Deep Fission ($FISN): a Money Pit

Please refer to our disclaimer at the bottom of the report.

Main findings

• Nasdaq-listed Deep Fission (FISN) is developing a small modular reactor (SMR) with a capacity of 5MWe-15MWe. The company plans to bury a miniature version of the widely used pressurized water reactor (PWR) one mile deep, inside a 40-60-inch water-filled borehole drilled into hard rock. 

• Like other SMR developers, FISN pitches itself as an AI data-center power solution. But data centers have turned to natural gas to accelerate deployment. Almost all of the roughly 75GW of booked data center on-site generation is natural gas-based. FISN has no regulatory approval and remains years away from commercialization.

• Before its June 2026 IPO, FISN guided to a levelized cost of energy (LCOE) of just $50-70/MWh. Then the company dropped the reference. Using figures disclosed in its June 2026 prospectus, we calculate an LCOE of $289/MWh for the first-of-a-kind (FOAK) reactor and $148/MWh for the nth-of-a-kind (NOAK) reactor, about 200% above prior guidance. This is not competitive with behind-the-meter gas solutions.

• Even this NOAK estimate is too optimistic, as capital cost is understated. Using figures from independent third parties, we estimate NOAK capital cost to be $112mn, 34% above FISN’s estimate, raising our estimated NOAK LCOE to $170/MWh. 

• FISN claims it can secure Nuclear Regulatory Commission (NRC) approval in less than a year. Yet it has already struggled with the less demanding Department of Energy (DOE) process, missing the program deadline while at least four peers have already completed it. Even if NRC approved, commercialization is far from assured: NuScale, which has a much more conventional design, has yet to sell a reactor three years after receiving its NRC license.

• FISN targets deploying its reactor in just six months versus 3-4 years for other advanced SMRs, citing fewer safety-related systems. Yet its own filing states a two-year development timeline. Peers pursue the same simplification strategy, but FISN adds a major step: drilling a one-mile borehole.

• FISN touts the industry’s largest customer pipeline. Most counterparties are undisclosed. Of those identified, one is an undeveloped industrial park with doubtful demand for a nuclear reactor. The other two (Endeavour Energy and Blue Owl) have signed only non-binding agreements. 

• FISN has less than twelve months of cash runway and faces ~$350mn of unfunded capital needs, potentially implying ~43% dilution to its existing shareholders. At the same time, the competition for SMR funding is intensifying: publicly traded SMR names have risen from three before October 2025 to nine today, with three more coming to the market, including a far more established player, Westinghouse.

• FISN is largely a repackaging of the founders’ prior venture, Deep Isolation, a deep-borehole nuclear waste disposal company that has failed to commercialize its concept after a decade.

Background and bull case

Deep Fission (NASDAQ: FISN) is a California-headquartered company founded in 2023. The company is developing a nuclear reactor called Gravity. The reactor is intended to be a miniature version of the conventional pressurized water reactor (PWR), using low-enriched uranium (LEU) as fuel. The company’s key innovation is that its reactor would be buried in a 40-60-inch water-filled borehole one mile (~1.6 km) underground. The reactor is designed with a power generation capacity of 5–15MWe.

Source: Deep Fission

Management claims that its reactor is cheaper to build because its innovative borehole design eliminates certain components required in above-ground PWR reactors. For example, the depth and surrounding geology provide natural containment, potentially eliminating the need for the reinforced concrete and steel containment structures typical of above-ground reactors.

FISN was founded by Richard Muller, a UC Berkeley physics professor, and his daughter Elizabeth, who serve as CTO and CEO of the company, respectively. The company became an unquoted public company in September 2025 through a reverse merger with shell company Surfside Acquisition, before listing on Nasdaq in June 2026.

FISN is participating in the Department of Energy’s (DOE) pilot reactor program. The company is seeking DOE authorization to build a pilot reactor at a site in Parsons, Kansas, by 1H 2027. Then, FISN plans to submit license applications to the Nuclear Regulatory Commission (NRC) to convert the pilot reactor into a commercial facility by the end of 2027, with high-volume deployment targeted to begin in 2028. Because its reactor is based on well-established PWR technology, its licensing pathway would be simpler than for novel advanced reactor designs. 

The company claims that its reactor can then be commercially deployed in as little as six months, compared with three to four years for other small modular reactors (SMR). 

FISN expects the surge in power demand from AI data centers will drive demand for its reactors and claims the largest pipeline of potential customers, representing 18.5GW of prospective capacity.

It’s time for small modular reactor supporters to admit it: natural gas, not SMRs, is the undisputed winner of the AI data center build-out

In early 2024, the market realized that AI would require enormous amounts of power. This kicked off a frenzy across nearly every corner of the power-generation trade, spanning conventional nuclear, gas-turbine manufacturers and the speculative (and often sketchy) SMR companies. Every player, including those with the most exotic concepts, was supposed to win.

