpunkt nun gekommen war, hatte jedoch einen unschönen Grund. Mehrere Menschen erpressten sie und drohten damit, ihre Geschichte zu veröffentlichen. Wer genau hinter der Erpressung steckte, verriet sie nicht. Die Erpresser argumentierten nach ihren Angaben, sie wolle es veröffentlichen, weil Nikkie de Jager angeblich lüge oder nicht die Wahrheit sage. Sie habe zu viel Angst davor gehabt, den Leuten zu erzählen, wer sie wirklich sei. Diese Behauptung wies sie entschieden zurück. In dem Video betonte sie, sie habe keine Angst. Sie wolle selbst bestimmen, wann, wie und in welchem Rahmen sie über ihre Identität spricht. Das ist ein zentraler Unterschied zwischen einem selbstbestimmten Coming-out und einem fremdbestimmten Outing. Ein Coming-out kann befreiend sein, wenn die betroffene Person den Zeitpunkt wählt. Ein Outing gegen den Willen der Person ist dagegen ein Eingriff in die Privatsphäre und kann erhebliche psychische und soziale Folgen haben. Im Fall von Nikkie de Jager wurde aus einer drohenden Enthüllung eine bewusste Entscheidung: Sie nahm den Erpressern die Möglichkeit, ihre Geschichte zu kontrollieren, indem sie sie selbst erzählte.

Erpressung, Privatsphäre und die Verantwortung der Öffentlichkeit

Der Fall macht sichtbar, wie verletzlich öffentliche Personen trotz großer Reichweite sein können. Nikkie de Jager hatte jahrelang über Make-up und Schönheit gesprochen, aber nicht über ihre Geschlechtsidentität. Diese Entscheidung war ihr gutes Recht. Die Erpressung griff genau diese Grenze an. Wer damit droht, private Informationen zu veröffentlichen, nutzt die Angst vor Ablehnung, vor beruflichen Konsequenzen und vor öffentlicher Debatte aus. Dabei geht es nicht um Transparenz oder um ein berechtigtes öffentliches Interesse, sondern um Kontrolle und Druck. Die Geschichte zeigt auch, wie wichtig ein respektvoller Umgang mit dem Thema Transgender ist. Transgender ist ein Oberbegriff für Menschen, deren Geschlechtsidentität nicht mit dem Geschlecht übereinstimmt, das ihnen bei der Geburt zugewiesen wurde. Die Wege, die Menschen dabei gehen, sind sehr unterschiedlich. Manche unterziehen sich medizinischen Behandlungen, andere nicht. Manche sprechen früh über ihre Identität, andere später oder nie öffentlich. Nikkie de Jager entschied sich für ein öffentliches Coming-out, weil sie dazu gedrängt wurde. Sie machte aber deutlich, dass die Entscheidung, über die eigene Identität zu sprechen, bei der betroffenen Person liegen sollte. Für Medien und Öffentlichkeit bedeutet das, nicht über Menschen hinweg zu berichten und private Informationen nicht als Sensation zu behandeln.

Reaktionen aus der Promi-Welt und die Bedeutung für Trans-Sichtbarkeit

Nikkie de Jagers Coming-out löste große Aufmerksamkeit aus. Zahlreiche andere YouTuber und Prominente kommentierten ihren Instagram-Beitrag, darunter Ariana Grande und der Designer Marc Jacobs. Die Reaktionen waren überwiegend unterstützend. Für viele Fans war es bemerkenswert, dass eine so bekannte Influencerin offen über ihre Geschichte sprach. Sichtbarkeit kann für transgender Menschen wichtig sein, weil sie zeigt, dass sie in allen Bereichen der Gesellschaft vertreten sind, auch in der Unterhaltungsbranche. Zugleich ist Sichtbarkeit kein Ersatz für Schutz. Wer öffentlich über seine Identität spricht, wird oft mit Kommentaren, Fragen und teils auch Anfeindungen konfrontiert. Nikkie de Jager hatte bereits vor ihrem Coming-out Erfahrungen mit Mobbing gemacht. Ihr Fall zeigt, dass Unterstützung im persönlichen Umfeld eine große Rolle spielt. Ihre Mutter habe sie früh verstanden, und ihr Verlobter Dylan stehe heute hinter ihr. Dylan habe zunächst nicht von Anfang an über ihre Geschichte Bescheid gewusst, unterstütze sie aber. Diese Details machten das Video für viele Zuschauer nahbar. Es ging nicht um eine perfekte Inszenierung, sondern um einen Menschen, der nach Jahren des öffentlichen Lebens einen privaten Teil seiner Biografie teilt. Dass sie dazu gezwungen wurde, macht die Geschichte traurig. Dass sie sich dennoch selbst äußerte, verleiht ihr eine besondere Klarheit. In dem Video sagte sie, sie habe keine Angst davor, die Wahrheit zu erzählen.

