Bloom Energy stock surges 400% as AI data centers scramble for faster power solutions

2 Sources

Share

Bloom Energy has emerged from decades of controversy to become a major winner in the AI boom, with its stock surging over 400% in the past year. The company's fuel cells deliver on-site power to data centers in months rather than years, attracting every major hyperscaler. With record revenue of $2.02 billion and a $20 billion backlog, Bloom is betting on AI's insatiable electricity demands.

Bloom Energy Emerges as Unlikely Winner in AI Infrastructure Race

Bloom Energy has transformed from a niche energy player into a critical supplier for the AI boom, with its stock price surging more than 400% over the past year—outpacing even AI bellwethers like Nvidia

1

. The San Jose-based company, founded 25 years ago by CEO KR Sridhar, spent years pitching fuel cells to a market that largely wasn't listening. Now, the intense power demands of AI data centers have turned Bloom into what one analyst calls a company that received "a huge gift" from hyperscalers desperate for electricity

1

.

Source: Benzinga

Source: Benzinga

The company closed 2025 with record full-year revenue of $2.02 billion, representing a 37.3% increase over 2024, as AI data centers and manufacturing hubs moved toward energy independence

2

. Bloom's stock price surged late Thursday after its earnings beat investor expectations, and the company issued aggressive revenue guidance of $3.1 billion to $3.3 billion for 2026, well above analyst estimates

2

.

Speed to Power Drives Hyperscaler Interest

What makes Bloom Energy uniquely positioned for the AI infrastructure buildout is speed. The company manufactures shipping container-sized fuel cell units that generate power with natural gas through a chemical reaction instead of combustion

1

. Bloom can provide at least some level of on-site power in as little as a few months, compared to multiple years for electricity connected to power grids

1

.

"We went from zero hyperscalers directly talking to us to pretty much every hyperscaler talking to us," Sridhar said in an interview during last month's World Economic Forum in Davos

1

. During the first half of 2025, "pretty much our phones wouldn't stop ringing, and we had to be selective on who we were talking to," the physicist and former NASA materials scientist added

1

.

Source: Axios

Source: Axios

Michael Thomas, founder and CEO of Cleanview, a market intelligence platform, explained that hyperscalers are willing to pay the relatively higher cost of Bloom Energy's technologies because "all that matters is speed to power"

1

. Roughly one-third of all planned new power capacity for data centers now envision skipping the grid entirely by building power on site, up from virtually none a year ago

1

.

Bring Your Own Power Becomes Infrastructure Standard

The electricity demands of AI have fundamentally shifted how tech companies approach power infrastructure. "Bring Your Own Power has become the mantra for data centers and power-hungry factories," Sridhar stated on the company's earnings call, adding that "on-site power has moved from being a decision of last resort to a vital business necessity"

2

.

Bloom has signed major deals with key players across the AI and energy landscape over the last year and a half, including Oracle, utility AEP, and Brookfield, a leading financier and owner of data center assets

1

. The shift is evident in the company's staggering backlog growth, with product backlog surging 140% year over year to $6 billion, and total backlog reaching approximately $20 billion

2

.

800-Volt DC Technology Targets Expanding AI Infrastructure

Bloom is positioning itself as the primary power layer for expanding AI infrastructure with its native 800-volt DC offering. "AI is a huge tailwind for the power industry and a big catalyst for Bloom's growth," Sridhar said, noting that AI computer racks consume nearly "100 times more power" than traditional predecessors

2

.

Because digital equipment runs on DC, traditional AC-to-DC conversion in data centers creates waste and complexity. Sridhar argued that "800 volts DC will soon be the data center standard because physics requires it," and boasted that "Bloom, and only Bloom, natively produces 800 volts DC today. No band-aids, or adapters needed"

2

.

Each fuel cell is roughly book-sized and runs continuously, with hundreds stacked together inside the container-sized power units

1

. The fuel cells reduce smog-forming air pollution compared to combustion but still emit carbon when fueled by natural gas. Bloom's fuel cells can also use biogas or hydrogen—fuels that remain limited and expensive but that longer term could be cleaner

1

.

Risks Remain Despite Momentum

While Bloom Energy has delivered results that validate Sridhar's long-running narrative, challenges persist. Analysts at Jefferies raised its 2026 outlook for Bloom on stronger order visibility but remain cautious longer term, citing competition, supply risks and limited margin for error at current valuations

1

.

The company has had a controversial run since its founding 25 years ago. Bloom's 2018 IPO filing included a surprising disclosure involving investors with ties to a firm sanctioned for misleading investors, and a 2019 investigation flagged a history of unreliable projections

1

. Sridhar acknowledged that Bloom requires higher up-front costs, though he says its fuel costs are comparable and that upcoming shifts in how data centers use electricity will make the technology even more cost-competitive

1

.

CFO Maciej Kurzymski pointed to $271.6 million in adjusted EBITDA as proof of the "operating leverage there is in the business as we start to scale"

2

. For now, Sridhar dismisses the risk of an AI bubble, confident that "we will deliver our power platform faster than you can build your greenfield facilities"

1

2

.

Today's Top Stories

TheOutpost.ai

Your Daily Dose of Curated AI News

Don’t drown in AI news. We cut through the noise - filtering, ranking and summarizing the most important AI news, breakthroughs and research daily. Spend less time searching for the latest in AI and get straight to action.

© 2026 Triveous Technologies Private Limited
Instagram logo
LinkedIn logo