Nuclear's New Gold Rush: Which 8 SMR Companies Are Worth Your Money?

The global race to power artificial intelligence (AI) data centers has quietly become an energy arms race — and small modular reactors (SMRs) are the weapon everyone wants. Tech giants Google and Amazon have taken equity stakes in nuclear startups. A company with White House connections has locked up 13 gigawatts of orders. And restarting decades-old nuclear plants has somehow become the fastest path to powering tomorrow's internet.

For investors, the question is no longer whether nuclear is back. It is which companies are worth the bet.

Analysts at Green Impact Academy have identified five critical variables for evaluating next-generation nuclear fission companies: technology pathway, grid interconnection capability, financing capacity, order-winning ability, and speed of regulatory clearance from the U.S. Department of Energy and Nuclear Regulatory Commission (NRC).

Of the eight companies examined in this analysis, the conclusions break down by investor priority:

• Lowest regulatory risk, fastest approval timeline: GE Vernova Hitachi Nuclear Energy (GEH) or NuScale
• Stable supply chain and manufacturing capacity: Rolls-Royce SMR or Holtec International
• Maximum safety profile and industrial heat applications: Kairos Power or X-energy
• ESG credentials and nuclear waste innovation: Oklo or TerraPower

The differences run deeper than marketing. The three main fission technology pathways — fast neutron (Generation IV), conventional light water (Generation 3.5), and TRISO fuel (Generation IV) — handle nuclear safety, proliferation risk, and waste management in fundamentally different ways.

Fast Neutron / Generation IV (Oklo, TerraPower)Oklo's liquid metal-cooled design operates at atmospheric pressure, making it inherently safer in many scenarios — but liquid sodium fire risk requires careful engineering. Its fuel, High-Assay Low-Enriched Uranium (HALEU), combined with potential plutonium breeding capability, demands stringent international oversight. The upside: fast reactors can technically burn high-level radioactive waste, compressing disposal timescales from tens of thousands of years to just hundreds.

Conventional Light Water / Generation 3.5 (GEH, NuScale, Rolls-Royce SMR, Holtec)These designs use standard Low-Enriched Uranium (LEU) and benefit from the most mature global regulatory framework. GEH has upgraded to natural circulation and gravity-driven cooling — a genuine safety improvement — but high-pressure systems retain residual risk. The waste picture is less encouraging: open-cycle fuel cycles produce the same long-lived radioactive waste as conventional nuclear plants, requiring geological disposal for over ten thousand years.

TRISO Fuel / Generation IV (Kairos Power, X-energy)TRISO fuel pellets represent the most robust fuel form available. A full reactor core meltdown is physically impossible with this design. The fuel uses higher-enrichment HALEU but is extremely difficult to repurpose for weapons. The trade-off: waste volumes are larger and harder to recover, though the material is highly stable and far less complex to engineer than alternatives.

The table below summarizes how all eight companies score across the three nuclear safety dimensions:

Table 1: Nuclear Safety — Three-Indicator Assessment Across Fission Technology Pathways. Source: John Wang / Green Impact Academy

Of the eight companies in this analysis, onlyNuScale andOklo are independently listed on public markets. The remaining six — TerraPower, GEH, Rolls-Royce SMR, Holtec International, Kairos Power, and X-energy — are either privately held or operate as divisions within large industrial conglomerates.

This matters enormously for investors. Both NuScale and Oklo are currently cash-burning operations. NuScale generates modest construction service revenue; Oklo has yet to record any revenue at all. Their share prices reflect market expectations about future potential, not present earnings — making conventional fundamental analysis essentially useless for valuing them.

The standard benchmark for comparing power generation economics is the Levelized Cost of Energy (LCOE): the full-lifecycle cost per megawatt-hour (MWh) that accounts for capital expenditure, operating costs, fuel, and decommissioning. For SMRs to compete with conventional thermal generation, they need to approach the current U.S. natural gas cost range of $48–$109 per MWh.

The data so far is instructive but incomplete. NuScale's initial project estimate came in at $58 per MWh — before cost overruns forced the entire project's cancellation. GEH's Canadian project currently carries an estimate of approximately $110 per MWh. Both figures represent design targets and research estimates, not operational reality.

The more useful comparison is the historical trajectory. Solar photovoltaic costs collapsed from $230 per MWh in 2010 to $34 per MWh by 2020. Wind power dropped from $440 per MWh in 1984 to $32 per MWh in 2020. First-of-a-Kind (FOAK) nuclear construction costs are high by definition; serial production is expected to bring SMR costs down to $50–$80 per MWh. For long-horizon investors, that trajectory is the thesis.

A recent industry survey of SMR and advanced reactor professionals identified the top three obstacles to commercial deployment: regulatory review processes (49%), FOAK construction and lack of precedent (45%), and access to capital and project financing (44%). The three are mutually reinforcing: regulatory uncertainty makes financing harder; financing difficulty delays construction; construction delays feed back into regulatory timelines.