Source: stern.de News

• Disney und der schwedische YouTuber Felix Kjellberg, bekannt als PewDiePie, gehen getrennte Wege.
• Grund sind antisemitische Videos, die der 27-Jährige veröffentlicht hatte.
• Sein Multi-Channel-Netzwerk Revelmode zählt mehr als 53 Millionen Abonnenten und arbeitet mit Maker Studios zusammen.
• Maker Studios wurde 2014 von Disney gekauft und beendet nun die Zusammenarbeit mit Kjellberg.
• Seit August 2016 soll der Schwede neun Videos produziert haben, in denen antisemitische Witze oder Darstellungen des Nationalsozialismus vorkamen.
• Besonders ein Video vom 11. Januar sorgte für Empörung: Zwei Inder wurden bezahlt, ein Pappschild mit der Aufschrift „Death to all Jews“ in die Kamera zu halten.
• Kjellberg erklärte, er habe mit dem Video lediglich zeigen wollen, wie „verrückt“ die moderne Welt sei, und unterstütze keine hasserfüllten Einstellungen.
• Das umstrittene Video und zwei weitere Januar-Clips wurden gelöscht; andere Clips mit antisemitischen Witzen sind weiterhin online.

Ein Bruch mit Folgen

Der Medienkonzern Disney hat die Zusammenarbeit mit einem der bekanntesten YouTube-Stars der Welt beendet. Felix Kjellberg, der unter dem Namen PewDiePie ein Millionenpublikum erreicht, ist nicht länger Teil des Netzwerks, das über Maker Studios mit Disney verbunden war. Der Schritt ist bemerkenswert, weil er nicht aus einem Vertragsstreit über Inhalte oder Reichweiten entstand, sondern aus einer politischen und moralischen Grenzüberschreitung. Kjellberg hatte wiederholt judenfeindliche Clips veröffentlicht. Damit geriet nicht nur der YouTuber selbst unter Druck, sondern auch die Unternehmen, die an seinem Erfolg verdienen.

Die Trennung zeigt, wie eng Unterhaltung, Werbung und politische Verantwortung in der digitalen Öffentlichkeit inzwischen verknüpft sind. Ein YouTube-Star ist kein isolierter Einzelunternehmer mehr. Hinter ihm stehen Netzwerke, Agenturen, Sponsoren und Plattformen, die an Reichweite und Aufmerksamkeit mitverdienen. Wenn Inhalte als antisemitisch kritisiert werden, geraten diese Strukturen ebenfalls in den Blick. Disney und Maker Studios mussten deshalb nicht nur entscheiden, ob Kjellbergs Witze geschmacklos waren. Sie mussten auch abwägen, welches Signal von einer weiteren Zusammenarbeit ausgehen würde.

Neun Videos seit August 2016

Es war nicht das erste Mal, dass Kjellberg wegen judenfeindlicher Clips in die Kritik geriet. Berichten zufolge hat der Schwede seit August 2016 neun Videos veröffentlicht, in denen antisemitische Witze oder Darstellungen des Nationalsozialismus vorkamen. Damit entstand der Eindruck einer Reihe von Grenzüberschreitungen, nicht eines einzelnen Ausrutschers. Für ein Publikum, das den YouTuber vor allem für Gaming, Comedy und schnelle Schnitte kennt, war diese Entwicklung nicht nur ein Imageproblem. Sie berührte die Frage, wo Provokation endet und wo Hass beginnt.