Fast forward two years, and the verdict is in: gas-fired generation has emerged as the clear preferred solution for powering AI infrastructure. Data centers have booked roughly 75GW of onsite generation capacity so far, with the overwhelming majority relying on natural gas.

Source: Semianalysis

While the largest and most efficient combined-cycle gas turbines face multi-year lead times, data centers have turned to smaller engines or aeroderivatives with lead times as short as twelve months. Gas generation is a proven technology and can meet AI power demand quickly, with the added advantage of cheap natural gas in the US. SMRs, on the other hand, sit at exactly the opposite end of the spectrum: they remain unproven, with commercialization not expected before the 2030s. 

The political environment has also become more favorable to fossil-fuel-based generation, removing one of the key roadblocks to new gas development.

The company’s own numbers show its reactors are not cost-competitive

Even FISN’s own numbers reveal the poor economics of its reactors. Before its June 2026 IPO, the company claimed that its reactor could achieve a levelized cost of energy (LCOE) of only $50-70/MWh, including in an investor presentation prepared in connection with the February 2026 private placement. The number excluded non-engineering, procurement, and construction (EPC) costs and contingencies.

Source: Deep Fission Jan 2026 Investor Presentation (p. 14)

FISN initially claimed it could reach this ultra-low LCOE with its first-of-a-kind (FOAK) reactor. The company subsequently changed its tune: the same LCOE could be achieved only with nth-of-a-kind reactors (NOAK) (p. 26). 

Then, when the company was listed, the reference to the LCOE was dropped. The prospectus filed in connection with FISN’s June 2026 Nasdaq debut no longer mentions an LCOE target, nor do the company’s subsequent investor presentations, news releases and filings. Yet the company itself recognizes that LCOE is a key driver of its success (p. 38):

“If the levelized cost of electricity generated by our Gravity Reactor is not competitive with other sources of electricity generation, or if alternative technologies become more cost-effective or widely adopted, commercialization and demand for our reactors and related services could be reduced or fail to materialize.”

Another claim made in the January 2026 presentation (p. 17) but absent after the IPO was the assertion that its reactors would achieve 70%-80% cost reduction versus newly built conventional PWRs. The company probably realized that there was no basis for such a claim. It seems that material walkbacks like those mentioned above are fairly common for the company. 

Source: Deep Fission Jan 2026 Investor Presentation (p. 17)

Source: Deep Fission Aug 2026 Investor Presentation

If we simply apply the financial assumptions provided by the company, there is no way FISN can achieve such a low LCOE, whether for a FOAK or a NOAK reactor. We calculated the LCOE using FISN’s own disclosed assumptions (p. 11). We estimate an LCOE of $289/MWh for FOAK reactors, declining to $148/MWh for NOAK. This is about 200% higher than FISN’s initial guidance. We conservatively applied Lazard’s 7.7% discount rate, despite FISN’s substantially higher risk profile. See Exhibit A for methodology. We strongly suspect a lawyer asked the company to drop the LCOE reference during the IPO process.

Source: Deep Fission (p. 11), Lazard

Bloomberg estimates that gas-based generation, including engines and turbines used in behind-the-meter (BTM) data centers, has an LCOE of $103-110/MWh, significantly below our  estimated $148–289/MWh for FISN. Even the most expensive fuel cells have an LCOE of $140/MWh, still below FISN. Our own calculation using publicly available information matches Bloomberg’s estimate. FISN’s reactors would therefore remain materially more expensive than even the costly BTM alternatives currently favored by data center developers. 

In any case, FISN does not seem to be pursuing BTM at all. Its first reactor in Kansas is still expected to connect to the grid, not directly to an end customer. In fact, the company has cited (17:00) grid interconnection constraints as the reason for reducing the capacity of its initial reactor to just 5MW:

“We’re looking at 5MWe, in large part because that’s what we can connect to the grid quickly without needing to go through an extensive process for grid interconnection.”

Note that the company projections include tax credits according to its prospectus: “Our business and financial model contemplates, in part, the potential monetization of available tax credits and other government incentives, including investment tax credits that may be available for qualifying clean energy projects.”

Canaccord Genuity conjured a NOAK LCOE estimate of $36/MWh in its 15 July 2026 initiation report, even lower than FISN’s pre-IPO LCOE. That would make FISN’s reactor one of the world’s cheapest energy sources. The problem is that Canaccord misapplied the LCOE formula: it discounted the numerator (lifetime cost) but not the denominator (lifetime electricity generation). Over forty years, that error materially understates the LCOE. Properly discounting the electricity generation using Canaccord’s own 11% discount rate yields an LCOE of $160/MWh, making Canaccord more bearish than us. 

