NuScale vs Oklo: Two Very Different Routes to Small Nuclear Power

NuScale and Oklo are often placed in the same “small modular reactor”
category. That label hides a major difference.

NuScale is developing a smaller version of familiar light-water
reactor technology, with several 77-megawatt-electric modules intended
to work together as one power station. Oklo is pursuing compact
sodium-cooled fast reactors and a business model in which it would own
the plants and sell their electricity and heat.

Both companies say their reactors could serve rising demand from data
centres, industry and electricity grids. Both have also made genuine
regulatory and project-development progress. But as of 30 August 2026,
neither has an operating commercial reactor of the type it is now
promoting.

The most useful comparison is therefore not which company has the
bigger announced pipeline. It is which technical, regulatory and
commercial problems each one has chosen to solve.

NuScale vs Oklo: how the reactor designs differ

Feature	NuScale	Oklo
Main design	Integral pressurised light-water reactor	Sodium-cooled, metal-fuelled fast reactor
Proposed output	77 MWe per module; six-module US460 plant totals 462 MWe	Current Aurora product line described as 15–75 MWe
Fuel	Conventional light-water-reactor fuel enriched below 5% uranium-235	Metal fuel; the Idaho demonstration plans to use recovered HALEU material
Deployment model	Sell modules and associated engineering, licensing and services	Design, build, own and operate plants; sell power and heat under agreements
Current US regulatory position	US460 received NRC Standard Design Approval in May 2025	Current work remains in pre-application engagement with the NRC; Idaho demonstration is also pursuing a DOE authorisation route

NuScale’s reactor is an integral pressurised-water
reactor. The core, pressuriser and steam generators are
contained within one reactor vessel, while modules are partly submerged
in a below-grade pool. Water cools the reactor and carries heat to the
steam system. The design uses natural circulation, reducing reliance on
large reactor coolant pumps.

The US Nuclear Regulatory Commission describes the approved US460
configuration as six 77 MWe modules with a combined output of 462 MWe.
NuScale says the modules use standard 17-by-17 light-water-reactor fuel
assemblies enriched to less than 5%, with refuelling intervals of up to
21 months. The
NRC completed its US460 technical design review in May 2025.

Oklo’s Aurora belongs to a different reactor family. It uses liquid
sodium rather than water as the coolant and metal fuel in a fast-neutron
spectrum. Oklo’s latest filing describes planned Aurora variants from 15
to 75 MWe. The company says the approach is intended to support long
operating periods between refuelling and, eventually, different fuel
sources.

For the first proposed Aurora at Idaho National Laboratory, the US
Department of Energy has made five metric tonnes of recovered high-assay
low-enriched uranium, or HALEU, available for fuel
fabrication. HALEU contains a higher proportion of uranium-235 than fuel
used in most existing US commercial reactors, but remains below 20%. DOE
retains ownership of the material and has approved the conceptual design
for the associated fuel-fabrication facility.

This creates a clear trade-off. NuScale can draw on decades of
light-water-reactor knowledge and a more established fuel supply chain.
Oklo aims for a smaller, potentially longer-running system, but must
demonstrate a less familiar reactor, develop its fuel arrangements and
establish a new operating record.

NuScale
has the stronger design approval—but not a plant licence

NuScale reached an important milestone in May 2025 when the NRC
issued Standard Design Approval for the 77 MWe US460 design. That review
covered the standard reactor design and allows it to be referenced in
later licence applications.

It does not authorise NuScale or a customer to build
and operate a specific power station. A real project would still need a
site, financing, project-specific environmental and safety reviews,
construction, fuel and the appropriate NRC licence.

NuScale’s most advanced international proposal is a six-module
project at Doicești in Romania, on the site of a former coal-fired power
plant. The Romanian government approved an investment decision in
February 2026, allowing further financing, licensing, geotechnical work
and site-specific design. NuScale’s latest quarterly filing says
additional pre-construction engineering remains dependent on financing
and new contracts. The
filing does not identify a firm construction start or operating
date.

In the United States, NuScale’s commercial partner ENTRA1 announced a
framework with the Tennessee Valley Authority contemplating up to 6
gigawatts of capacity. The scale attracted attention, but NuScale’s
filing describes the arrangement as non-binding and
says the next phase depends on one or more power-purchase agreements
between ENTRA1 and TVA. It should not be counted as a firm reactor
order.

The company’s earlier Carbon Free Power Project with Utah Associated
Municipal Power Systems is an important warning about first-of-a-kind
economics. The parties terminated it in November 2023 after it became
unlikely to secure enough participating subscriptions. NuScale had
previously raised its target electricity price to $89 per megawatt-hour,
citing inflation and material-cost pressure. The
termination showed that regulatory progress alone does not guarantee a
financeable project.

Oklo
is advancing a demonstration, but key approvals remain

Oklo’s regulatory story requires more careful reading.

