CFS’s Billion-Dollar Play for the Grid

Last week, Axios reported that Commonwealth Fusion Systems (CFS) is raising a massive Series B2 extension round, targeting $1-1.5 billion to finance its first commercial fusion power plant. If successful, the raise could push the company’s valuation north of $8 billion. Interestingly, the round is anchored by a yet-unnamed “hyperscale data center developer” (more thoughts on who that might be later).

It’s a stark contrast to our story last week about General Fusion’s urgent cash crunch and deep layoffs. While CFS is attracting growth-stage capital to break ground on its ARC plant, General Fusion is seeking rescue financing just to stay afloat. The juxtaposition reveals deeper signals about where fusion investors are putting their money, and why.

From Concept to Construction

CFS, the MIT spinout developing a high-field tokamak, is no stranger to headline-grabbing raises. But this latest round is different: it’s explicitly intended to fund ARC, a 400 MW net-electric power plant slated for early 2030s deployment. Unlike prior raises that funded R&D, this is infrastructure capital intended to bring fusion power to the grid.

The company’s credibility rests on a strong track record. In 2021, CFS demonstrated the world’s strongest high-temperature superconducting magnet, a breakthrough that enabled its compact tokamak design and catalyzed its previous funding. Since then, it has invested more than $2 billion into its SPARC prototype, with first plasma expected in 2026 and net energy shortly after.

The Bifurcation of Fusion Investment

The contrast between CFS and General Fusion is not just about execution; it’s about how investor priorities are evolving.

In 2021’s fusion boom, capital flowed broadly: tokamaks, stellarators, field-reversed configurations, Z-pinches, and General Fusion’s magnetized target approach all secured meaningful funding. Now, that breadth is narrowing.

Capital is concentrating around approaches seen as either more technically validated or more capital-efficient. High-field tokamaks like CFS’s benefit from decades of tokamak science and recent magnet breakthroughs, giving late-stage investors confidence that a power-producing device is within reach. 

On the other end, some venture investors are gravitating to novel concepts that promise smaller, faster iteration cycles. For example, magnetic mirror fusion startup Realta just raised $36M, with backers touting that its simpler, modular 50 MW design could lower deployment costs compared to huge tokamaks. 

Similarly, other U.S. startups like Helion and Zap Energy are pursuing compact, pulsed fusion systems aiming to deliver net energy on accelerated timelines (Helion notably inked a deal to supply electricity by 2028). These “sooner, smaller” bets often require less upfront capital per step, which is appealing in a tighter funding climate.

But the middle is thinning. Technologies that require hundreds of millions without clear short-term results, like General Fusion’s approach, are falling out of favor. General Fusion’s magnetized target system is still unproven at scale, and despite years of work, the company hasn’t demonstrated net energy or a clear path to commercialization. In a tighter funding climate, that’s proving fatal.

Who might the “Hyperscale Data Center Developer” be?

This is just an educated guess, of course, but my money’s on Microsoft. For a handful of reasons:

• In 2023, Microsoft signed a power purchase agreement with Helion Energy, the first such agreement with a fusion startup.

• Microsoft currently provides cloud services to CFS.

• Microsoft founder Bill Gates previously invested in CFS’s Series B round.

• Microsoft operates one of the world’s largest cloud infrastructures (Azure) and has set ambitious CSR targets, aiming to be 100% “carbon negative” by 2030.

We’ll see how well that bet ages.