A reactor component has a manufacturing contract. That is progress. It also leaves a fairly important question: who pays for the finished power plant?
Small modular reactor announcements can put design approvals, potential electricity sales and factory orders into the same optimistic paragraph. They describe different commitments, with different consequences for cash. The useful finance question is which company has promised to do what, and which conditions still stand between that promise and payment.
This assessment uses disclosures checked on 30 October 2026. It compares three business positions, rather than ranking reactor technologies or predicting share prices.
Fact: a factory contract moves one part of the project
On 29 October, NuScale Power (NYSE: SMR) announced a contract covering manufacture of the first upper reactor coolant system article for its power module. Its manufacturing announcement describes a defined component and production work. It does not establish that a customer power station is operating or that its electricity is being sold.
The regulatory achievement also needs its proper label. The US Nuclear Regulatory Commission completed the standard design approval for NuScale’s US460 in May 2025. The approved configuration comprises six modules rated at 77 megawatts electric each, or 462 MWe in total, according to the NRC project record.
Design approval · May 2025
Capacity on the approved design
Six modules × 77 MWe
Design rating; not operating output
That approval can be referenced in subsequent licensing applications. The approval itself expressly does not commit the regulator to issue a permit or licence. An approved design, a manufacturing order and permission for a particular plant therefore remain separate pieces of evidence.
NuScale’s August second-quarter update described its strategic partner’s discussions towards a definitive power purchase agreement with the Tennessee Valley Authority. That is a dated description of negotiations, not a claim here that a completed agreement exists today. Its June-quarter business note said significant revenue generated to date came from engineering and licensing fees and services for potential customers.
Fact: a customer pipeline contains several kinds of promise
Oklo (NYSE: OKLO) illustrates why the agreement’s name matters. Its June-quarter filing distinguishes non-binding customer letters from a prepayment mechanism intended to fund fuel and deployment work. It also describes a fuel-supply letter of intent whose pricing, volumes and other material terms still required negotiation. These are different commitments; adding their proposed capacity together would not produce a comparable revenue backlog.
The same filing reports first criticality at its Groves Isotope Test Reactor on 5 August. Criticality means a sustained nuclear chain reaction; here the milestone concerns a low-power test reactor. Oklo’s 13 October update describes a further isotope production and processing pilot project. Neither announcement should be relabelled as a commercial electricity plant delivering contracted power.
Interpretation: follow the obligation and the direction of cash
For each announcement, identify the buyer, seller, deliverable and trigger for payment. A developer ordering equipment can be making a cash commitment. A supplier receiving an equipment order may be gaining future revenue. A prospective electricity buyer signing an initial letter may still have made neither commitment. The same project can therefore look quite different in three sets of accounts.
A practical reading sequence starts with the contract’s binding scope. Then ask which permissions and financing conditions remain, whether payment is refundable, and who absorbs cost overruns. Finally, look for evidence of work performed and cash collected. Where commercial terms are undisclosed, the sensible entry is “unknown”. An announcement cannot fill that box for us.
BWX Technologies (NYSE: BWXT) provides a useful supplier comparison. In January 2025 it announced a reactor pressure vessel manufacturing contract for the Darlington small-reactor project, saying the order had been booked in the second quarter of 2024. That is evidence of manufacturing work, rather than merely potential demand for electricity.
Its June-quarter report defines backlog as remaining performance obligations under qualifying contracts, while warning that projects can be cancelled or modified. It also describes revenue recognition as work progresses, generally measured using costs incurred against estimated total costs. A booked order still needs execution; reported revenue and cash collection need not arrive together.
Who could benefit?
Interpretation: NuScale could gain licensing, engineering and module-related business if customer projects turn into executable delivery commitments. The useful milestones over coming reporting periods are disclosed binding scope, funded project steps and payments. The new factory work supports readiness; it does not quantify eventual customer revenue or profitability.
Oklo’s potential benefit depends on converting development arrangements into plants that can deliver paid power. Its fuel and deployment commitments make financing and execution central to that longer-term opportunity. Evidence of satisfied contract conditions would strengthen the case; headline capacity alone would not.
BWXT could earn manufacturing revenue during construction, before a completed reactor sells electricity, if contracted work proceeds. The opportunity must be measured against its wider business: the second-quarter results discussion covers both Government and Commercial operations. Treating every dollar of group growth as a small-reactor outcome would overstate the connection.
Who’s at risk?
Interpretation: NuScale and Oklo face the possibility that spending advances faster than customer commitments or plant receipts. Delays can extend the period before the intended business pays for itself. That is an execution and financing mechanism, not a forecast that either company will fail or that either share price must fall.
BWXT faces a different timing problem: changes to scope, estimated costs or customer schedules can alter the margin on booked work. Its accounting disclosures make that sensitivity explicit. Contract awards reduce one kind of uncertainty without removing the obligation to deliver economically. None of the sources used here establishes a reliable project-level return that can be applied across all three companies.
Speculation: what would count as progress?
Repeated builds might eventually make factory production more predictable and spread development costs across more projects. They might also reveal bottlenecks that a first component cannot expose. Those are scenarios, not measured outcomes, and this article assigns neither probabilities nor a future electricity price.
For the next update, the test is concrete: a named obligation, conditions satisfied, work completed and the associated revenue or cash disclosed. A production milestone deserves recognition. The business case still needs the rest of the receipts.
This article is for information and analysis, not investment advice.
Photo by Melih Türkmenoğlu on Pexels. Electricity pylon at sunset.