Brookfield and Bloom Energy Join Forces with $5 Billion Plan to Power the AI Boom
Silicon Valley runs on electricity the way athletes run on adrenaline—and lately, both are in short supply. On October 13, a major move sent ripples through the tech and energy worlds. Brookfield, a global infrastructure heavyweight controlling over $550 billion in assets, teamed up with fuel cell innovator Bloom Energy in a $5 billion agreement to build massive “AI factories.” These aren’t ordinary data centers. They’re purpose-built hubs designed to feed the insatiable appetite of artificial intelligence.
This partnership isn’t just symbolic. It officially launches Brookfield’s new AI Infrastructure strategy, an ambitious attempt to merge power generation, computing capacity, and capital into a seamless ecosystem. Instead of relying on the aging electric grid, Bloom Energy will deploy modular solid-oxide fuel cells that generate power directly onsite. Imagine quiet power boxes converting natural gas into electricity right next to the servers—no waiting for grid approval, no transmission delays. KR Sridhar, Bloom’s CEO, cut right to the point: “AI infrastructure must be built like a factory—with purpose, speed, and scale.” Legacy power systems, he argued, simply can’t keep up.
The urgency is real. Analysts predict U.S. data center electricity demand could more than triple—from 35 gigawatts today to as much as 123 gigawatts by 2035. Generative AI models gulp power during training runs that spike without warning. Tech giants like Google have already signed billions in renewable power deals with Brookfield. They want campuses that can fire up gigawatts of energy at a moment’s notice. But the grid moves slowly. Interconnection queues stretch out for years. Transmission upgrades crawl. Builders can’t wait. They need power in months—not 2030.
That’s where onsite energy becomes a game-changer.
Sikander Rashid, who leads Brookfield’s AI Infrastructure division, calls onsite generation essential to “closing the grid gap.” Under the deal, Bloom becomes Brookfield’s preferred fuel cell provider, with up to $5 billion earmarked to deploy projects around the world. One major European site is expected to break ground by year-end—likely in a region where natural gas is abundant and permits are easier to snag. Brookfield is no stranger to digital infrastructure. The company holds massive stakes in Compass Datacenters, Duke Energy Florida, and other power-heavy assets. Bloom, meanwhile, already powers facilities for Equinix, Oracle, and American Electric Power. Both companies know the terrain—and the stakes.
Wall Street noticed. Bloom Energy’s stock exploded, jumping nearly 30 percent in a single day. Shares closed at $109.91 on October 13, up more than $23 from the previous session. At one point, the price skyrocketed from $88.82 to almost $118 intraday. Investors no longer saw Bloom as just another clean energy company. They saw the backbone of AI power.
Fuel cells offer big advantages. They’re modular, so builders can start small and expand as needed. They react quickly to AI’s ever-shifting loads. When paired with batteries, they boost reliability and avoid outages. Best of all, they can sidestep the painful delays of utility interconnections. One analyst said it best: “Months, not years.” Developers can build wherever pipelines exist. Over time, these systems may even blend hydrogen to reduce emissions.
But the path isn’t risk-free. Natural gas still emits CO2—cleaner than diesel, sure, but environmental scrutiny is rising. New methane rules and hourly carbon reporting could complicate operations. Gas price volatility could pinch margins. Permitting in dense European markets may trigger local opposition. And competitors are circling. Gas turbines, small modular nuclear reactors, long-term power purchase agreements—everyone is racing to supply firm, on-demand electricity to AI campuses.
In this race, the real currency isn’t megawatts—it’s time.
AI companies will likely adopt hybrid energy strategies: use fuel cells for reliable baseload power, batteries for spikes, and keep the grid as backup. Utilities could lose revenue if customers go off-grid, so they may counter with fast-track energy tariffs or joint microgrid partnerships. Policy could swing the playing field. U.S. hydrogen tax credits or EU sustainability rules might supercharge—or stall—projects. One industry observer summed it up nicely: “The real product is calendar time.” Whoever delivers power before servers arrive wins.
Predictions are already taking shape. By mid-2026, expect at least two more major infrastructure firms to launch AI energy platforms with built-in transitions from natural gas to hydrogen. New contracts will likely include fuel-indexed pricing and hourly-matched low-carbon guarantees. Bloom’s future success will hinge on this Brookfield pipeline—watch for service contracts and hydrogen trials by 2027.
In the end, this partnership isn’t just another deal. It’s a bold attempt to solve one of the most pressing challenges of the AI era: how do you power intelligence at planetary scale when the grid can’t keep up? Brookfield and Bloom believe the answer lies in building power like factories—fast, flexible, and fiercely independent.
If they succeed, the next wave of AI will rise on foundations they built. If they stumble, the grid’s limitations will cast a long shadow over innovation.
For now, the lights are still on—and the race to power the future has officially begun.
NOT INVESTMENT ADVICE