Europe is expanding its infrastructure, computing capacity and domestic artificial intelligence capabilities, but it continues to lag behind the United States and China in the most advanced technologies. According to Polish MEP Tobiasz Bocheński, the European Union has become particularly effective at creating regulations, but rules alone will not be enough to improve Europe’s competitiveness or ensure cybersecurity in an era of AI-powered threats.
“Virtually all of the latest AI models are being developed primarily in the United States and, secondly, in East Asia, mainly China. Among the ten biggest players in this industry, there is only one European company, and it is French. The European Union is vulnerable both commercially and in terms of the most advanced technologies,” Tobiasz Bocheński, a Law and Justice MEP and member of the European Conservatives and Reformists Group, told the Newseria news agency.
He argues that regulation cannot compensate for Europe’s limited role as a producer of cutting-edge technologies.
“No legislation adopted by the European Commission will change this, because Europe is essentially a technology taker rather than a technology provider,” he said.
European AI companies remain relatively few
This year’s Forbes AI 50 ranking of promising companies developing and using artificial intelligence includes five European businesses: France’s Mistral AI, Swedish companies Lovable and Legora, Germany’s Black Forest Labs and UK-based Synthesia.
Among them, Mistral AI has raised the most funding, with more than $3.1 billion.
Bocheński believes that the European Commission has placed excessive faith in its ability to shape the digital market through regulation.
“The European Commission promotes the idea that by introducing new directives and regulations it can force large foreign companies, through administrative decisions, to protect certain values it considers essential. But one important fact must be recognised: none of those major providers is backed by European capital,” he said.
According to the MEP, there is also a risk that increasingly restrictive regulation could reduce European consumers’ access to the most advanced digital services.
“This works both ways, because it could end with Europeans having limited access to the highest-quality digital services available to Americans and Chinese users,” Bocheński said.
He argues that Europe failed to invest enough in innovation when it had the opportunity.
“The European Union spent tens, if not hundreds, of billions of euros on objectives that turned out to be wasteful at a time when it should have been developing innovation and modern technologies. Now we have a serious problem because, when it comes to cutting-edge technology, we risk becoming a technological museum.”
Brussels wants to build an AI continent
In April 2025, the European Union unveiled its coordinated AI Continent Action Plan, based on five pillars: infrastructure investment, better use of data, sectoral deployment of AI, skills development and regulatory simplification.
The objective is to make better use of the economic and technological opportunities created by artificial intelligence.
In 2026, the European Commission presented further initiatives linked to the plan, including a cybersecurity and AI action plan.
Brussels wants to create a more coordinated system for responding to threats involving advanced AI models used in cyberattacks.
The proposals include building European capabilities for assessing AI models from a cybersecurity perspective, developing rules with ENISA for cybersecurity organisations to access advanced models, and establishing secure platforms where AI systems could be tested in simulated environments.
Chips, cloud infrastructure and AI gigafactories
In June, the Commission also presented a European technological sovereignty package.
It includes two major legislative initiatives: the Chips Act 2.0 and legislation intended to support the development of cloud computing and artificial intelligence.
The aim is to reduce Europe’s dependence on suppliers from outside the EU while encouraging investment in data centres, computing infrastructure and AI deployment.
The Commission has also announced further development of AI gigafactories and plans to create European financing capacity dedicated to investments in technological sovereignty.
Despite those initiatives, Bocheński remains deeply sceptical about Europe’s chances of catching up with the global leaders.
“I do not see a realistic chance of us catching up with the world. I am absolutely pessimistic about it. That train has already left the station,” he said.
“When we look at companies leading in quantum computing, the most advanced AI models, energy prices – which are essential for developing modern data centres – major corporations and robotics, the European Union is far behind China and the United States. At this point, we are becoming subcontractors.”
US leads in advanced AI models, China dominates research output
China currently leads the world in the number of AI-related publications, citations and patents granted, while US patents tend to have greater technological impact.
According to Stanford University’s 2026 AI Index Report, the United States developed 59 notable AI models last year, compared with 35 in China.
That gap in frontier models is particularly important because the most advanced AI systems require huge investments in computing power, specialised chips, electricity, data infrastructure and highly skilled personnel.
Bocheński believes Europe’s response should therefore focus much more strongly on deregulation.
“I hope the new EU strategy will not be adopted. The European Union should focus on deregulation. We are being mocked around the world because of the AI Act and other regulations,” he said.
He also criticised discussions about possible new forms of taxation affecting the digital and AI sectors.
“There are proposals concerning a digital tax – not simply the one discussed in relation to corporations such as Facebook, but potentially a mechanism under which AI companies would transfer part of their revenues to intellectual-property owners. No country that is successfully developing AI has anything comparable.”
Europe has around 5% of global AI computing capacity
A report prepared by the EU AI Office, Enhancing Competitiveness, Sovereignty and Security of the European Union in Frontier AI, argues that the EU should aim for a share of global computing capacity proportional to its approximately 15% share of global GDP.
Europe’s current share of global AI computing capacity is estimated at only around 5%.
Closing that gap would require major investment in data centres, chips, electricity generation and transmission infrastructure.
Energy is becoming particularly important because training and operating increasingly advanced AI models requires enormous and rapidly growing amounts of electricity.
The AI Office report identifies computing infrastructure and the energy required to power it as two of Europe’s most urgent priorities for the next two years.
Talent and start-ups continue to leave Europe
Bocheński argues that Europe’s regulatory environment also affects its ability to retain entrepreneurs, researchers and innovative businesses.
“Our problem is that there are too many regulations, restrictions and various legal mechanisms that cause the most talented people, the most creative think tanks and the most innovative start-ups to leave Europe,” he said.
He pointed to the competitiveness report prepared by former European Central Bank president Mario Draghi as evidence that the structural problem has already been widely recognised.
“This phenomenon was described in the Draghi report published at the end of 2024. We are now in 2026 and nothing has changed. The speed of our response simply does not match the speed of change in today’s world.”
In Bocheński’s view, Europe’s policy debate is increasingly disconnected from the scale of technological ambitions seen in the United States.
“If we look at Elon Musk’s companies and their plans involving robotics, autonomous passenger vehicles, trucks and public transport, attempts to build data centres in orbit and transmit data from space, then what we are discussing in Europe simply looks ridiculous,” the MEP said.
“The gap will continue to widen. We have stopped while they are continuing to grow.”
Frontier AI is advancing at unprecedented speed
Experts cited by the EU AI Office warn that frontier AI models have developed at an unprecedented rate.
Within just three years, the most advanced systems have moved from struggling with relatively basic tasks to achieving results close to the limits of many existing benchmarks.
At the same time, Europe’s position in frontier AI remains weak relative to the size and potential of its economy.
The AI Office report stresses that development of the most advanced models is driven primarily by private investment.
For Europe to improve its position, the experts argue, it needs not only public funding but also better conditions for private capital, fewer obstacles to investment and access to internationally competitive operating conditions.
That means the European AI debate is increasingly moving beyond regulation alone.
For the next stage of the global technology race, access to computing power, affordable energy, capital, advanced chips, research talent and an environment in which companies can scale quickly may prove far more important than Europe’s ability to write new rules.