Cost reduction between FOAK and NOAK is extremely aggressive

The actual economics are even worse because FISN appears to materially understate its NOAK capital cost. As mentioned, the NOAK LCOE of $148/MWh, while still uncompetitive, represents a substantial improvement over the FOAK case. This improvement is driven by an assumed 45% reduction in capital costs from $152mn to $84mn.

At first glance, this cost reduction seems reasonable. The Tennessee Valley Authority (TVA) estimates (p. 172) that PWR capital costs decline by 49% from FOAK to NOAK. An Idaho National Laboratory (INL) study (p. 70) more conservatively estimates a ~34% decline. TVA’s estimate is particularly relevant because it comes from a prospective utility customer rather than a reactor vendor, reducing the incentive to rely on aggressive assumptions to promote the technology. TVA also has direct experience evaluating potential SMR deployments at its Clinch River Nuclear Site. 

However, TVA assumes no increase in unit size between FOAK and NOAK, consistent with the conventional approach. FISN, by contrast, increases reactor capacity substantially by ~88%, from 8MWe for FOAK to 15MWe for NOAK. The nuclear industry typically applies a 0.6 scaling exponent, which reflects that cost rises less than proportionally with capacity (see Exhibit B). Applying this factor, a FOAK reactor scaled to 15MWe would cost $222mn. Using TVA’s 49% cost reduction from FOAK to NOAK, we estimate FISN’s NOAK at $112mn, 34% higher than FISN’s NOAK estimate. 

Using this revised capital cost, we estimate FISN’s NOAK LCOE at $170/MWh.

Source: Deep Fission (p. 11), Lazard, Iceberg

At this cost level, FISN would be more expensive than most firm energy sources (except new conventional nuclear).

Source: Iceberg, Lazard

Finally, let’s remember that the nuclear industry is prone to cost overruns. FISN is particularly at risk as its design is still evolving. One illustration is the changing diameter of the borehole. Initially, the company envisioned a 15MWe reactor in a 30-inch borehole. It later revised (p. 8) the design to 8MWe in a 42-inch borehole. More recently, at an August 2026 conference, FISN’s CEO stated (~24:58) that the company is now targeting just 5MWe in a 40-inch borehole and 15MWe in a 60-inch borehole.

The trend is clear: FISN now requires substantially larger boreholes to deliver the same, or even lower, power output than originally envisioned. This matters because larger boreholes materially increase drilling costs. Our estimates above assume a 42-inch borehole for 8MWe and a 50-inch borehole for 15MWe because these are the diameters underlying the company’s disclosed financial projections. They therefore do not reflect FISN’s latest design changes. Factoring in the larger boreholes would make the project economics even worse than our current estimates.

Source: Fire2Fission, NRC, Canaccord Genuity

The certification timeline is wildly optimistic

For AI data centers, energy readiness or speed-to-power is a top priority. FISN’s marketing pitch is cleverly built around this requirement. 

The company claims that, unlike many SMR vendors pursuing novel reactor designs, it could follow a simpler and faster licensing process. This is because its reactor is based on well-established technologies, including PWR, the dominant reactor design used in operational nuclear power plants today, which could reduce regulatory uncertainty and streamline the path to approval. As for the unconventional idea of burying a reactor one mile underground, the company would simply recycle technologies developed by the oil & gas/geothermal industries. 

The company targets NRC approval and its first commercial reactor by the end of 2027. However, progress to date has been lackluster and nothing seems to support this aggressive timeline. 

FISN’s R&D and licensing strategy is as follows:

• Participate in the DOE Reactor Pilot program. The program provides an accelerated pathway for advanced reactor developers to build and test demonstration reactors. It is less onerous than the NRC’s licensing process, but it does not authorize commercial electricity generation. 

• Use the DOE process to support a subsequent NRC application for a commercial license. After testing the pilot reactor, FISN plans to apply to the NRC, leveraging work completed through the DOE pilot program.

a) DOE Reactor Pilot program

FISN was selected by the DOE in August 2025 to participate in the reactor pilot program. The program provides no financing, and at least nine other companies were also selected. 

If the bull case is that using existing technology should accelerate licensing, then one would expect passing the DOE review to be a walk in the park. FISN targeted achieving criticality by July 2026 (criticality is when a nuclear reactor achieves a self-sustaining nuclear chain reaction). In a November 2025 podcast, the CEO was enthusiastic about reaching the target: 

“But it [getting regulatory approval to build the reactor] doesn’t have to take a long time to get there, so the target date for having the first reactor turned on is July 4, 2026…we’re counting down because it’s a tremendous opportunity. We had expected to build our reactor in 2029, and now all of a sudden we have this fast path to build it in 2026, which is very exciting…”

However, the company dropped the target in 2026. Meanwhile, four SMRs developed by Antares Nuclear, Valar Atomics, Deployable Energy and Aalo Atomics reached criticality by July 2026. None of these reactors uses PWR technology.