The U.S. government's track record of budget shutdowns — 23 since the nation's founding, including a 43-day closure in 2025 — has historically made long-term nuclear funding commitments unreliable. The Consolidated Appropriations Act of 2024 (the "Minibus Package") addressed this directly by decoupling critical long-term nuclear funding from the annual budget cycle, providing a more stable financial foundation for regulatory approvals, FOAK construction, and capital deployment simultaneously.

Since traditional valuation metrics cannot guide investment in pre-revenue companies, order volume becomes the clearest proxy for future potential. The table below summarizes where each company stands.

Table 2: Partnership and Order Summary — Eight Next-Generation Fission Technology Companies. Source: John Wang / Green Impact Academy

Oklo's order lead is striking and explained largely by its proximity to power in Washington. The company's political connections have produced 13 GW in commitments from Switch Data Centers — though "non-binding" is the operative word. TerraPower trails closely with a more geographically diversified client base spanning data centers, industrial hydrogen, and international government partners, though most agreements remain letters of intent.

GEH stands out for contract quality over quantity. Its partnerships with Ontario Power Generation in Canada and Orlen in Poland represent the most commercially certain deals in the group. Kairos Power and X-energy have attracted something arguably more valuable than raw order volume: equity investment from Google and Amazon respectively — direct validation from the companies most desperate to solve their own power problem.

There is a structural bottleneck that no breakthrough technology can engineer around: grid interconnection.

Generating electricity and delivering it to the grid are two entirely separate challenges. Before any power plant — nuclear, solar, or otherwise — can feed electricity into the U.S. grid, it must obtain a feeder line allocation from the relevant grid operator. Those allocations are scarce. The U.S. interconnection queue is severely congested, with projects waiting years for approval.

The practical implication: an SMR that cannot connect to the grid is an expensive demonstration project. The fastest path to actual grid-connected nuclear power is not building new plants — it is restarting existing ones that already have feeder connections in place.

Two U.S. cases illustrate this clearly. The Duane Arnold Nuclear Power Plant is being restarted and expanded, with 6 GW of additional capacity planned at the existing site and a further 3 GW at a new adjacent facility that includes SMR technology. The Palisades Nuclear Power Plant — previously decommissioned — is being brought back online with federal loan support specifically to supply Microsoft's data centers. Both plants already hold feeder line allocations and bypass the interconnection queue entirely.

For sites with relatively stable demand profiles — particularly industrial facilities and large data centers — a third option exists: microgrid design. An SMR supplying a self-contained microgrid can bypass interconnection requirements entirely, adjusting output to match on-site demand and eliminating the curtailment losses that affect solar and wind installations.

Given that neither NuScale nor Oklo can be valued through conventional earnings analysis, the investment framework here is explicitly speculative. The recommended approach: allocate 1%–5% of a portfolio to capture the upside of the price-to-dream ratio — buying at a point of relative undervaluation and holding for the long term. The asymmetry is attractive: if the thesis plays out, returns of ten times or more are plausible; if it does not, the loss is contained.

For investors who prefer more defensible fundamentals, two adjacent plays offer exposure to the nuclear supply chain without the binary risk of pre-revenue startups:

• BWX Technologies (ticker: BWXT) — holds a near-monopoly in nuclear components manufacturing
• Centrus Energy (ticker: LEU) — the only U.S. company currently producing HALEU

Both trade at elevated valuations, with price-to-earnings ratios of 50–60 as of this writing. Neither is a buy at current prices; both are worth watching for a correction.

South Korea's trajectory offers a pointed lesson. Both South Korea and Taiwan built their manufacturing sectors in the same postwar generation, and both made early strategic bets — Taiwan on semiconductors, Korea on a broader industrial base. Today, Taiwan's economy is concentrated almost entirely in chips while other manufacturing sectors have weakened. South Korea, meanwhile, has become a global power in nuclear energy, with Korean reactor designs serving as the price-performance benchmark for Western buyers. X-energy and NuScale have both turned to Seoul for supply chain partnerships. No equivalent outreach has come to Taipei.

Taiwan has allowed its hard-won nuclear engineering talent to migrate elsewhere over the past two decades. But it still holds two strategic assets: a dominant position in AI hardware, and a $500 billion U.S.-directed investment commitment emerging from the current trade environment. The question now being asked in some policy circles is whether Taiwan could use those assets as an entry ticket into the nuclear supply chain — the way it once used its manufacturing base to enter semiconductors — before the window closes.

That would require a serious, sustained commitment of talent across new energy, electrical engineering, power systems, and finance. It would require building foundational infrastructure piece by piece, the way energy storage capacity was developed. And it would require treating nuclear power not merely as a low-carbon energy option, but as a dimension of the technology arms race and national strategic positioning.

Whether Taiwan will move in that direction remains an open question. What is not open to question is that the race is already underway.

About the Author |John Wang

（Related： The TSMC Effect: How Taiwan’s Pension Funds Became Bound to a Single Tech Giant ｜ Latest ）