The NRC denied the company’s first combined licence application
without prejudice in January 2022 because the application did not
contain enough information for the agency to make its safety findings.
That application concerned an older, much smaller heat-pipe reactor
concept—not the present 75 MWe sodium-cooled Aurora proposal. The
NRC closed that original review but allowed Oklo to submit a future
application.

Oklo is now working with the NRC through pre-application reviews for
its newer design. In April 2026, the NRC approved a Principal Design
Criteria topical report for Aurora. This is useful because it
establishes fundamental safety and design criteria that can be
referenced later. It is not a construction permit, combined licence or
operating approval.

The first Aurora project at Idaho National Laboratory is also moving
through the Department of Energy’s Reactor Pilot Program. DOE approved a
Nuclear Safety Design Agreement in March 2026 and a Preliminary
Documented Safety Analysis in June. These are intermediate steps in the
federal authorisation process, not final permission to construct and
operate the reactor.

Oklo held a “groundbreaking” event in September 2025, but DOE’s
environmental documents limited the approved work to non-nuclear site
preparation and excluded reactor components. Oklo still targets 2028 for
the first powerhouse, although its latest SEC filing calls that timeline
ambitious and identifies supply-chain, construction and design risks. Oklo’s
second-quarter filing also confirms that no commercial Aurora is
operating yet.

Selling reactors
versus selling electricity

The companies also expose themselves to different business risks.

NuScale plans to sell reactor modules and provide engineering,
licensing, start-up, testing, fuel and refuelling services. A utility,
government-backed developer or project consortium would typically
arrange the power station’s ownership and financing.

Oklo plans to design, build, own and operate its powerhouses, then
sell electricity and heat under long-term agreements. That could create
recurring revenue if the plants work as intended, but it also places
more responsibility for development, financing, construction and
operations on Oklo.

Several large announcements remain preliminary. Oklo’s agreement with
Meta supports development work for a planned 1.2 GW nuclear campus in
Pike County, Ohio, with a first phase targeted as early as 2030. Its
much-cited 12 GW master agreement with data-centre operator Switch is
explicitly non-binding; individual binding power-purchase agreements are
expected only as projects meet later milestones. The
Meta announcement describes the Ohio dates as development targets, not
guaranteed completion dates.

Fuel is another constraint. A June 2026 letter of intent with Centrus
contemplated HALEU deliveries from 2029 for up to five Aurora units, but
volume, price, timing and prepayment terms still require a definitive
agreement. A
letter of intent is not a completed fuel-supply contract.

Which approach
is closer to commercial reality?

There is no single answer because the companies are closer in
different ways.

NuScale is further ahead on standardised US reactor-design review.
Its light-water technology and conventional fuel are also more familiar
to regulators, utilities and suppliers. Its unresolved challenge is
converting an approved design into a financed, licensed and constructed
plant at a competitive cost. As of its August 2026 filing, NuScale had
not delivered a module or secured a binding module-delivery
contract.

Oklo is attempting a more vertically integrated model aimed at
smaller, dedicated power users. It has made progress on Idaho site
preparation, DOE safety documentation and NRC design criteria. But the
newer Aurora design still lacks final construction and operating
authorisation, HALEU supply remains a bottleneck, and its major
deployment schedules are company targets rather than demonstrated build
times.

The next decisive evidence will not be another memorandum or headline
capacity figure. For NuScale, it would be a fully financed customer
project with a binding reactor order and site-specific licence. For
Oklo, it would be final authorisation, completed fuel fabrication and
successful operation of the first Aurora.

Small nuclear power may ultimately need both approaches: larger
multi-module plants that resemble scaled-down utility nuclear stations,
and compact powerhouses serving individual industrial or data-centre
sites. The race will be decided by licensing, construction performance,
fuel availability and delivered electricity—not by announced gigawatts
alone.

Reporting note

This article reflects public information available on 30 August 2026.
Neither NuScale’s Standard Design Approval nor Oklo’s approved Principal
Design Criteria and DOE safety milestones constitutes permission to
operate a commercial plant. Several cited project agreements are
non-binding, contingent on financing or subject to future definitive
contracts. Company targets may change.

Google’s agreement with Fervo explores another route: enhanced geothermal power for data centres.

Sources and further reading

1. US
   Nuclear Regulatory Commission: NuScale US460 Standard Design
   Approval
2. NuScale
   Q2 2026 Form 10-Q
3. NuScale
   and UAMPS terminate the Carbon Free Power Project
4. US
   Nuclear Regulatory Commission: Oklo Aurora pre-application
   activities
5. Oklo
   Q2 2026 Form 10-Q
6. US
   Department of Energy: Oklo fuel-fabrication facility design
   concept
7. Oklo
   and Meta: planned 1.2 GW Ohio development
8. Oklo
   and Centrus: HALEU letter of intent

FutureTechDose covers biotechnology, AI, data-centre and
energy-sector research and industry progress for a general audience.
This article is informational and does not provide medical or investment
advice.