Die Kritik richtete sich zunächst an Kjellberg selbst. Sie traf aber auch die Unternehmen, die seine Kanäle vermarkteten. Wer mit einem Star wirbt, übernimmt Mitverantwortung für dessen öffentliches Auftreten. Das gilt besonders dann, wenn Inhalte nicht in einem Nischenforum erscheinen, sondern vor einem Millionenpublikum. Kjellberg war kein Randphänomen. Sein Netzwerk Revelmode zählte mehr als 53 Millionen Abonnenten. Diese Reichweite machte ihn für Werbetreibende attraktiv und für seine Partner wertvoll. Zugleich erhöhte sie die Wirkung jedes einzelnen Videos.

Das Video mit dem Pappschild

Besonders ein Video vom 11. Januar rief Empörung hervor. Darin ist zu sehen, wie Kjellberg weltweit über eine Website Menschen beauftragt, gegen Bezahlung Dinge für ihn zu tun. Zwei Inder bezahlt er dafür, ein Pappschild zu basteln, auf dem „Death to all Jews“ steht, und es in die Kamera zu halten. Das Video wurde millionenfach angeschaut. Die Aufnahme war kein abstrakter Scherz über Geschichte. Sie nutzte eine NS-Vernichtungsdrohung als Pointe und inszenierte sie als bezahlte Dienstleistung.

Gerade darin lag die besondere Schärfe des Falls. Antisemitismus wird nicht dadurch harmlos, dass er in einem Unterhaltungsformat auftaucht. Die Verwendung eines Tötungsslogans gegen Juden ist auch dann nicht einfach Ironie, wenn sie als Kommentar über die Absurditäten der digitalen Welt gemeint sein soll. Die Bilder waren für ein internationales Publikum sichtbar, darunter auch junge Zuschauer, die mit historischen Zusammenhängen möglicherweise nicht vertraut sind. Wer solche Inhalte verbreitet, muss damit rechnen, dass sie nicht nur als Provokation, sondern als Botschaft verstanden werden.

Reaktionen und Statements

Nach Berichten schwedischer Medien kündigte Maker Studios an, die Zusammenarbeit mit dem YouTuber zu beenden. Er habe „seine Witze zu weit getrieben“, hieß es in einer Stellungnahme. Der Satz markiert den Kern des Konflikts. Kjellbergs Marke beruhte seit Jahren auf Grenzüberschreitung, auf schneller Sprache, auf bewusst albernen Situationen und auf der Rolle des freundlichen Provokateurs. Doch nicht jede Provokation lässt sich mit dem Hinweis auf Comedy erklären. Sobald ein Witz eine historisch aufgeladene Vernichtungsdrohung gegen eine Minderheit verwendet, verändert sich die Bedeutung.

Aron Verständig, Präsident des Zentralrats der Juden in Schweden, wollte sich zunächst nicht ausführlich äußern. Er sagte lediglich, er halte das Vorgehen für unangebracht, wolle sich diesmal aber eines Kommentars enthalten. Diese Zurückhaltung zeigt, wie schwierig die Einordnung für jüdische Gemeinschaften ist. Einerseits geht es um eine klare rote Linie. Andererseits steht die Frage im Raum, ob jedes provokative Video dieselbe öffentliche Aufmerksamkeit erhalten sollte. Der Fall war bereits zu einem internationalen Medienthema geworden, bevor sich viele Beteiligte positioniert hatten.

In den sozialen Netzwerken wurde Kjellberg vor allem vorgeworfen, sich bislang nicht entschuldigt zu haben. Am Sonntag verteidigte er sich in einem schriftlichen Statement gegen die Vorwürfe, antisemitische Inhalte verbreitet zu haben. Die Kritik beziehe sich auf eines seiner Videos, mit dem er lediglich habe zeigen wollen, wie „verrückt“ die moderne Welt sei. In keiner Weise unterstütze er „irgendwelche hasserfüllten Einstellungen“. Diese Erklärung ist ein typischer Versuch, die Verantwortung auf die Rezeption zu verschieben. Nicht der Inhalt, sondern die Interpretation sollte demnach das Problem sein.