Rather than acknowledging the missed target, the CEO shifted the goalposts, stating (~8:33) that the company had never been focused on criticality and characterizing the milestone as merely a “science experiment.” :

“Our focus has been commercial deployment from the beginning. We are not focused on science experiments. We are not focused on criticality tests. We are not focused on test reactors. From the start, we have said we want to build a commercial reactor that will generate electricity and start to bring in revenue”

The company is not even close to reaching its initial target. It has completed only the first stage of the process: the Nuclear Safety Design Agreement (NSDA), which simply establishes the framework (p. 7) and ground rules for the safety review that follows.

Source: Deep Fission, DOE

It took FISN nearly a year to receive its NSDA in early August 2026. By comparison, peer Antares, which is developing a novel sodium heat-pipe-cooled reactor, completed the same step in just three months. 

If the company’s competitive advantage is its reliance on well-known technologies, why is the process taking so long? We believe that placing the reactor underground introduces a host of new safety issues that regulators must consider. The CEO hinted (18:28) at this herself:

“This is a document that lays out the foundation of our safety case that we will then build on for the Detailed Safety Analysis, but it’s really a significant milestone because it covers everything that’s different about the approach that we’re planning on using, and it has now been approved by the Department of Energy.” (emphasis added)

This again runs contrary to the bull case that using conventional PWR technology should automatically translate into a faster licensing process.

b) NRC licensing

FISN targets an NRC licensing submission in the first half of 2027 and approval by the second half, implying a review period of less than a year. Such a timeline is unprecedented. As shown below, NuScale took 74 months to secure approval for its first 50MWe reactor, while even at the faster end, Kairos took 16 months just to obtain a construction permit for its test reactor – not a full commercial license.

Source: DOE, NRC | Start date reflects the application submission date per the NRC

Recent regulatory reforms could accelerate the licensing process for the industry. But it is very doubtful that FISN can meet its stated regulatory timeline even under a more streamlined framework.

In July 2024, FISN met with the NRC to discuss its reactor design as part of its pre-licensing engagement. NRC staff provided observations and feedback on the design and raised a laundry list of potential issues, including the unproven use of geology as containment and heat sink, an unclear disposal pathway and neutron leakage.

And even if FISN obtained an NRC commercial license, would that matter? NuScale offers a warning. NuScale’s reactor is far closer to conventional nuclear technology than FISN’s: essentially a miniaturized, above-ground PWR. This conventional design helped NuScale receive the first NRC certification for an SMR in February 2023. But has NuScale sold any reactors since? More than three years later, it still has not sold a single reactor despite a very supportive environment for power demand. If NuScale, with a more conventional design and a multi-year regulatory head start, still cannot secure a commercial order, we see no reason to expect an easier path for FISN.

And if a data center developer truly wants to build a PWR-based nuclear plant, we believe it will turn to an established player such as Westinghouse or GE Hitachi, which offer a more proven solution than either NuScale or FISN.

A 6-month commercial deployment timeline is also wishful thinking

Once licensed, FISN claims that its reactor could be deployed in as little as six months, versus an estimated three to four years for peer SMR designs. FISN is trying to convince AI data centers (or perhaps retail investors) that, unlike other SMRs, it can offer a solution almost immediately rather than in the distant future.

Source: Deep Fission

This aggressive claim is already undermined by FISN’s own prospectus (p. 10), which states a two-year development timeline (not just deployment). This is probably an admission that the project is more complex than simply drilling a hole and placing a reactor inside it. For example, even before drilling can begin, extensive geological characterization is required to establish that the site is suitable. The NRC raised (p. 10) precisely this issue in its response to FISN’s 2024 white paper.

FISN stated:

“The ability to ship a prepackaged reactor to a site where boreholes can be drilled within days with emplacement and operation of the reactor within weeks of delivery”

The NRC responded:

“The basis for this statement is not clear. Based on experience with Yucca Mountain, investigations that included excavations at depth to evaluate the geologic conditions took a significant amount of time – – well beyond the days for drilling and weeks for emplacement and operation stated above.”

The claimed six-month deployment advantage is itself dubious. The company attributes (~1:17) this advantage to a reactor design that eliminates or simplifies several safety-related systems used in conventional PWR, such as the reactor pressure vessel, containment structure, and emergency core cooling system. But this strategy is not unique to FISN. Other advanced SMR developers have similarly adopted novel reactor designs that eliminate or simplify many of the safety systems required in conventional PWRs. For example, Kairos Power’s fluoride salt-cooled high-temperature reactor operates at near-atmospheric pressure and uses TRISO fuel, enabling it to eliminate a number of conventional PWR safety systems. 

