Fusion power got another boost on Monday as Google announced it will buy half the output of Commonwealth Fusion System’s first commercial power plant.
Commonwealth Fusion Systems (CFS) will send Google 200 megawatts of electricity from its Arc power plant, which is expected to come online in the early 2030s. Meanwhile, Google is sending CFS a check as part of a new, unannounced funding round.
The new round will be “comparable” to the previous one, CFS co-founder and CEO Bob Mumgaard said. CFS’s most recent funding round, a Series B that Google participated in and that brought in $1.8 billion, closed in 2021. The company has raised the most of any fusion startup.
“That’s a very strong demand signal,” said Mumgaard. “This investment allows us to do some of the R&D that will enable us to go into Arc faster.”
CFS is building a demonstration reactor, known as Sparc, just outside of Boston. That facility will be completed in 2026, according to Mumgaard. Arc, the company’s commercial power plant, will be built near Richmond, Virginia.
The new deal marks only the second time a major company has agreed to buy power from a fusion startup. The first was signed in 2023 when Microsoft agreed to buy the output of Helion’s first commercial power plant, which is scheduled to come online in 2028.
Like other hyperscalers, Google has been scouring the globe for new sources of electricity. AI and cloud services have ignited a surge in data center construction, driving a new wave of electricity demand along with it. One forecast expects that data center power demand will double by the end of the decade.
“To power all this, we know that we’re going to need to make big bets in this next frontier of energy innovation,” said Michael Terrell, Google’s head of advanced energy.
Google thinks of its energy investments across three time horizons, Terrell said. In the short term, the company has prioritized solar, wind, and batteries. A bit farther out, it’s betting on geothermal and small modular nuclear reactors, as evidenced by its investments in geothermal startup Fervo Energy and nuclear startup Kairos Power.
Fusion energy is a bit further out, and that “would certainly put this [CFS investment] in the long term category,” he said.
Google bought 8 gigawatts of renewable power in 2024, twice what it puchased in 2023. And while solar, wind, and batteries have been the main additions to Google’s portfolio in recent years, Terrell told TechCrunch that the company will need other power sources to allow its data centers to operate 24/7.
“There’s definitely a path with wind and solar and storage in regions where the resource is very strong,” he said, citing places like the Midwest, which has consistent winds, and the Southwest, which has mostly cloud-free skies. But other places, like the Southeastern U.S. and many counties in the Asia-Pacific region might be too cloudy or their power grids too fragmented to make traditional renewables work.
To compensate for those shortfalls, one approach is to overbuild wind and solar, but that can get expensive quickly.
Technologies like fusion “actually bring the cost down of achieving high penetrations of carbon free energy,” Terrell said. “If you have these clean, firm technologies — even if they’re more expensive on a per megawatt-hour basis — if you’re sort of folding those into the portfolio, it actually brings your overall portfolio costs down.”
Mumgaard is confident that CFS can deliver power to Google in less than a decade, and when that happens, he thinks that demand for fusion will skyrocket.
“It doesn’t depend on geography or weather, doesn’t depend on access to special materials. It’s something that you could run 24/7,” he said. “We expect that fusion can have a really big payoff because once it’s shown that you can do this and you have a first power plant up and running, you could scale it. You could build this around the world.”