The pre-revenue developer drew a going-concern warning from its auditors even as it courts data-center demand for its borehole-based power design.

June 22, 2026

Deep Fission, Inc. has priced its initial public offering at $16.00 per share, selling 2.5 million shares of common stock to raise about $40 million before expenses as it pushes to commercialize an unconventional nuclear reactor buried roughly a mile underground. The Berkeley, California company expects its stock to begin trading on the Nasdaq Global Market under the ticker FISN, its first listed public offering after years as a private development-stage business.

After underwriting discounts and estimated expenses, the company expects net proceeds of approximately $32.9 million, or about $38.5 million if underwriters exercise an option to buy up to 375,000 additional shares within 30 days to cover over-allotments. The deal is being run on a firm commitment basis by Benchmark, a StoneX company, alongside Seaport Global Securities and Maxim Group. A group of existing shareholders the company calls cornerstone investors signaled interest in buying up to $10 million of stock on the same terms as other purchasers, though that interest is not binding. Shares are expected to settle on or about June 22, 2026.

A reactor built for the borehole

Deep Fission is developing what it calls the Gravity Reactor, a small modular reactor based on established pressurized water reactor technology but placed inside a deep, cased vertical borehole rather than a conventional surface containment building. The company argues that emplacing the reactor about one mile down lets the surrounding geology and a tall column of water do much of the work that steel and reinforced concrete handle at traditional plants. The water column is expected to generate roughly 160 atmospheres of hydrostatic pressure, supporting reactor operating pressure and cooling, while the rock provides shielding and confinement.

The pitch is that this approach could cut capital and operating costs, speed up deployment, and reduce exposure to surface hazards such as extreme weather, seismic events, and sabotage. The company is targeting later single-reactor installations that take as little as about six months from the start of construction, far faster than the multi-year timelines it associates with conventional plants and above-ground modular designs.

The reactors are designed to use standard low-enriched uranium and conventional fuel assemblies, which the company frames as a way to avoid the supply-chain and licensing risk tied to specialty fuels used by some advanced reactor rivals. An initial four-assembly configuration is targeted to produce up to about 8 megawatts of electric power, with a later nine-assembly version aimed at up to 15 megawatts and stronger per-reactor economics. Internal estimates put per-reactor deployment costs falling from roughly $152 million for the early design to about $84 million for the larger one, though the company stresses these are conceptual figures built on assumptions that may not hold.

Demonstration ahead of commercial licensing

Deep Fission is advancing on two regulatory tracks. It was selected in August 2025 as one of eleven projects across ten companies eligible for the U.S. Department of Energy Reactor Pilot Program, and in November 2025 it signed an Other Transaction Agreement under which the DOE would review the safety basis for a demonstration reactor and approve testing milestones. That authorization does not permit commercial operation and does not replace the need for a license from the Nuclear Regulatory Commission. The company intends to apply for an NRC commercial license in the first half of 2027 and is pursuing the two processes in parallel rather than in sequence.

Field work is centered on a long-term lease covering about 100 acres at the Great Plains Industrial Park in Parsons, Kansas. The company has drilled a first data acquisition well to gather subsurface data down to 6,000 feet and plans to demonstrate a commercial-scale borehole and a prototype reactor before moving toward clustered, higher-volume deployments. The DOE program provides authorization and oversight but no direct funding, so Deep Fission must finance the demonstration work itself.

The company is positioning the technology for customers with large, continuous power needs, led by hyperscale data centers and artificial intelligence workloads, and extending to industrial users, utilities, and government and defense sites. It has lined up early relationships, including a memorandum of understanding with Blue Owl Digital Infrastructure Advisors that came with a $20 million equity investment, plus arrangements with Halliburton, Urenco USA, Day & Zimmermann, R-V Industries, and Sparx Engineering. Most of these are non-binding or preliminary.

Losses, a restatement, and a capital gap

The prospectus is candid about the financial picture. Deep Fission has no revenue and remains in the development stage. It reported a net loss of $21.3 million for the first quarter of 2026, $61.0 million for 2025, and $5.7 million for 2024, with an accumulated deficit of $88.1 million as of March 31, 2026. The 2025 loss was driven largely by a roughly $39.8 million charge tied to the conversion of earlier SAFE Notes.

The company’s auditors included a going-concern paragraph, citing recurring losses and the expectation of continued operating losses. Cash stood at $84.8 million at the end of March but had fallen to about $64.3 million by May 31, 2026, an amount management said would not have been enough to fund operations for the next twelve months without this offering. Even with the IPO, the company estimates it needs:

• about $138 million in additional capital to develop, license, and commercially deploy a first commercial reactor; and
• roughly $67 million to build and begin operating an initial test reactor.

Deep Fission also disclosed that it recently restated prior financial statements to correct errors involving SAFE Note valuations and stock-based compensation, and that it identified material weaknesses in internal control over financial reporting. Management concluded that disclosure controls were not effective as of March 31, 2026 and said it is working to remediate the problems. Separately, the company is a plaintiff in litigation challenging aspects of the NRC’s licensing framework, which could affect its regulatory path.

The company traces its public-market structure to a September 2025 merger in which a former shell company, Surfside Acquisition, combined with the private Legacy Deep Fission and took on its name. After the offering, about 60.3 million shares will be outstanding. New investors buying at $16.00 face immediate dilution of $14.08 per share, since the as-adjusted net tangible book value works out to roughly $1.92 a share. Deep Fission does not plan to pay dividends and intends to direct proceeds toward engineering, research and development, licensing, and construction of its first pilot reactor.