FISN also says its reactor uses commercially available low-enriched uranium (LEU) as fuel, while competitors rely on high-assay low-enriched uranium (HALEU), next-generation fuels that currently lack an established supply chain. The US has the technical capability to enrich uranium to HALEU levels. In fact, Centrus Energy, a domestic LEU enricher, has already demonstrated this capability and enriched well over one metric ton of HALEU as of end-2025. The reason HALEU supply remains limited in the US is simply that demand for it is negligible today, given that virtually all operating US reactors run on LEU. 

Then, of course, FISN’s design introduces an additional requirement that no other SMR has: drilling a one-mile borehole through hard rock (e.g. granite) with an unusually large diameter. FISN claims that its reactor’s design leverages “existing capabilities in the oil, gas and geothermal industries”, implying that all technologies are off-the-shelf, mature and could be easily applied to constructing its reactor. It is true that geothermal and oil & gas industries commonly drill (p. 2) deep boreholes, but these are much narrower, only up to about 10 inches in diameter, versus FISN’s 40-60 inches. At a conference (~05:38), the CEO admitted that drilling a wider borehole is difficult: 

“We are drilling wells that are 1 mile deep, not that deep when it compares to some other oil and gas wells. But we are pushing the limits when it comes to the diameter. We are looking at building 38-inch diameter wells and getting even wider.” (emphasis added) 

The “largest pipeline of any advanced reactor company” is hot air

The company claims an eye-popping customer pipeline of up to 18.5GW. In fact, the company boasted (12:33) that it has the “largest pipeline of any advanced reactor company in terms of the customers that are signing up to work with us and to purchase electricity.” 

Of FISN’s claimed 18.5GW pipeline, only about 4GW can be tied to identifiable counterparties and even these warrant significant skepticism. One is a modest industrial park with no clear need or means to support a nuclear reactor. The other two, like the industrial park, have signed only non-binding agreements, despite flattering press releases. 

Great Plains Industrial Park

The company entered into a non-binding letter of intent (LOI) with Great Plains Industrial Park (GPIP) in Parsons, Kansas, for the potential development of up to 2GW of generating capacity. GPIP has also been selected as the site of FISN’s pilot reactor, which the company ultimately intends to convert into a commercial facility. FISN has described GPIP as “an established site for industrial and energy development.” 

Source: Great Plains Industrial Park

The park is large with a 6,800-acre footprint but its current economic activity is limited. At a May 2026 board meeting, GPIP reported revenue of $219,000 in March and $197,000 in April. Annualizing those two months implies revenue of just $2.5mn.

A substantial portion of the site remains either undeveloped or devoted to relatively low-energy uses, such as agriculture. The Wall Street Journal described the park as “mostly overgrown”. 

The park has none of the large electricity consumers needed to support a multi-gigawatt nuclear development. There also does not appear to be a committed pipeline of large electricity consumers, such as large data centers or heavy industrial facilities, at GPIP. FISN’s own management has acknowledged that the scale of the eventual project depends on attracting those customers. At a community town hall in May 2026, FISN COO Mike Brasel told local residents that:

“build out will depend on them attracting load; industrial loads, AI data centers, That’s up to them.”

In addition, the project is facing opposition from the local community.

Blue Owl

In a Feb 2026 press release, the company announced a strategic relationship with Blue Owl under which the two parties would “collaborate to deploy Deep Fission SMR projects for Blue Owl’s digital infrastructure portfolio”. Blue Owl also invested (p. 7) $20mn in the company’s equity.

However, FISN filings suggest that the partnership is less substantive than the press release implies. The company states that the partners’ memorandum of understanding (MoU) “contemplates potential collaboration relating to customer and utility outreach, project development activities, and evaluation of potential power offtake and project-level financing opportunities associated with Blue Owl’s digital infrastructure portfolio and developments.”

The filing further states: 

“The MOU does not require either party to enter into any specific project, power offtake arrangement, financing transaction, or other commercial arrangement…”

Endeavour Energy, LLC

FISN announced its partnership with Endeavour, a credible counterparty with exposure to data centers, in January 2025. Endeavour is also an investor in FISN. The initial announcement described the partnership as follows:

“Endeavour and Deep Fission have committed to co-developing 2 gigawatts (GW) of nuclear energy to power Endeavour’s expanding global portfolio of Edged data centers, with the first reactors expected to be operational in 2029.” (emphasis added)

However, the most recent prospectus (p. 17) clarified that the relationship is non-binding: “The Endeavour relationship is governed by a non-binding term sheet and does not create commitments to purchase electricity, finance projects, construct facilities, grant exclusivity, or deploy a specified number of reactors.”

