Analysis
Private investors are still pouring capital into nuclear-power startups even as the sector's publicly traded names have taken a bearish turn, according to a Thursday Crunchbase News report. The split is recent. Crunchbase says recent quarters were also strong for nuclear IPOs, but optimism about IPO prospects has taken a hit as public-market enthusiasm weakened, even as funding to nuclear startups hit an all-time high on expectations of surging AI-driven energy demand.
## Why Private And Public Have Diverged Venture investors backing both fission and fusion startups are underwriting a multi-year thesis: that AI data centers' power demand will eventually need firm, carbon-free baseload generation that wind and solar alone can't provide, and that whoever gets a next-generation reactor design licensed and built first captures outsized value. On the public side, at least three nuclear companies went public at initial valuations above $1 billion in the past six months, but most now trade far below their prior highs. X-energy hit a $12 billion valuation in initial trading after its April debut and has since shed about half its value, and Oklo is down about two-thirds from a high about a year ago. Crunchbase says it's unclear whether those setbacks reflect diminished optimism about the feasibility of nuclear startups' ambitions or a sense that valuations warranted a cutback, and U.S. public backlash against massive data center buildouts has also contributed to investor caution.
“Both are part of more than $6 billion invested in fission and fusion companies so far in 2026.”
## The Numbers Investors Are Betting Against A single hyperscale AI data center campus can require several hundred megawatts to a gigawatt of continuous power -- the kind of load that's pushed Microsoft, Google and Amazon to all sign nuclear-power agreements of their own since 2024, including Amazon's capital investment in X-energy, one of the nuclear IPOs Crunchbase describes. That corporate demand is the demand-side case private investors are underwriting; whether it's big enough to justify today's startup valuations is a separate question from whether AI needs more power at all.
## The Counterweight Nuclear projects have a long history of running years behind schedule and well over budget -- Georgia's Vogtle reactors are the most visible recent example -- and a public market pricing in that execution risk today isn't necessarily wrong just because venture capital is pricing a longer-dated bet differently. The two markets can both be right, over different time horizons.
## The Clock That Matters Crunchbase's report did name the biggest deals: Commonwealth Fusion Systems secured $1 billion in a July equity financing, and Valar Atomics picked up $1 billion across two Series B equity tranches. Both are part of more than $6 billion invested in fission and fusion companies so far in 2026. The signal to watch is whether private rounds keep coming at that scale while newly public nuclear names trade well below their peaks.