Nuclear’s AI Boom Just Made Short Sellers $2bn

London, 19 August 2026 — EBM Newsdesk Analysis —Anthony Gill

Short sellers have made an estimated $2.1bn betting against three of the most prominent companies in America’s small nuclear reactor boom, as investors retreat from valuations built on future demand rather than current revenues.

NuScale Power, Nano Nuclear Energy and Sam Altman-backed Oklo have collectively lost approximately $30.3bn in market value since their shares peaked last October. The decline represents a sharp reversal for a sector that became one of Wall Street’s favourite ways to invest in the artificial intelligence revolution.

The underlying nuclear opportunity remains compelling. What has collapsed is the assumption that every company promising to supply that power deserves a multibillion-dollar valuation today.

AI created an irresistible investment story

Artificial intelligence requires vast quantities of electricity. Data centres need power around the clock, making intermittent wind and solar generation insufficient without substantial storage or backup capacity.

Small modular reactors appeared to offer an ideal solution. Their components could theoretically be manufactured in factories, transported to individual sites and assembled more quickly than conventional nuclear plants. Developers promised dependable, low-carbon power without the extreme construction costs associated with traditional reactors.

Government support strengthened the argument. The Trump administration made nuclear energy central to its plans for American energy security, while technology groups began signing agreements with reactor developers.

Investors responded by pushing early-stage nuclear companies to extraordinary valuations—even when they lacked operating reactors, regulatory approval or meaningful revenue.

The difficulty is that growing electricity demand does not automatically translate into near-term earnings for reactor developers.

Short sellers targeted the delivery gap

NuScale has achieved an important advantage by securing US regulatory approval for its reactor design. However, it has not yet delivered a commercially operating plant.

Oklo intends to build and operate reactors before selling electricity directly to customers. That could eventually create an attractive stream of recurring revenue, but the company still requires regulatory approval and has yet to produce power commercially.

Nano Nuclear recently reported its first revenue following an acquisition, but remains loss-making and years away from demonstrating its core reactor technology at scale.

The companies are therefore exposed to what might be called the nuclear timetable problem. Investors can identify a genuine demand emerging in the 2030s while still paying too much for businesses attempting to meet it in 2026.

Only two commercial small modular reactors are currently operating globally, in China and Russia. More than 80 designs are under development, but licensing, financing, fuel procurement and construction remain formidable obstacles.

Up to 30 per cent of Nano Nuclear’s shares have reportedly been on loan, illustrating the scale of bearish positioning against the company.

A correction, not nuclear’s funeral

The market’s reversal should not be interpreted as evidence that small reactors will fail.

TerraPower, founded by Bill Gates, received a construction permit in March for its Natrium project in Wyoming and is targeting commercial operation in 2031. The company has also reached a preliminary agreement with South Korea’s SK Innovation to collaborate on international projects.

Amazon, Meta and other technology companies continue to pursue nuclear-power agreements. Forecasts indicate that electricity demand from data centres could triple by 2035, leaving governments and utilities with an urgent requirement for dependable generation.

The more important distinction is between nuclear exposure and speculative nuclear exposure. Established engineering groups with operating businesses, manufacturing capacity and regulatory experience offer a different risk profile from pre-revenue developers dependent on repeated capital raising.

Investor appetite will soon face further tests as companies including Holtec International and Westinghouse consider public-market transactions.

The market has rediscovered time

The short sellers’ success does not disprove the case for nuclear energy. It demonstrates that even strategically important technology can become dangerously overvalued.

AI enthusiasm encouraged investors to compress a decade of regulatory and construction risk into present-day share prices. The companies were valued as though large reactor fleets were approaching operation rather than remaining on engineering drawings and licensing schedules.

Small reactors may eventually become essential infrastructure for the digital economy. But demand for electricity cannot accelerate nuclear approvals, manufacture specialist fuel or eliminate construction risk.

The $2.1bn gained by short sellers is ultimately the price investors paid for forgetting that a convincing long-term story is not the same thing as a functioning business.