Europe's AI Compute Tripling Plan: Key Recommendations, Legislative Steps, and Criticisms

New Supercomputers and 'Gigafactories'
The report calls for deploying at least 9 new AI-optimised supercomputers under the EuroHPC Joint Undertaking, which alone would more than triple the bloc's current AI computing capacity . It also recommends creating 'AI Gigafactories'—next-generation facilities roughly 4 times more powerful than the current AI Factories, with €20 billion mobilised through the InvestAI initiative .

Broader Ecosystem and Coordination
Beyond hardware, the experts recommend boosting AI uptake across sectors, setting up public-private partnerships for sectoral AI ecosystems, and ensuring world-class research and talent attraction . Crucially, they call for a coordinated EU-wide approach to overcome the current fragmentation where 27 member states largely act alone on procurement and infrastructure .

Legislative and Policy Steps Already Taken

The EU has not waited for the report to act. Several major measures are already in force or under development.

The AI Act (Regulation (EU) 2024/1689), which entered into force on 1 August 2024, is the world's first comprehensive horizontal AI law, establishing harmonised rules for trustworthy AI across the EU . Its rules are being phased in, with obligations such as the prohibition of unacceptable-risk AI practices already applicable .

The AI Continent Action Plan, published by the European Commission in April 2025, goes beyond regulation to combine oversight with investment, infrastructure, and skills development . It sets five main objectives: building large-scale AI data and computing infrastructure, improving access to data, fostering innovation, accelerating AI adoption, and ensuring a skilled workforce .

The AI Factories programme, with a €10 billion budget for 2021–2027, has already procured or planned at least 13 operational AI factories by 2026, including 9 new AI-optimised supercomputers . In June 2024, the EuroHPC regulation was amended to create a dedicated AI pillar for the Joint Undertaking, enabling it to acquire, upgrade, and operate these AI factories .

The InvestAI / Gigafactories initiative aims to mobilise €20 billion to deploy up to 5 AI Gigafactories, each roughly 4 times more powerful than current Factories . This was announced at the AI Action Summit in Paris in February 2025 .

Other steps include the proposed Cloud and AI Development Act (CADA) (June 2026), which aims to increase the EU's share of the global cloud market from 15% to 30% ; the Data Union Strategy (November 2025) to make more high-quality data available for AI ; and the launch of an AI Act Service Desk in October 2025 to help ensure smooth implementation of the rules .

Criticisms: Big Tech Windfall and Market Fragmentation

Despite the ambitious plans, significant criticism has emerged along two main lines.

Risk of a Big Tech Windfall
The EU's flagship €200 billion (~$230 billion) AI investment plan relies heavily on public-private co-investment. Critics warn that much of this funding could flow primarily to large US hyperscalers—Amazon, Microsoft, Google—that already dominate the European cloud and AI infrastructure market . EU regulators are currently evaluating whether Amazon and Microsoft should be classified as 'gatekeepers' under the Digital Markets Act for their cloud and AI services . Dutch non-profit SOMO published a report in July 2025 warning that European governments, in their rush to invest, risk 'playing into the hands of Big Tech' as leading AI start-ups like OpenAI, Anthropic, and Mistral are deeply dependent on a small group of powerful US firms for hardware, infrastructure, and platforms .

The Cloud and AI Development Act (CADA) has also drawn fire. The Computer & Communications Industry Association (CCIA Europe) argued that the Act 'introduces discriminatory measures that directly undermine the EU's own digitalisation goals', by excluding trusted international technology providers based on their headquarters location, which could create advantages for favoured European providers while potentially rewarding large incumbents that can most easily comply .

Market Fragmentation Risks
A second major line of criticism concerns the EU's own internal market. Bruegel, a leading Brussels-based think tank, notes that EU demand is fragmented across 27 member states, each acting largely alone, with no coordinated mechanism to direct public procurement towards European AI hardware or to generate revenues that fund R&D . The think tank warns that unlike China, the EU has limited tools to generate a captive market that powers a domestic improvement cycle .

The European Commission itself, in its 2026 competitiveness report, admitted that Europe 'risks losing its edge in the race for innovation' due to labour shortages, problems scaling up, low patent numbers, and R&D spending below the 3% of GDP target . The Commission's own CADA proposal acknowledges that divergent national approaches to data centre capacity expansion, sustainability requirements, and permitting procedures are creating 'regulatory disparities that could undermine the internal market' .

The Stark Scale Gap
Even with the tripling target, the scale gap between the EU and its competitors is enormous. Europe hosts just 5% of global AI compute compared to the US's 80% . The EU's flagship €200 billion plan is dwarfed by the more than $400 billion US tech firms spent largely on expanding AI infrastructure in 2025 alone . One forecast projects that under current trends, Europe's share will briefly rise to 8% by 2028 before declining again to just over 5% by 2031, underscoring the gap between ambition and execution .