Diese Verteidigung griff jedoch zu kurz. Wer eine Website nutzt, um Menschen für absurde Aufgaben zu bezahlen, und dabei ein Schild mit einer Todesdrohung gegen Juden produzieren lässt, kann sich nicht allein auf die Absicht berufen. Die Wirkung solcher Bilder ist Teil der Tat. Zuschauer sehen nicht den inneren Gedanken des Comedians. Sie sehen den Slogan, die Kamera, die Bezahlung und die Bereitschaft der Beteiligten. Das Video und zwei weitere vom Januar wurden am Wochenende gelöscht. Andere Clips mit antisemitischen Witzen sind hingegen weiter online.

Plattform, Reichweite und Verantwortung

Der Fall macht ein grundlegendes Problem der digitalen Unterhaltungsindustrie sichtbar. YouTube-Stars agieren oft wie kleine Medienunternehmen, sind aber rechtlich und redaktionell nicht immer so organisiert. Sie produzieren täglich, reagieren auf Trends und leben von Aufmerksamkeit. Multi-Channel-Netzwerke bündeln viele Kanäle, vermarkten Werbeplätze und bieten technische Unterstützung. Für Disney war Maker Studios einst ein strategischer Kauf, um am wachsenden Onlinevideomarkt teilzuhaben. Dass nun eine Zusammenarbeit wegen antisemitischer Inhalte beendet wird, zeigt, wie schnell solche Strategien in einen Reputationskonflikt geraten können.

Plattformen und Netzwerke stehen vor einer schwierigen Aufgabe. Sie sollen Meinungsfreiheit und künstlerische Freiheit nicht unnötig einschränken. Gleichzeitig müssen sie verhindern, dass ihre Infrastruktur für Hasspropaganda genutzt wird. Die Entscheidung, einzelne Videos zu löschen, löst das Problem nur teilweise. Andere Clips bleiben online. Das verweist auf die Fragmentierung der Öffentlichkeit: Ein gelöschtes Video verschwindet nicht aus dem Gedächtnis, und seine Inhalte können über Kopien, Reaktionen und Diskussionen weiter zirkulieren. Gerade deshalb ist die Haltung der beteiligten Unternehmen so wichtig. Sie setzt ein Zeichen, auch wenn sie die Verbreitung nicht vollständig kontrollieren kann.

Der schmale Grat zwischen Provokation und Hass

Kjellbergs Karriere war von Beginn an mit dem Versprechen verbunden, dass Unterhaltung unmittelbar und ungefiltert sein darf. Viele seiner Fans schätzten genau diese Mischung aus Nähe, Chaos und Selbstironie. Doch die Grenze zwischen einem harmlosen Scherz und einer verletzenden Botschaft verläuft nicht dort, wo der Urheber sie zieht. Sie verläuft dort, wo historische Gewalt, Gruppenfeindschaft und Diskriminierung berührt werden. Antisemitische Witze sind kein Stilmittel unter vielen. Sie knüpfen an eine Geschichte der Verfolgung an, die nicht als Kulisse für schnelle Pointen dienen darf.

Der Fall PewDiePie ist deshalb mehr als eine Personalie. Er betrifft die Frage, wie viel Provokation eine Gesellschaft aushält, ohne dass sie in Normalisierung umschlägt. Er betrifft die Verantwortung von Unternehmen, die mit Reichweite Geld verdienen. Und er betrifft die Rolle von Plattformen, die zwischen offener Teilhabe und Schutz vor Hass abwägen müssen. Die Trennung von Disney ist ein deutliches Signal. Sie ersetzt aber keine Debatte darüber, wie digitale Öffentlichkeit künftig mit solchen Grenzüberschreitungen umgeht. Während das umstrittene Januar-Video und zwei weitere Clips gelöscht wurden, sind andere Videos mit antisemitischen Witzen weiterhin abrufbar. Genau darin liegt die offene Spannung dieses Falls.