Large stock dilution is inevitable

As discussed above, we doubt FISN will ever achieve large-scale commercialization given the poor economics of its reactor. Even if commercialization were possible, it would occur much later than management’s current guidance suggests. In the meantime, the company will continue burning cash. FISN itself has disclosed that its current cash runway is less than twelve months:

“…,management’s current operating plan indicates that the proceeds from the offering, together with existing cash and cash equivalents, will not be sufficient to fund the Company’s planned operations through the one-year period following the issuance of these interim unaudited condensed consolidated financial statements. Accordingly, these circumstances continue to raise substantial doubt about the Company’s ability to continue as a going concern within one year after the date these interim unaudited condensed consolidated financial statements are issued.” (emphasis added)

In a January 2026 investor presentation, the company projected (p. 30) its spending plan for 2026 and 2027. 

Source: Deep Fission (p. 30)

The company assumed that it would burn $261mn of cash in 2026 and 2027 before it could begin constructing its first commercial project. Against that plan, the company has spent only $42mn in the first half of 2026, potentially reflecting schedule slippage and funding constraints. Assuming the original spending plan remains broadly intact, FISN would therefore need to spend another $218mn before reaching the commercial construction stage.

Separately, the company estimates approximately $750mn of development and construction costs for its first commercial multi-reactor project, before financing. The company wants to build and operate the plant itself, rather than sell a reactor design. Assuming the project is financed with 30% equity, FISN would need to contribute $225mn of equity capital.

Taken together, this implies a total cash requirement of $443mn through 2027: $218mn of remaining pre-commercial spending plus $225mn of equity funding for the first commercial project.

Against this requirement, the company has only about $93mn of cash as of June 2026, implying a funding gap of $350mn. That amount is equivalent to 76% of FISN’s current $459mn market capitalization. The shortfall is likely to be funded entirely through equity, which at current share price implies about 43% dilution to existing shareholders.

Source: Deep Fission

The company has warned investors in its prospectus:

“Given the magnitude of our anticipated capital requirements, future equity financings could result in substantial dilution to stockholders.”

An avalanche of newly listed SMR companies compete for funding

FISN’s need for additional fundraising comes at a time when investor appetite for SMR stocks has materially weakened, as the broader SMR investment theme has cooled and SMR-related stocks have lost momentum. After years of warnings about the technology’s readiness and economics, institutional investors are noticing that natural gas is the actual solution being adopted by AI data centers, while SMRs remain a distant and highly speculative prospect. Retail investors, who largely fueled the bubble, are turning their attention to SpaceX’s space-based data center concept. 

Source: Google Finance, as of 23 September 2026 | SMR, NNE, OKLO, NKLR, IMSR, HDRN, NWCL are benchmarked to their share prices as of October 1, 2025. XE and FISN are benchmarked to their respective share prices on April 4, 2026 and June 18, 2026, their respective first trading days.

Case in point: FISN had to cut its IPO price from its initial target range of $24-$26 per share to just $16 per share. Similarly, newly public SMR-adjacent stock Standard Nuclear (NYSE: STDN) slashed the size of its IPO by more than 50%. Holtec had to suspend its IPO.

Just as investor demand for SMR stocks has declined, the supply of publicly traded nuclear and SMR names has risen rapidly and is set to increase further. Before October 2025, there were only three listed SMR stocks, but that number has since risen to nine. Importantly, Westinghouse, an established nuclear reactor company, is finally expected to list, giving investors better alternatives in the SMR theme. Two smaller SMR companies, NuCube and ONE Nuclear, are also expected to go public, bringing the total to twelve. 

Deep Fission recycles the founders’ past venture that never achieved commercialization 

FISN was founded by a father-daughter duo, Richard Muller and Elizabeth Muller, who serve as the company’s CTO and CEO, respectively. Richard Muller is an emeritus physics professor at UC Berkeley. His prolific academic career spans elementary particle physics, astrophysics, and geophysics but he is not a nuclear engineer. Elizabeth Muller holds a bachelor’s degree in mathematics from UC San Diego and a master’s degree in International Management from ESCP Business School in Paris.

The two co-founders’ most relevant experience is their first startup, Deep Isolation. In fact, Deep Fission is largely a repackaging of Deep Isolation. Founded by the duo in 2016, Deep Isolation aims to commercialize deep-borehole nuclear waste disposal. The company wants to drill a borehole ~1 km deep, place high-level nuclear waste in canisters, and bury them in the borehole. 

