A New Kind of Network, Built for and by AI
For decades, building a networking company required a large team of engineers and years of development. A new startup is betting that AI changes that equation entirely. The company is launching an Agentic Defined Networking platform that lets users and their AI agents build private networks. It also announced a $38 million Series A round led by Andreessen Horowitz, with participation from Animo Ventures and Focal.vc.
The founder is Barrett Lyon, a serial networking entrepreneur who has founded and sold multiple vendors over the past two decades. His previous companies include Prolexic, acquired by Akamai; defense.net, acquired by F5 Networks; and Netography, acquired by Vectra AI. Now he has turned his attention to AI with a platform that aims to make network creation and operation autonomous.
The project has its roots in a hobby called Darkflare, which Lyon created as a TCP tunnel over a major content delivery network. The tool was designed to tunnel through public networks in a way that helps avoid detection and potential censorship. The response to that project led Lyon to consider a VPN that cannot be stopped and is autonomously managed by AI. 'The hobby turned into an obsession, and the obsession turned into a business,' Lyon said.
How the Agentic Defined Network Works
The platform combines custom VPN transports, a company-run mesh, and a separate control plane. Each layer is designed to give AI agents and human users fine-grained control while maintaining resilience and privacy.
Covert Transports
The client supports WireGuard and custom transports for VPN traffic over WebSocket, Session Initiation Protocol (SIP), ping, QUIC, a major content delivery network, and HTTPS. These transports are designed to get through firewalls. By blending VPN traffic with common protocols, the system aims to avoid blocking and keep connections alive even in restrictive network environments.
Mesh Networking
The mesh connects every site to every site with encryption. It runs over IPv4 and IPv6 and is orchestrated through an application programming interface (API). This means that once a node joins the mesh, it can communicate directly with other nodes without complex manual configuration. The mesh is designed to be resilient and scalable, allowing new sites and devices to be added programmatically.
Control Plane
The client application is a device, not a control plane. Routing, port forwarding, and firewall rules are managed in a separate portal. This separation helps keep the control plane secure and allows administrators to manage policies centrally while devices remain lightweight.
Addressing
Devices on the mesh are not behind network address translation (NAT) or carrier-grade NAT, according to Lyon. That allows peer-to-peer (P2P) applications to work more reliably. Direct addressing can improve performance for real-time communication, file sharing, and other decentralized applications that struggle behind NAT.
DNS and Certificates
The company runs its own Domain Name System (DNS) root with 196 domains and its own certificate authority. Lyon said the team wrote its own Border Gateway Protocol (BGP) daemon and runs internal BGP across the mesh. Running its own DNS and certificate infrastructure gives the platform greater control over identity, routing, and trust. It also reduces dependence on external services that could be blocked or compromised.
Identity and Discovery
The address book holds no phone numbers, email addresses, or names. Users exchange cryptographic keys and assign an alias to each contact. A directory service seeds a new device with the data needed to find its contacts. After that initial call, devices sync with each other. This design prioritizes privacy and minimizes the metadata that could identify users.
How AI Runs the Network
The need for the platform to be run by AI is pragmatic. Lyon said he turned to AI because he could not manage the network by himself. 'The whole thing is run completely by AI, so behind the scenes, I mean, it's a network with 31 locations around the world, and internally there's a mesh network, and I can't, as a human being, manage all of this by myself,' Lyon said.
Lyon said the team modeled every site, down to the wire and the optic, in a virtual model before building. An infrastructure management system running the company's own AI on its own hardware then ordered the installation. It orchestrated shipping and delivery through data center APIs. Human technicians performed the remote smart hands installations. This hybrid approach combines AI-driven planning and orchestration with human execution at the physical layer.
The company also built tools for agents to work with users and with the network. An agent gateway gives an AI agent an identity in the company's chat app. The user pastes a credential into the agent. The agent obtains its certificate and appears in the user's chat. Users can create group chats with several agents. In one example, Lyon said one agent runs BGP while others handle other tasks.
The API is self-teaching. An agent given the configuration URL can learn how to build networks without further instructions. It then asks what the user wants to do, such as link an office to a house or give a Linux box a private IP address. Agents can also create their own domains and host names, sign certificates, and run web servers and applications inside the network.
Administrators issue read-only or admin tokens for each agent. Tokens can expire or be revoked. An administrator can also cut an agent's gateway access or delete the agent in the app. 'You don't just give them your account credentials to the master account,' said Lyon. This token-based permission model is intended to limit the blast radius if an agent is compromised or misbehaves.
Why the Name Is a Reclamation
The term 'doxxing' usually refers to publishing someone's private information, but that is the exact opposite of what the platform is about. As it turns out, Lyon registered the domain more than 30 years ago, as it was his hacker name at the time. 'So, I'm just reclaiming the word, and it is kind of funny that we're doing the exact opposite of what doxing is. And, I don't know, it just seems fitting for me,' he said.
That history gives the project a distinct identity. Rather than exposing personal data, the platform aims to protect communications and make networks more private and resilient. The name is a nod to early hacker culture and a statement about privacy in an era of increasing surveillance and censorship.
Enterprise Plans and Open Source
Looking forward, the plan is to build out more enterprise offerings as well as release some code as open source. Fundamentally, the technology is not just about using AI for networking but also about using AI to build the technology itself. That distinction matters: the company is not simply adding AI assistants to existing network tools. It is using AI to design, deploy, and operate the network from the ground up.
'You can have a vision of where it's going to go, and then that vision hits a wall when you hit, you know, your own personal limitations,' Lyon said. 'The AI just takes it well beyond, running this with an agentic partner, so to speak, turns this into fire.'
The $38 million Series A round led by Andreessen Horowitz, with participation from Animo Ventures and Focal.vc, suggests strong investor interest in AI-driven network automation. The funding will likely support expansion of the mesh, further development of agent tools, and enterprise features. The company is also expected to open source some components, which could accelerate adoption and allow developers to build on its protocols.
As AI agents become more capable, the idea of letting them manage network infrastructure raises important questions about security, governance, and trust. The platform's approach includes token-based access, revocation, and separate control plane management. These safeguards are designed to keep humans in ultimate control while delegating routine and complex tasks to AI. The balance between autonomy and oversight will be a key challenge for the company and its customers.
Source: Network World News