Europe's push to become a self-sufficient AI superpower has gained momentum as the continent's long-standing alliance with the US falters. European countries are seeking inventive ways to close the gap with American rivals that have dominated the field for years.
The US currently outstrips Europe across various aspects of AI production, including processor design and manufacturing, data center capacity, model development, and application development. The country has also captured a significant proportion of the money pouring into AI, reflecting in the performance of its homegrown stocks and the growth of its economy.
However, some quarters believe that the US-based leaders are already so entrenched as to make it impossible for European nations to break their dependence on American AI. This sentiment is echoed by the head of Belgium's national cybersecurity organization, who recently stated that Europe had "lost the internet" and should accept a degree of reliance on US infrastructure.
Yet, governments in the UK and EU are not ready to give up on this goal. They have committed hundreds of millions of dollars to minimizing their reliance on foreign AI suppliers, while also pursuing alternative methods for developing competitive products built around imaginative model design.
One possible advantage that European AI labs hold over large American firms is a willingness to develop out in the open. By publishing models for anyone to use or modify, breakthroughs achieved by European labs can compound as they are further refined by collaborators.
In recent months, European leaders have clashed with the Trump administration over various issues, including policing American tech firms and trade policies. The two sides have openly criticized each other's approaches, leading to speculation about a deterioration in the NATO alliance.
Against this backdrop, Europe's reliance on American-made AI begins to look more like a liability. In a worst-case scenario, the US could choose to withhold access to AI services and crucial digital infrastructure. More plausibly, the Trump administration could use Europe's dependence as leverage in trade negotiations.
European nations are attempting to hedge against these risks by bringing the production of AI onshore through funding programs, targeted deregulation, and partnerships with academic institutions. Some efforts focus on building competitive large language models for native European languages, such as Apertus and GPT-NL.
However, bridging the performance gap remains a challenge. American firms like Nvidia, Google, and Meta continue to dominate the field, leaving smaller European competitors struggling to keep up. Even with significant investments in AI research and development, it is unclear whether Europe can catch up with its US rivals.
The push for "digital sovereignty" in Europe raises questions about what exactly this means. Does it demand total self-sufficiency across the entire AI supply chain, or only an improved capability in specific disciplines? Is it possible to achieve digital sovereignty without excluding US-based providers?
Despite these challenges and uncertainties, many experts believe that it is still possible for European labs to make significant progress in AI research and development. The success of China-based AI lab DeepSeek, which shattered the dogma that control over the largest fleet of AI processors determines which firm wins out, has inspired a new wave of European initiatives.
One such initiative is SOOFI, an open-source model development project aiming to put out a competitive general-purpose language model with roughly 100 billion parameters within the next year. According to its lead, Wolfgang Nejdl, this project intends to make Europe's DeepSeek.
The US currently outstrips Europe across various aspects of AI production, including processor design and manufacturing, data center capacity, model development, and application development. The country has also captured a significant proportion of the money pouring into AI, reflecting in the performance of its homegrown stocks and the growth of its economy.
However, some quarters believe that the US-based leaders are already so entrenched as to make it impossible for European nations to break their dependence on American AI. This sentiment is echoed by the head of Belgium's national cybersecurity organization, who recently stated that Europe had "lost the internet" and should accept a degree of reliance on US infrastructure.
Yet, governments in the UK and EU are not ready to give up on this goal. They have committed hundreds of millions of dollars to minimizing their reliance on foreign AI suppliers, while also pursuing alternative methods for developing competitive products built around imaginative model design.
One possible advantage that European AI labs hold over large American firms is a willingness to develop out in the open. By publishing models for anyone to use or modify, breakthroughs achieved by European labs can compound as they are further refined by collaborators.
In recent months, European leaders have clashed with the Trump administration over various issues, including policing American tech firms and trade policies. The two sides have openly criticized each other's approaches, leading to speculation about a deterioration in the NATO alliance.
Against this backdrop, Europe's reliance on American-made AI begins to look more like a liability. In a worst-case scenario, the US could choose to withhold access to AI services and crucial digital infrastructure. More plausibly, the Trump administration could use Europe's dependence as leverage in trade negotiations.
European nations are attempting to hedge against these risks by bringing the production of AI onshore through funding programs, targeted deregulation, and partnerships with academic institutions. Some efforts focus on building competitive large language models for native European languages, such as Apertus and GPT-NL.
However, bridging the performance gap remains a challenge. American firms like Nvidia, Google, and Meta continue to dominate the field, leaving smaller European competitors struggling to keep up. Even with significant investments in AI research and development, it is unclear whether Europe can catch up with its US rivals.
The push for "digital sovereignty" in Europe raises questions about what exactly this means. Does it demand total self-sufficiency across the entire AI supply chain, or only an improved capability in specific disciplines? Is it possible to achieve digital sovereignty without excluding US-based providers?
Despite these challenges and uncertainties, many experts believe that it is still possible for European labs to make significant progress in AI research and development. The success of China-based AI lab DeepSeek, which shattered the dogma that control over the largest fleet of AI processors determines which firm wins out, has inspired a new wave of European initiatives.
One such initiative is SOOFI, an open-source model development project aiming to put out a competitive general-purpose language model with roughly 100 billion parameters within the next year. According to its lead, Wolfgang Nejdl, this project intends to make Europe's DeepSeek.