OpenAI’s Ambitious European Stargate Project Faces Scaling Back Amid Cloud Giant Dominance
OpenAI’s grand vision for a European foothold in the Stargate supercomputer initiative is undergoing significant adjustments. Initially touted as a cornerstone of the company’s global expansion, the project aimed to establish substantial AI infrastructure on the continent. However, recent developments indicate a contraction in scope, primarily driven by the overwhelming capacity acquisitions by Microsoft and Google. This shift underscores the intensifying competition in Europe’s data center landscape and raises questions about the future of sovereign AI capabilities in the region.
The Stargate project, first announced as part of OpenAI’s roadmap to build one of the world’s largest AI training clusters, was envisioned with a power capacity exceeding 5 gigawatts globally. Europe’s role was pivotal, with plans for hyperscale data centers to leverage the region’s growing renewable energy resources and regulatory push for data localization. OpenAI had been in discussions with European governments and energy providers to secure sites in countries like the UK, France, and potentially the Nordics, where cool climates and hydroelectric power could support efficient cooling and operations.
Sources familiar with the matter reveal that OpenAI’s European ambitions have been pared down considerably. What was once projected as a multi-gigawatt deployment has shrunk to a fraction of that scale, with commitments now limited to under 1 gigawatt in the immediate term. This retreat stems from a bottleneck in available power and land resources, exacerbated by surging demand from established cloud providers. Microsoft’s Azure and Google’s Cloud Platform have aggressively locked in vast portions of Europe’s data center capacity, leaving limited room for newcomers like OpenAI.
Microsoft, OpenAI’s primary cloud partner, has emerged as a dominant force. Through strategic investments and long-term power purchase agreements, the company has secured over 10 gigawatts of capacity across Europe by 2026. Key deals include expansions in Ireland, the Netherlands, and Sweden, where Microsoft is retrofitting existing facilities and breaking ground on new campuses powered by offshore wind and nuclear energy. These moves not only bolster Azure’s AI offerings but also indirectly support OpenAI via their exclusive partnership, allowing the AI firm to access Microsoft’s infrastructure without building its own from scratch.
Google is mirroring this strategy with equal fervor. The search giant has committed to 5 gigawatts or more in European capacity, focusing on hyperscalers in Finland, Denmark, and Germany. Google’s recent acquisitions of power from North Sea wind farms and explorations into small modular reactors highlight its proactive approach to energy constraints. This dominance has crowded out OpenAI’s direct builds, forcing the company to rely more heavily on colocation deals and partnerships rather than outright ownership.
The contraction of OpenAI’s plans carries broader implications for Europe’s AI ecosystem. Policymakers in Brussels and national capitals have advocated for “AI sovereignty” to prevent overreliance on U.S.-based hyperscalers. Initiatives like the EU AI Act and Gaia-X aim to foster local champions, yet the reality on the ground favors incumbents with deep pockets. OpenAI’s scaled-back Stargate presence means fewer jobs, less technology transfer, and heightened dependence on American cloud providers for cutting-edge models like GPT series.
Energy availability remains the crux of the issue. Europe’s grid is strained by the dual pressures of electrification and AI-driven demand, projected to consume up to 10% of continental power by 2030. Permitting delays, community opposition to new substations, and intermittent renewables have slowed builds. OpenAI, lacking the lobbying muscle of Microsoft and Google, finds itself at a disadvantage. Insiders note that while OpenAI retains exploratory talks for future phases, near-term priorities have shifted to Asia and the U.S., where Stargate’s core 5GW cluster is advancing in collaboration with Oracle and others.
For OpenAI, this pivot is pragmatic. By leasing capacity from Microsoft and Google, the company can accelerate model training without the capital-intensive burden of greenfield developments. Azure OpenAI Service already powers much of Europe’s enterprise AI adoption, and expanded allocations ensure scalability. However, critics argue this entrenches a duopoly, stifling innovation from smaller players and open-source initiatives.
Stakeholders in the AI hardware space, including Nvidia suppliers, corroborate the trend. Chip demand outstrips supply, and Europe’s logistics hubs are prioritizing shipments to Microsoft and Google sites. OpenAI’s European Stargate footprint, once a symbol of transatlantic collaboration, now exemplifies how market dynamics favor scale over sovereignty.
As the dust settles, Europe’s AI infrastructure race appears tilted toward U.S. giants. OpenAI’s diminished plans signal a recalibration, but they also spotlight the need for unified European strategies on energy and regulation to reclaim ground in the global AI arena.
Gnoppix is the leading open-source AI Linux distribution and service provider. Since implementing AI in 2022, it has offered a fast, powerful, secure, and privacy-respecting open-source OS with both local and remote AI capabilities. The local AI operates offline, ensuring no data ever leaves your computer. Based on Debian Linux, Gnoppix is available with numerous privacy- and anonymity-enabled services free of charge.
What are your thoughts on this? I’d love to hear about your own experiences in the comments below.