The EU plans to bring 3-5 AI gigafactories online by 2028 through its EuroHPC Joint Undertaking. With a €7bn budget, this initiative represents Europe's bid to compete in the global AI race against the United States and China. Eastern and Central Europe are uniquely positioned to host these facilities, offering critical advantages in energy costs, regulatory agility and strategic value.

Europe faces familiar obstacles in AI development. It cannot match America's venture capital firepower. According to Stanford’s HAI report, the US captured 67% of global private AI investment in 2023 compared to just 3% for the EU. Nor can it replicate China's state-directed approach, which has channelled an estimated $7bn into AI through government programmes.

European electricity costs remain stubbornly high, and regulatory complexity — while promoting safety and ethics — slows innovation. The EU AI Act, which took force in August 2024, could cost firms €100,000-€500,000 per high-risk system for initial compliance, according to Brookings Institution analysis. This could prove especially punitive for start-ups and smaller businesses. 

However, despite all this, Eastern Europe has some significant advantages on its side. First, lower energy costs. Poland and Czechia, both under consideration for gigafactories, offer industrial electricity at €0.21 and €0.20 per kWh respectively — compared to €0.23 in Germany and €0.24 in the Netherlands, according to 2024 Eurostat data. For a 150-megawatt facility consuming 1.3 terawatt-hours annually, this translates to tens of millions in annual savings, even before bulk power purchase agreements.

AI systems themselves can amplify these savings. Google's DeepMind reduced datacenter cooling energy by 40% using machine learning, while Microsoft reported 15-20% overall energy reductions through AI-driven optimisation. Combined with Eastern Europe's lower baseline costs, this creates meaningful competitive advantage.

Eastern Europe’s second advantage lies in faster regulatory processes. Eastern European regulatory frameworks, built more recently with digital systems in mind, offer faster approval timelines. Poland's data centre construction approval takes 18-24 months compared to 36-48 months in Germany, where legacy consultation processes add complexity and cost.

The Baltic states, particularly Estonia with its e-governance model, have demonstrated regulatory innovation. While claims of Tallinn being the "Silicon Valley of Europe" oversimplify reality (London and Berlin attract more VC investment), these countries show genuine adaptability for emerging technologies.

Finally, the eastern half of the continent benefits from strategic EU support. EU cohesion funding prioritises Eastern European development, potentially covering 40-50% of infrastructure costs. The European Commission's 2024 Digital Decade report notes Poland and Czechia now rank above  the EU average in terms of 5G coverage and fibre availability — critical for AI facilities requiring high-bandwidth connections.

However, when considering Europe’s approach to AI, any honest assessment must acknowledge challenges, which include: 

• Semiconductor dependence: Europe relies entirely on imports of advanced NVIDIA GPUs, subject to US export controls
• Talent drain: Poland alone loses 15,000-20,000 STEM graduates annually to emigration, requiring competitive compensation to reverse
• Grid upgrades: Czech operator ČEPS proposed €2.3bn in transmission improvements through 2030, though plans predate AI facility requirements
• Water resources: Each facility needs 3mn-5mn litres daily for cooling — sustainable access must be demonstrated

Whilst these constraints shouldn’t dishearten those hoping from a brighter future for AI in Europe, they are important to keep in mind. 

On a continent wide level Europe's strength lies in embedding AI into real-world systems. McKinsey estimates 70% of AI's economic value will come from integration into existing business processes — precisely where European manufacturing, automotive, and pharmaceutical strengths can shine. This stands in contrast to US focus on frontier models or China's surveillance applications, 

The gigafactories can also pioneer AI-driven regulatory compliance tools. While the EU AI Act's requirements for human oversight and legal accountability cannot be fully automated, AI can reduce compliance costs by 30-50% through automated documentation, risk assessment, and continuous monitoring. If European facilities develop and export these tools as global standards converge, compliance becomes competitive advantage rather than mere overhead.

When EuroHPC finalises site selections, the evidence supports Eastern and Central European locations. Poland, Czechia and potentially Romania offer the combination of lower costs, regulatory flexibility, EU funding access, and strategic diversification value that make them strong candidates.

Success requires sustained investment beyond the current €7bn, addressing talent retention, upgrading grid infrastructure, and securing semiconductor supply chains. But if executed well, these gigafactories can establish Europe — particularly its eastern region — as a credible third pole in global AI, focused on trustworthy, industrially-embedded systems rather than competing directly on model scale.

Europe cannot out-spend Silicon Valley or out-plan Beijing. But it can build the infrastructure for practical, governed AI deployment at scale. Eastern Europe is where that vision should take root.

Maria Taylor is the founder of Verulam Capital.