Source: Jüdische Allgemeine News

• Huda Kattan zählt zu den erfolgreichsten Beauty-Influencerinnen der Welt und hat mehr als 37 Millionen Follower.
• Ihre Kosmetiklinie Huda Beauty wird mit einem geschätzten Vermögen von rund 490 Millionen Euro in Verbindung gebracht.
• Kattan ist Tochter irakischer Einwanderer und wuchs unter anderem im US-Bundesstaat Tennessee auf.
• Bereits mit zwölf Jahren setzte sie Make-up bewusst ein, um ihr Selbstbewusstsein zu stärken.
• Im Jahr 2010 startete sie einen Beauty-Blog; später baute sie gemeinsam mit ihren Schwestern Mona und Alya ein Kosmetikimperium auf.
• Dubai bezeichnet sie als entscheidenden Standort für ihren unternehmerischen Erfolg.

Von Make-up-Videos zu einem globalen Kosmetikunternehmen

Huda Kattan gehört zu den bekanntesten Gesichtern der internationalen Beauty-Welt. Mit ihren Schminkvideos in sozialen Netzwerken und mehr als 37 Millionen Followern wurde sie zu einer der erfolgreichsten Influencerinnen überhaupt. Aus ihrer Leidenschaft für Make-up entstand die eigene Kosmetiklinie Huda Beauty, mit der die in Dubai lebende Unternehmerin ein geschätztes Vermögen von rund 490 Millionen Euro erreicht hat. Ihr Weg dorthin ist alles andere als geradlinig verlaufen und erzählt von familiären Brüchen, dem Gefühl des Fremdseins und einer sehr frühen Entdeckung: Make-up war für sie nicht nur Schönheitspflege, sondern ein Mittel, um sich selbst zu behaupten.

Kattan ist die Tochter irakischer Einwanderer. Ihr Vater war Professor, die Familie zog häufig um, weil er bessere berufliche Möglichkeiten suchte. Die Kindheit beschreibt sie rückblickend als wenig unbeschwert. Es sei oft ein kleiner Überlebenskampf gewesen, während der Vater studierte und die Familie versuchte, über die Runden zu kommen. Für Huda Kattan bedeutete das ständige Anpassung: neue Wohnorte, neue Schulen, neue Umgebungen. Diese Erfahrungen hätten sie geprägt und auf eine Weise mutig gemacht, die sie heute als Grundlage für ihre unternehmerische Härte betrachtet. Wer früh lernt, mit Unsicherheit umzugehen, entwickelt möglicherweise eine Widerstandskraft, die im Geschäftsleben zum Vorteil wird.

In ihrer Highschool in Tennessee wurde sie nach eigener Darstellung wegen ihres exotischen Aussehens ausgegrenzt. In dieser Zeit begann sie, sich intensiv mit Make-up zu beschäftigen. Getreu dem Credo „viel hilft viel“ legte sie schon mit zwölf Jahren mehrere Schichten auf und entwickelte einen Look, der ihr Halt gab. Die Schminke stärkte ihr Selbstbewusstsein. Was zunächst wie ein jugendliches Experiment wirkte, wurde später zur beruflichen Leidenschaft. Kattan erkannte früh, dass Schönheit nicht nur mit Äußerlichkeiten zu tun hat, sondern auch mit Selbstwahrnehmung, Kontrolle und der Möglichkeit, sich in einer Umgebung zu behaupten, in der man sich sonst nicht zugehörig fühlt.

Als junge Frau arbeitete Huda Kattan zunächst als Finanzanalystin in Dubai. Doch der Wunsch, sich beruflich mit Beauty zu beschäftigen, blieb. Sie begann eine Ausbildung zur Visagistin und lernte dort Techniken, die sie später in Video-Tutorials weitergab. Im Jahr 2010 startete sie ihren Blog hudabeauty.com, auf dem sie Schminktipps und Produktvorstellungen veröffentlichte. Der Blog wurde zu einem der meistgelesenen Beauty-Blogs im Internet. Besonders Frauen im Nahen Osten verehrten sie für ihren Erfolg und ihre Make-up-Philosophie. Kattan verband praktische Anleitungen mit einer persönlichen Geschichte: Sie zeigte nicht nur, wie man sich schminkt, sondern auch, wie man Selbstvertrauen aufbauen kann.

