All-In Summit: Nuclear fusion and the potential for energy abundance

Helion announced Microsoft as its first customer, under a power-purchase agreement for a 50-MW fusion plant targeted for 2028. Kirtley said that output is roughly enough for 40,000 homes, giving Helion just five years from prototype development to promised grid electricity.

Helion’s core economic bet is 95% electrical-energy recovery, which it says can bypass steam turbines and cooling towers and drive power toward 1¢/kWh. Kirtley said a pickup truck’s worth of fuel could run a 50-MW plant for a decade, making capital cost the dominant expense.

Commonwealth Fusion Systems is betting that new superconducting magnets can shrink proven tokamak physics enough to commercialize it. Mumgaard said the magnets made equivalent machines 10× smaller, while SPARC was halfway built and designed for about 100 MW of heat, potentially approaching 10× fusion gain.

Fusion has already become a multibillion-dollar private-capital race rather than primarily a laboratory program. Commonwealth had raised more than $2 billion, Helion about $2.2 billion, and Sam Altman personally invested $375 million in Helion’s recent $500 million round.

The founders openly identified different failure points: Commonwealth sees Helion’s plasma physics as its central uncertainty, while Helion worries about preserving extremely high energy-recovery efficiency. Kirtley said even a five-point efficiency loss would require more fusion, making Helion’s machines larger and more expensive.

A U.S. regulatory decision may remove one of fusion’s biggest commercialization barriers. Mumgaard said the NRC chose to regulate fusion like particle accelerators rather than fission plants, cutting estimated regulatory overhead from roughly $1 billion to $10 million; Helion said its existing state permits take six to nine months.

Helion’s bottlenecks are industrial as well as scientific. Kirtley said 90% of the capacitors for its current machine were purchased overseas because internal manufacturing could not scale quickly enough, even as Helion began rebuilding U.S. capacitor production for its next-generation machine.

Both founders rejected using future fusion as a reason to slow today’s clean-energy buildout. Mumgaard said another decade at roughly 50 gigatons of annual carbon emissions is too much, while Kirtley argued solar, wind and other zero-carbon sources still need aggressive deployment because future energy demand may exceed current forecasts.