Cerebras targets Europe with multibillion-dollar AI data center expansion challenging Nvidia

Cerebras Plans Massive European Expansion with 200MW of AI Data Center Capacity

AI chip startup Cerebras announced an aggressive European expansion strategy that will bring 200MW of AI data center capacity to the continent by the end of 2027, marking a multibillion-dollar investment to challenge sector leader Nvidia 1

1

. The California-based company, which currently operates three AI data centers in France, Finland, and Norway, plans to bring its first expanded European capacity online by the end of 2026, followed by rapid deployment across the Nordics 2

2

. CEO Andrew Feldman described the initiative as "massive expansions" worth several billion dollars during the RAISE Summit in Paris 1

1

.

Soaring Demand for AI and Data Sovereignty Concerns Drive Investment

The European expansion addresses soaring demand for AI compute capacity as businesses, research institutions, and governments seek alternatives to infrastructure concentrated in the US and Asia 1

1

. Data sovereignty concerns have intensified amid transatlantic tensions, making European organizations wary of overreliance on US providers 1

1

. Feldman told AFP that demand for computing power to run generative AI in Europe is "extraordinary... growing very, very quickly," with growth "faster than we can keep up" 1

1

. The company's strategic positioning aims to meet unique European requirements around data control and compliance with local regulations 2

2

.

Low-Latency AI Inference Infrastructure Powers OpenAI Partnership

Cerebras specializes in AI inference chips designed to process user queries and generate responses from trained AI models, a segment experiencing explosive growth alongside the adoption of AI agents 2

2

. The company's wafer-scale processor architecture enables faster AI processing while reducing communication delays that occur when multiple smaller chips are linked together 2

2

. Part of the planned AI data center capacity is expected to support OpenAI workloads under the companies' existing partnership 3

3

. Cerebras recently disclosed a multi-year deal with OpenAI valued at more than $20 billion, covering 750 megawatts of high-speed inference compute through at least 2028 3

3

.

Scaling Production and Challenging Nvidia's Market Position

To support its expansion, Cerebras announced an expanded manufacturing partnership with Flex to scale production of its CS-3 systems at facilities in Milpitas, California, with production capacity expected to increase approximately 7x through 2026 5

5

. The move positions Cerebras to challenge Nvidia, whose technology powers more than 90% of Europe's announced AI factory projects 1

1

. The AI infrastructure boom helped Cerebras raise $5.5 billion in its initial public offering in May, making it one of the 15 largest IPOs in Wall Street history and giving the company a market capitalization of roughly $40 billion 1

1

2

2

. The company also launched a multi-year partnership with AWS to bring fast inference to global distribution, combining AWS's Trainium 3 chips with Cerebras CS-3 systems 3

3

. Recent financial results showed revenue of $193.40 million, beating consensus estimates, with the company projecting fiscal-year revenue between $855 million and $865 million 3

3

. Feldman argued that concerns about an AI bubble are misplaced because customer demand continues to outpace available computing supply, with broader economic benefits of artificial intelligence only beginning to emerge 2

2

.