Gemeinsam mit ihren beiden Schwestern Mona und Alya baute Huda Kattan schließlich ein Kosmetikimperium auf. Das erste Produkt waren falsche Wimpern aus einer synthetischen Faser, die Nerz-Pelz imitieren sollte. Die Wimpern trafen offenbar einen Nerv: Die US-Prominente Kim Kardashian wurde Kundin, was dem jungen Unternehmen zusätzliche Aufmerksamkeit brachte. Später wurde das Sortiment erweitert, und die Wimpern waren auch in einer Luxus-Version aus echtem Nerz erhältlich. Aus einem Blog und ersten Produkten entstand eine Marke, die heute international präsent ist und zu den bekanntesten Namen der Beauty-Branche zählt.

Der wirtschaftliche Erfolg blieb nicht unbemerkt. Huda Kattan wurde auf einer viel beachteten Liste erfolgreicher Unternehmerinnen geführt, auf der auch Beyoncé und Oprah Winfrey stehen. Für Kattan war das eine besondere Ehre, denn sie bezeichnet die beiden als ihre Vorbilder. Zugleich betont sie, dass solche Anerkennung nicht täglich erfahrbar sei. Man arbeite ständig hart und bekomme nicht permanent Bestätigung. Der Platz auf einer solchen Liste sei deshalb nicht nur ein Symbol für Erfolg, sondern auch eine Erinnerung daran, dass sich Ausdauer langfristig auszahlen kann.

Der Einstieg in die Kosmetikindustrie war nach ihren Worten ein Fluch und ein Segen zugleich. Der Markt sei bereits sehr heiß gewesen, als sie mit ihrem Unternehmen begann. Das habe einerseits für viel Entwicklung und Dynamik gesorgt, andererseits aber auch für einen intensiven Wettbewerb. Alle hätten versucht, im Beauty-Geschäft mitzuspielen. Kattan unterscheidet jedoch zwischen denen, die nur am Geschäft teilhaben wollen, und denen, die wirklich Teil dieser Welt sind. Sie selbst beschreibt sich als besessen davon, Tipps und Tricks zu finden, um noch schöner auszusehen. Der Antrieb sei bis heute kindlich geblieben: Sie fühle sich wie ein kleines Kind, wenn sie mit Make-up spiele und mit Produkten experimentiere, um herauszufinden, wie man das eigene Aussehen verändern und diese Erkenntnisse mit anderen teilen kann.

Auf die Frage, welchen Rat sie ihrem jüngeren Ich geben würde, antwortet Kattan mit einer einfachen, aber schwer umzusetzenden Einsicht: Man werde es nie allen recht machen können. Sie fragt sich, warum Menschen sich so oft auf die konzentrieren, die sie vielleicht nicht mögen, und darüber nachdenken, was diese über sie denken könnten. Wichtiger sei doch, an die Menschen zu denken, die einen mögen. Warum interessiere man sich so sehr für die Meinung anderer? Aber genau so seien Menschen nun einmal. Diese Haltung zieht sich durch ihre öffentlichen Auftritte: Kattan betont Selbstbestimmung, ohne den Druck durch soziale Medien zu leugnen.

Als starke Frau und Vorbild für viele junge Mädchen nimmt sie ihre Rolle sehr ernst. Sie spricht von einer großen Verantwortung, gerade in sozialen Netzwerken. Als das Geschäft erste Erfolge zeigte, habe sie kurzzeitig sogar etwas Motivation verloren, weil sie begann, über den Sinn ihres Tuns nachzudenken. Tut sie es für ihre Kinder, für ihre Tochter, für die nächste Generation? Dabei denkt sie nicht nur an Millennials, sondern auch an die Generation Alpha, also an die zwischen 2000 und 2010 Geborenen. Sie macht sich Gedanken über die Welt, in der diese Kinder aufwachsen. Soziale Medien setzten Menschen unter Druck, und sie wisse, wovon sie spreche. Ständig würden Inhalte und Anreize auf die Nutzer einprasseln. Diese Entwicklung beschäftigt sie intensiv.