Deep Isolation has never commercialized its deep-borehole waste-disposal business, despite its ten-year existence. The company has never achieved profitability. Deep Isolation did generate $6.1mn of revenue in 2025. But 59% of that revenue was unrelated to deep-borehole nuclear waste disposal and instead came from its subsidiary Freestone, an environmental and water-resources consulting firm acquired in 2021. 

Almost all of the remaining revenue (39%) was derived from so-called “strategic analysis” contracts. The company describes these contracts as “typically relat[ing] to smaller consulting contracts or services performed under grants … used by customers to have us review potential geology formations for suitability to DBD…”. In other words, these are government-supported feasibility studies exploring deep-borehole waste disposal, not contracts for commercialization.

Conclusion

FISN’s bull case ultimately rests on lower cost and faster deployment. 

The economics are unattractive. Even on FISN’s own assumptions, power costs exceed competing firm-power alternatives. If capital costs are adjusted to a more realistic level, the economics deteriorate further. 

FISN’s use of established PWR technology is supposed to accelerate licensing, yet the company already trails its peers in the less demanding DOE process. FISN claims its simplified design enables much faster deployment, but peers pursue the same simplification strategy. Burying a reactor will clearly add technical and regulatory challenges.

Founded only three years ago, FISN is still just a concept, as the prospectus itself underlines (p. 8): “The Gravity Reactor remains in conceptual and early engineering stages”. The fact that its LCOE and capital cost claims have already disappeared speaks volumes. How much is a concept worth? A lot, when you are in the middle of a bubble. But that time has clearly passed.

With no advantage in either speed or cost, FISN will not be a winner in this increasingly crowded market. Pressing ahead with the IPO despite having to slash the offer price looks like a late attempt to ride the SMR hype, and dump a stranded asset onto retail investors.

Exhibit A: LCOE estimate

We estimate FISN’s LCOE using the following formula:

In simpler terms, the formula estimates the total after-tax present value of CAPEX and OPEX per unit of the present value of electricity generated over the system’s lifetime. 

Key inputs include:

Source: Deep Fission, Lazard, Iceberg

FISN estimates (p. 11) average annual operating expenses over the first ten years at $4mn for FOAK and $8mn for NOAK, assuming annual cost inflation of 3%. We therefore estimate Year 0 OPEX at $3.388mn for FOAK and $6.775mn for NOAK by dividing the reported 10-year average OPEX by the average inflation factor over the period.

The 10-year average inflation factor is calculated as:

We assume a five-year MACRS depreciation schedule:

We assume full and immediate monetization of the tax benefits, which is generous for FISN given it probably does not have taxable income to utilize the tax benefits.

Tax rate (federal and state) is assumed to be 40% to align with Lazard’s assumption to ensure comparability. 

The LCOE model for FOAK is shown below:

Figures are expressed in US$mn, unless otherwise stated.

Deployment occurs over one year in Year 0. OPEX begins at $3.489mn in Year 1 and escalates at 3% annually over the subsequent 40 years. Annual electricity generation is calculated as 8MWe of capacity multiplied by 8,760 hours per year and a 93% capacity factor. CAPEX is based on FISN’s estimates. CAPEX, OPEX, depreciation, and electricity generation are discounted to Year 0 using a WACC of 7.7%.

The model outputs are as follows:

The LCOE model for NOAK is shown below:

Figures are expressed in US$mn, unless otherwise stated.

Deployment occurs over two years, in Year 0 and Year 1, consistent with FISN’s assumption. Total CAPEX of $84mn is assumed to be split evenly across the two years. OPEX begins at $6.978mn in Year 2 and escalates at 3% annually over the subsequent 40 years. Annual electricity generation is calculated as 15MWe of capacity multiplied by 8,760 hours per year and a 93% capacity factor. CAPEX, OPEX, depreciation, electricity generation are discounted to Year 0 using a WACC of 7.7%.

The model outputs are as follows:

Exhibit B: Scaling factor

We use the common economics of scale relationship for nuclear reactors as expressed below:

The value of the scaling factor k can range from 0 to 1, where k=1 implies that there are no economies of scale, while a lower value implies larger economies of scale. The rule of thumb for the scaling factor is 0.6.

Disclaimer

Our research and reports express our opinions, which we have based upon generally available public information, field research, inferences and deductions through our due diligence and analytical process. To the best of our ability and belief, all information contained herein is accurate and reliable, and has been obtained from public sources we believe to be accurate and reliable, and who are not insiders or connected persons of the issuer of the security covered herein or who may otherwise owe any fiduciary duty or duty of confidentiality to such issuer. We strive for accuracy and completeness to support our opinions, and we have a good faith belief in everything we write, however, all such information is presented “as is,” without warranty of any kind – whether express or implied. Iceberg Research (“Iceberg”) makes no representation, express or implied, as to the accuracy, timeliness, or completeness of any such information or with regard to the results to be obtained from its use. You agree that the use of Iceberg’s research is at your own risk. In no event will Iceberg or any person affiliated therewith be liable for any losses caused by, or arising out of, any information available in this report. 

