Swedish universities are preparing bids under a government scheme that aims to muster the nation's fledgling capabilities in artificial intelligence (AI) and turn them into a world-leading ecosystem capable of breakthroughs that subvert Silicon Valley's supremacy. This ambitious effort, known as the Clusters of Excellence strategy, represents a bold gamble for a country that acknowledges its current weakness in AI.
Background and ambition
Already considered a world leader for the intensity of its science and innovation, Sweden has set an ambition to rank among the top five nations in 10 areas of science within a decade, with fundamental AI breakthroughs its number one priority. The government surprised innovation experts with its AI ambitions, given that Sweden is self-consciously weak in this field. In its Clusters of Excellence strategy, it plans to build university-led, techno-industrial ecosystems in fields where it has established strengths. Each cluster will receive 10 years of funding, totalling about kr1.7bn (€160m), to become not only a world-leading centre of scientific research but also ranked similarly for its ability to turn research into innovative technology, startups, private capital, talent attraction, and a self-sustaining industrial ecosystem. The cluster programme demands breakthroughs in fundamental AI technology that it says should “leapfrog … today’s computational paradigms, and create the next generation of intelligent systems”.
Architect's perspective
Marika Edoff, secretary general for research infrastructures at the Swedish Research Council and one of the architects of the programme, explained the top-down requirement for fundamental AI: “We were supposed to be bottom-up, with the exception of fundamental AI, because our assignment from the government said that specifically. I think that was part of a political negotiation.” She clarified that fundamental AI means game-changing breakthroughs in AI itself, not merely applying AI to other fields like materials or biotech, as many preparatory cluster plans have proposed. However, she expressed caution: “To be world-leading in AI, I don’t think we can do that. I think we could be world-leading in a particular part of AI.” Fredrik Heintz, who heads AI research at Linköping University and is involved in a bid to form an AI cluster, described “the dream” as making breakthroughs that render large language models (LLMs) used by OpenAI and US tech giants obsolete. “Whether it’s realistic to get that far is unclear,” Heintz admitted. “The desire and the goal is to come up with the next paradigm. But I don’t think that’s necessary to be successful. What’s important is to find an area where we can be top-five in the world. It needs some form of specialisation to have a chance of actually succeeding.” Sweden possesses strengths in computer vision, robotics, reasoning and agentic AI, but the cluster programme’s demand that all fields prioritise fundamental AI has stretched expertise thin.
Reactions from cluster experts
Norwegian cluster consultant Christian Rangen, while hosting a workshop with cluster bidders for Swedish innovation agency Vinnova, called the initiative “freakin’ ambitious” because it aspires not merely to excellence but to global leadership in both research and innovation. He warned bidders: “Top-five in the world by 2035. That is really freakin’ ambitious! You can’t do this with a small ecosystem mindset. A lot of the clusters that we have in Europe, they struggle because they’re too local.” Mats Nordlund, research director at AI Sweden, who is helping mount a bid for a secure and trustworthy AI cluster, pointed out a crucial oversight: AI breakthroughs are typically driven by industry, not academia. “AI today is not driven by universities,” he said. “Universities have a really hard time keeping up with what goes on in industry. It’s driven today by the big tech and automotive companies. Industry gets stuck. Then it becomes a research problem.” Yet only universities can mount bids to form clusters, creating a structural tension. Nordlund helped establish a similar cluster programme in Moscow before Russia invaded Crimea.
Star researcher recruitment and startup challenges
Sweden plans for each cluster to build a team of world-class researchers around a “superstar” who will sit at the centre of a concentration of companies, entrepreneurs, and financiers. In March, as Sweden prepared to issue its call for bids, French AI superstar Yann Le Cun launched a robotics AI startup with $1bn seed funding from notable backers including Nvidia, Eric Schmidt, Tim Berners-Lee, Toyota, and the French state. This highlights the scale of competition. In contrast, Swedish startup Superintelligence Computing Systems (Sicsai) is raising €10m for robotics AI, claiming breakthroughs in fundamental AI. CEO Karim Nouira said Sicsai is “way ahead” of Le Cun in physical AI but acknowledged that Swedish firms need clusters to raise capital and build industry connections. “Europe is lagging behind because there’s almost no interest for significant deep tech investments,” Nouira said. “There are some smaller VCs doing small tickets, but it doesn’t take you anywhere. It just dilutes your company. It’s really difficult to get pilot projects going. The large companies say that they support startups, but when you want to do a pilot project, they have very restricted budgets and very little wiggle room.”
US dominance and Swedish computing resources
US dominance in LLM-based AI is overwhelming. Market research non-profit EpochAI charts the race for AI among US computing giants, showing that “compute-poor” nations like Sweden struggle to compete. OpenAI spent €16bn on compute last year; US big-tech firms collectively spent $700bn building infrastructure, roughly the size of Sweden's entire economy. Microsoft is finishing a €3bn AI compute facility in Sweden with 20,000 GPUs, coinciding with Sweden's own Arrhenius supercomputer (€68m, 1,500 GPUs) and the Mimer centre (€30m, 400 GPUs) under the EU AI Factory scheme. Bruno Lecointe, head of HPC at French firm Bull, noted Mimer will allocate half its budget for software and expertise to help small firms re-engineer algorithms, serving as a stepping stone to larger EU factories.
Swedish strengths in innovation and research
Despite its AI weakness, Sweden ranks second worldwide for innovation, has the most researchers per capita, and is among the largest in education spending. R&D spending as a proportion of GDP is fifth globally. Its strengths in bioscience, chemistry, and materials are complemented by two major sub-atomic experiments comparable to CERN. German clusters, used as exemplars, succeed by building on existing strengths, said Bastian Krieger of ZEW Mannheim. Sweden's plan for 10 clusters “ambitious, but not impossible,” he said, and the top-down diktat on fundamental AI might work if they attract scientific stars. James Wilson of the Basque Institute of Competitiveness noted clusters can be built in weak fields, as the Basque region did with biosciences, but only with existing institutes and lead firms to support them.
Path forward
Sweden's ambition to lead in AI is a high-risk, high-reward strategy. The cluster programme's success hinges on attracting top talent, fostering industry-academia collaboration, and specialising in niche areas where Sweden can genuinely excel. The next decade will reveal whether this small, innovation-rich nation can indeed leapfrog the computing paradigms of Silicon Valley.
Source: ComputerWeekly.com News