Auch das Thema Mutterschaft hat ihren Blick verändert. Kattan sagt, sie habe nie Mutter werden wollen. Als sie dann ein Kind bekam, habe sie die Aufgabe sehr ernst genommen. Sie habe ihre Tochter oft gefragt, ob sie wolle, dass sie nur Mama sei. Die Antwort habe gelautet: Nein, sie wolle, dass sie Huda Beauty ist. Diese Haltung beeindruckt Kattan bis heute. Ihre Tochter sehe Frauen heute ganz anders, als es frühere Generationen taten, für die Feminismus oft etwas Negatives hatte. Für die Tochter sei der Begriff positiv besetzt. Sie erlebe eine andere Welt, und Kattan möchte, dass es eine gute Welt ist, in der Menschen wieder zu sich selbst finden.

Die Zusammenarbeit mit ihren Schwestern beschreibt sie als herausfordernd, aber produktiv. In Familien sage man sich oft Dinge, die man anderen Menschen nie sagen würde. Grenzen würden definitiv überschritten. Dennoch arbeiten die drei schon lange zusammen und haben gelernt, sich anzupassen. Jeder habe seine eigene Art, die man verstehen müsse. Es gebe Zeiten, in denen man sich auf den anderen einstelle, und Zeiten, in denen das nicht gelinge. Letztlich verfolgen sie aber ein gemeinsames Ziel und versuchen, sich gegenseitig voranzutreiben. Diese Mischung aus familiärer Nähe und unternehmerischer Disziplin prägt das Unternehmen bis heute.

Dubai spielt für Kattan eine zentrale Rolle. Sie bezweifelt, dass sie in den USA erreicht hätte, was sie in Dubai erreicht hat. Die Vision von Scheich Mohammed und dessen Fähigkeit, ohne Grenzen zu denken, hätten sie inspiriert. Dubai sei ein neues Land, die Vereinigten Arabischen Emirate seien erst 40 Jahre alt. Welches Land schaffe in so kurzer Zeit Vergleichbares? Für Menschen, die hierherziehen, sei der Anfang sicher schwer, sie müssten einen Weg finden, sich anzupassen. Kattan selbst will nach eigenen Worten nie wieder weg. Dubai ist für sie nicht nur ein Standort, sondern ein Teil ihrer Erfolgsgeschichte.

Was sie am Boden hält, ist ihre Familie, insbesondere ihre Tochter. Sie beschreibt sie als sehr geerdet und einfach. Die Tochter interessiere sich nicht für den ganzen Rummel, sie wolle nur mit Barbies spielen, Schleim machen oder bunte Regenbogen-Spaghetti. Wenn Kattan nach Hause komme, erfülle das ihre Seele. Inmitten eines globalen Unternehmens und einer Marke, die ständig wächst, bleibt das private Familienleben ein Anker. Diese Erdung hilft ihr offenbar, die öffentliche Aufmerksamkeit und den Druck des Beauty-Geschäfts auszubalancieren.

Für die Zukunft hat Huda Kattan größere Pläne als nur eine Kosmetikmarke. Sie will die Welt der Schönheit mit anderen starken Marken mitgestalten und eine Pionierrolle übernehmen. Schönheit soll als etwas Warmes und Wohliges verstanden werden, sodass sich Menschen durch sie gestärkt fühlen. Der Anspruch ist, die Beauty-Welt nicht nur durch Produkte zu prägen, sondern durch eine Haltung: Selbstbewusstsein, Fürsorge und die Überzeugung, dass Schönheit mehr ist als Oberfläche. Kattan will zeigen, dass ein Unternehmen aus Dubai globale Maßstäbe setzen kann, ohne die familiären Wurzeln und die persönliche Geschichte zu verleugnen.

Source: euronews News