Think critically about our opinions and do your own research and analysis before making any investment decisions. This material and any opinions expressed herein should in no way be viewed as advice on the merits of any investment decision with respect to any issuer or any transaction or other matter. This material is not (and may not be construed to be) legal, tax, investment, financial, or other advice. You should consult your own legal counsel and tax and financial advisers as to legal and other matters concerning the information and statements contained herein. This material does not purport to be all-inclusive or to contain all of the information that may be relevant to an evaluation of the issuer, its securities, or the matters described herein. 

This material contains forward-looking statements. Forward-looking statements are statements that are not historical facts and may include projections and estimates and their underlying assumptions, statements regarding plans, objectives, intentions and expectations with respect to future financial results, events, operations, services, product development and potential, and statements regarding future performance. Forward-looking statements are generally identified by the words “expects”, “anticipates”, “believes”, “intends”, “estimates”, “plans”, “will be” and similar expressions. Although we believe that the expectations reflected in forward-looking statements contained herein are reasonable, investors are cautioned that forward-looking information and statements are subject to various risks and uncertainties—many of which are difficult to predict and are generally beyond the control of Iceberg or the issuer—that could cause actual results and developments to differ materially from those expressed in, or implied or projected by, the forward-looking information and statements. In addition, the foregoing considerations and any other publicly stated risks and uncertainties should be read in conjunction with the risks and cautionary statements discussed or identified in the issuer’s public filings with the U.S. Securities and Exchange Commission, including those listed under “Risk Factors” in the issuer’s annual reports on Form 10-K and quarterly reports on Form 10-Q. The forward-looking statements speak only as of the date hereof and, other than as required by applicable law, we do not undertake any obligation to update or revise any forward-looking information or statements. 

You should assume that as of the publication date of our reports and research, Iceberg (along with or through our members, partners, affiliates, employees, consultants, clients, and/or third-parties) has a short position in the securities (and/or options, swaps, and other derivatives related to the securities) covered herein and therefore stands to realize gains in the event that the price of the covered securities declines. We (including our members, partners, affiliates, employees, consultants, clients, and/or third-parties) may continue transacting in the securities of the company covered in this report, and we may buy, sell, cover or otherwise change the form or substance of our position in the issuer regardless of our initial views set out herein. 

This is not an offer to sell or a solicitation of an offer to buy any security, nor shall Iceberg offer, sell or buy any security to or from any person through this site or reports on this site. Iceberg is not registered as an investment advisor in any jurisdiction. You agree to do your own research and due diligence before making any investment decision with respect to securities discussed herein. You represent to Iceberg that you have sufficient investment sophistication to critically assess the information, analysis and opinions in this report.

We are entitled to our opinions and to the right to express such opinions in a public forum. We believe that the publication of our opinions about public companies that we research is in the public interest. This report and all statements contained herein are the opinion of Iceberg and are not statements of fact. You can publicly access any piece of evidence cited in this report or that we relied on to write this report. All expressions of opinion are subject to change without notice, and Iceberg does not undertake to update or supplement any reports or any of the information, analysis and opinion contained in them.

You agree that use of Iceberg’s research is at your own risk. In no event will you hold Iceberg or any affiliated party liable for any losses caused by, or arising out of, any information on this site. You further agree to do your own research and due diligence before making any investment decision with respect to securities covered herein. You represent to Iceberg that you have sufficient investment sophistication to critically assess the information, analysis and opinion on Iceberg’s site and in this report. You further agree that you will not communicate the contents of this report to any other person unless that person has agreed to be bound by these same terms of service.

By downloading, opening and/or reading this report you knowingly and independently agree: (i) to abide by the terms of service of our website, which are hereby fully incorporated herein, (ii) that any dispute arising from your use of this report or viewing the material herein shall be governed by the laws of the State of New York, United States, without regard to any conflict of law provisions; (iii) to submit to the personal and exclusive jurisdiction of the superior courts located within the State of New York and waive your right to any other jurisdiction or applicable law; and (iv) that regardless of any statute or law to the contrary, any claim or cause of action arising out of or related to use of this website or the material herein must be filed within one (1) year after such claim or cause of action arose or be forever barred. The failure of Iceberg to exercise or enforce any right or provision of this disclaimer shall not constitute a waiver of this right or provision. If any provision of this disclaimer is found by a court of competent jurisdiction to be invalid, the parties nevertheless agree that the court should endeavor to give effect to the parties’ intentions as reflected in the provision and rule that the other provisions of this disclaimer remain in full force and effect, in particular as to this governing law and jurisdiction provision.