Bloom Energy and Nebius will deploy fuel-cell systems at its U.S. data centers.

For years, the AI race has been framed around chips. Nvidia became the face of that boom. Startups and tech giants rushed to secure GPUs. Investors poured billions into AI infrastructure. Now another bottleneck is moving to the center of the conversation: electricity. Bloom Energy announced Wednesday that European AI cloud provider Nebius will deploy Bloom’s fuel-cell systems at its U.S. data centers as part of a deal that could generate up to $2.6 billion in service fees over the life of the agreement. The partnership comes at a moment when AI companies are scrambling to secure enough electricity to keep new data centers online. Massive AI workloads are pushing utilities and power grids to their limits, forcing cloud providers to search for faster ways to bring energy capacity online. Under the agreement, Bloom Energy will install, operate, and maintain the fuel-cell systems, and Nebius will purchase the electricity generated onsite. The rollout is expected to happen in three phases across 10-year terms, delivering roughly 250 megawatts of guaranteed power capacity and 328 megawatts of installed capacity, according to a filing with the SEC. Investors reacted quickly. Bloom shares rose in premarket trading after the announcement, and Nebius shares jumped more than 7%, CNBC reported. The AI boom is turning electricity into the next tech battleground “Power remains a key constraint for AI infrastructure build-outs,” Nebius Chief Product and Infrastructure Officer Andrey Korolenko said in a statement. “We chose Bloom because their fuel cells solve that directly: Clean power with virtually no pollutants is deployed onsite, on the timelines our customers need, with the availability AI workloads require.” “We expect to put this technology to work alongside our infrastructure as we continue to scale our capacity,” he added. The news comes less than a year after Brookfield invested $5 billion in Bloom Energy to help supply electricity for AI data centers using fuel-cell technology. Brookfield, one of the world’s largest infrastructure investors, brought financing muscle and large-scale project experience to the partnership. Bloom is no stranger to the data center market. Its fuel-cell systems already support digital infrastructure operated by companies including American Electric Power, Equinix, and Oracle. The Brookfield agreement marked a much larger push into AI infrastructure, positioning Bloom as one of the companies working to meet the growing energy demands of AI computing. The latest deal highlights how energy access is becoming one of the defining issues in the AI economy. Data centers built for generative AI require enormous amounts of electricity, and many projects are facing delays tied to grid connections, permitting, and energy shortages. AI cloud startup Nebius turns to Bloom Energy as power shortages threaten data center growth Nebius has been moving aggressively to position itself as one of Europe’s major AI compute providers. Earlier this year, the company secured a $2 billion investment from Nvidia and signed a reported $27 billion AI infrastructure agreement with Meta. The company recently unveiled plans for what it says will become Europe’s largest AI data center in Finland, a facility expected to reach 310 megawatts of capacity and begin serving customers in 2027. Europe’s AI ambitions face a tougher energy picture than the United States. Electricity costs remain higher across much of the region, and many large infrastructure projects are facing delays due to aging grids and limited power availability. That reality is creating new opportunities for companies like Bloom Energy. Fuel cells are no longer being pitched as backup systems or sustainability projects. They are increasingly being marketed as a direct answer to one of AI’s biggest problems: getting enough electricity online before demand outruns the grid.

Physicists at the University of Oxford show they can assemble Schrödinger’s cat-like superpositions from far more exotic quantum ingredients.

Selection marks the next phase of construction on Micron’s New York project, expected to generate 50,000 jobs.

OneAdvanced has launched what it bills as the UK’s first private, sovereign healthcare large language model trained on NHS primary-care data.