The small modular reactor sector experienced a dramatic reversal in 2026, with short sellers profiting an estimated $2.1 billion from betting against three major companies as the market wiped $30.3 billion off their collective value. NuScale Power, Nano Nuclear, and Oklo, which surged last year on investor enthusiasm about powering artificial intelligence data centers with nuclear energy, have now become cautionary tales about speculative bubbles in emerging technology.

Why Did Small Modular Reactor Stocks Soar and Then Crash?

The sector's meteoric rise and subsequent collapse followed a textbook pattern. Throughout 2025, these companies attracted massive investor interest as tech giants like Meta, Amazon, and others sought new power sources to fuel their energy-hungry artificial intelligence operations. Regulatory support from the Trump administration, which pledged tens of billions in loans and promised to cut red tape for nuclear projects, further fueled optimism. The Department of Energy announced $17.5 billion in loans to rebuild the U.S. nuclear supply chain, signaling serious government backing.

However, the three companies at the center of the collapse share a critical characteristic: they are lossmaking with little or no revenue. NuScale reported a $96.7 million loss in the first half of 2026 alone and now faces a shareholder class-action lawsuit alleging it misled investors. Nano Nuclear had reported zero revenue until acquiring a nuclear logistics company in the second quarter of 2026, then posted an operating loss of $15.8 million for that quarter. Oklo, backed by Sam Altman, has yet to secure full regulatory approval to build and operate its reactors.

"The stocks were overinflated in price, based on speculation. The sector went through a textbook hype cycle last year. It is very typical of a company that is in this early pre-consistent revenue phase," said Adam Stein, director of nuclear energy innovation at the Breakthrough Institute, a climate and energy think-tank.

Adam Stein, Director of Nuclear Energy Innovation at the Breakthrough Institute

How Are Short Sellers Profiting From the Collapse?

Short sellers, who profit when stock prices fall, have positioned themselves aggressively against these companies. According to data from S3 Partners, funds made an estimated $2.1 billion shorting NuScale, Nano Nuclear, and Oklo over the past year. The scale of these bets reflects investor conviction that the companies' valuations had become detached from reality.

The short positions remain substantial. Around 18 percent of Oklo and NuScale's outstanding shares remain out on loan, a proxy for short selling activity, while almost 30 percent of Nano Nuclear's shares are on loan. X-energy, another reactor developer backed by Amazon and Ken Griffin, has shed $5.8 billion in market value since its April initial public offering, with short sellers earning an estimated $67 million from bets against the company since mid-May.

What's the Core Problem With Small Modular Reactors?

Small modular reactors, or SMRs, are designed to be assembled from factory-built modules to save time and money, producing around 300 megawatts or less compared with more than 1,000 megawatts for traditional reactors. The technology promises flexibility and lower upfront costs, making it theoretically ideal for powering distributed data centers. However, the reality is far more complicated.

Only two commercial SMRs are currently operational worldwide, located in Russia and China. More than 80 designs exist in various stages of development, but the path from design to operation involves significant regulatory hurdles, capital requirements, and timeline uncertainties. The earliest some reactors will come online is mid-to-late 2028 if manufacturers can accelerate delivery while satisfying regulators, though the majority are expected during the 2030s. Additionally, analysts at BNP Paribas have raised concerns about potential shortages of high-assay low-enriched uranium, a specialized nuclear fuel vital for SMRs.

Steps to Understanding the Nuclear-AI Investment Landscape

• Revenue Reality Check: Evaluate whether companies have demonstrated actual revenue generation or are purely pre-revenue. NuScale, Nano Nuclear, and Oklo had minimal to zero revenue despite multi-billion-dollar valuations, a red flag that should have tempered investor enthusiasm earlier.
• Timeline Verification: Assess realistic deployment timelines against investor expectations. Most SMRs won't deliver power until the 2030s, yet markets priced in near-term revenue contributions, creating a fundamental mismatch between expectations and physics.
• Capital Requirements Analysis: Consider the enormous capital expenditure needed to build nuclear infrastructure. These companies require tens of billions in funding to move from design to operation, raising questions about whether venture capital and equity markets alone can sustain them.
• Regulatory Approval Status: Distinguish between companies with full regulatory approval and those still seeking it. Oklo and others lack complete licenses from the U.S. Nuclear Regulatory Commission, meaning their timelines remain speculative.

Investor sentiment has shifted dramatically. Christian Putz, founder and chief executive of investment firm ARR Investment Partners, who previously shorted Oklo but has since unwound that position, observed the changing market dynamics. He noted that while government support and the AI demand story drove bullish sentiment in 2025, the reality of near-zero revenue and massive capital requirements has become impossible to ignore.

"The sentiment has changed this year, people are far more critical. These companies have almost zero revenue for the foreseeable future and, on top of that, there are very high capex requirements," said Christian Putz, founder and chief executive of ARR Investment Partners.

Christian Putz, Founder and Chief Executive of ARR Investment Partners

What Does This Mean for AI's Power Future?

The collapse of SMR stock valuations does not eliminate the underlying energy crisis facing artificial intelligence. U.S. data center power demand is projected to climb from 34.7 gigawatts in 2024 to 106 gigawatts by 2035, according to BloombergNEF data. This explosive growth remains real and urgent.

What has changed is the market's willingness to fund speculative nuclear ventures at premium valuations. Big Tech companies like Meta have made strategic bets on SMR technology, striking deals with Oklo and TerraPower to support reactor development. However, these partnerships now operate in a far more skeptical investment environment. The sector still has government backing, with the Trump administration vocally supporting nuclear expansion and pledging tens of billions in loans, but private capital has become far more cautious.

Siegfried Eggert, chief executive of activist short seller Grizzly Research, characterized the situation bluntly: "What we have seen in 2025 seems to me like an industry bubble that is already deflating. I believe most knowledgeable investors understood for a while that the valuations seemed rather extended given the timeline of this industry".

As Holtec International and Westinghouse prepare to list their SMR divisions in the coming weeks, the market will face a critical test of investor appetite. Whether these established industrial companies can restore confidence in the sector, or whether the nuclear-AI narrative has permanently lost its luster, remains to be seen. What is certain is that the era of uncritical enthusiasm for pre-revenue reactor companies has ended.