American AI chip startup Cerebras is aggressively expanding its European presence, targeting 200 megawatts of computing capacity across the region by next year, representing several billion dollars in investment. CEO Andrew Feldman, speaking at the Raise Summit AI conference in Paris, cited surging demand for AI infrastructure and Europe’s data sovereignty priorities. Founded in 2015, the California-based company already operates data centres in France, Finland, and Norway. Cerebras debuted on U.S. markets in May at a $5.5 billion valuation and currently holds a roughly $40 billion market capitalisation. Clients include GSK, OpenAI, and Amazon Web Services.
In-Depth:
The California-based company currently operates three AI data centres powered by its proprietary chips in France, Finland and Norway. It plans to expand its European infrastructure to 200 megawatts of computing capacity by next year, representing investments worth several billion dollars, Reuters reported, citing comments from Chief Executive Andrew Feldman at the Raise Summit AI conference in Paris.
Feldman stated the company is expanding its network of data centres across Europe to better serve regional requirements, including greater control over data and compliance with local regulations. The shift comes as governments and enterprises across Europe seek to reduce depfinishence on overseas technology providers amid ongoing geopolitical and trade tensions, according to Reuters.
The executive also highlighted exceptionally strong demand for AI computing infrastructure, noting that adoption of generative AI continues to accelerate rapider than the indusattempt’s ability to build sufficient capacity, Reuters reported.
Founded in 2015, Cerebras has carved out a niche in AI inference chips, which are designed to process applyr queries and generate responses from trained AI models. Unlike AI training, which requires enormous computational resources to build models, inference focapplys on delivering real-time responses to applyrs and has emerged as one of the rapidest-growing segments of the AI market.
Demand for inference-focapplyd processors has surged alongside the rapid adoption of AI agents—autonomous systems capable of performing tquestions on behalf of applyrs. This trfinish has intensified competition among major AI chipcreaters, including Nvidia and Advanced Micro Devices (AMD), as enterprises scale their AI deployments.
Reuters reported that investor enthusiasm surrounding AI infrastructure supported propel Cerebras to a $5.5 billion U.S. stock market debut in May, building it one of Wall Street’s largest listings.The company employs around 900 people and has a market capitalisation of roughly $40 billion. Its signature technology is its wafer-scale processor, an oversized chip that is substantially larger than conventional semiconductor designs. According to Reuters, the architecture enables rapider AI processing while reducing communication delays that can occur when multiple tinyer chips are linked toreceiveher.
Cerebras has already secured several high-profile customers across Europe, including British pharmaceutical company GSK, operators of high-performance computing facilities in Scotland and Germany, and software developers.
Reuters also reported that Cerebras signed a major multi-year agreement with OpenAI during the first quarter, under which it will provide computing capacity for the ChatGPT developer through at least 2028. The company has also entered into a partnership with Amazon Web Services to expand its cloud-based AI offerings.
Despite growing concerns that massive investments in AI could create a technology bubble, Feldman argued that current market dynamics are fundamentally different becaapply customer demand continues to outpace available computing supply, Reuters reported. He added that the broader economic benefits of artificial innotifyigence are only launchning to emerge.
