Cerebras plans 200MW of AI compute capacity in Europe by end of 2027

Cerebras Systems is building out a major European data center footprint, with CEO Andrew Feldman announcing plans for 200 megawatts of AI compute capacity across the continent by the end of 2027, a portion of which is earmarked for OpenAI workloads under the companies' existing partnership.

What's new

Per Cerebras' announcement: "Cerebras will bring its first European data center capacity online by the end of 2026, with rapid build-out across France and the Nordics. The company plans to expand total capacity to 200 MW by the end of 2027, with a portion of that capacity expected to support OpenAI workloads as part of the companies' existing partnership."

Specifics from the announcement:

• First European data center capacity comes online by the end of 2026, with rapid build-out continuing across France and the Nordics.
• Total planned capacity reaches 200 MW by the end of 2027.
• Data centers are slated specifically for Norway and Finland, chosen in part for their abundant, low-carbon hydroelectric power.
• A portion of the capacity is committed to serving OpenAI workloads, extending the two companies' existing compute partnership into Europe.
• CEO Andrew Feldman announced the plans on stage at the RAISE Summit in Paris on July 9, appearing alongside OpenAI's Sachin Katti.

Feldman framed the buildout around customer demand: "Our customers don't just want AI compute. They want it close to home, powered responsibly, and available fast. This expansion and capacity plan reflects our confidence in Europe as a long-term growth market for Cerebras."

Context

Cerebras has spent 2026 aggressively scaling beyond its original wafer-scale-chip niche: the company went public in May with a $5.5 billion IPO that valued it near $66 billion on debut, and has since announced infrastructure and partnership expansions at a rapid clip, including a real-time speech-to-speech AI pipeline built with Hugging Face on Gemma 4. This European buildout follows a broader industry pattern of AI infrastructure providers racing to establish regional capacity as European enterprises, research institutions, and governments increasingly seek alternatives to compute concentrated in the U.S. and Asia — both for latency reasons and out of a desire for data-sovereignty options outside American and Chinese cloud providers.

Why it matters

For European AI developers and enterprises, dedicated low-latency Cerebras capacity — paired with hydro-powered, lower-carbon-footprint siting in Norway and Finland — offers a concrete alternative to routing inference workloads through U.S. data centers. For Cerebras, the buildout is a bet that regional compute sovereignty will be a durable purchasing preference rather than a passing trend, and the explicit tie to OpenAI workloads signals the partnership between the two companies is deepening rather than staying U.S.-only. It also puts Cerebras in more direct infrastructure competition with Nvidia, AMD, and hyperscaler cloud regions in the fight for AI compute market share on the continent, at a moment when European governments and companies are actively looking to reduce dependence on both American and Chinese AI infrastructure providers.