How the NRC Actually Works, and Why It Takes So Long

You cannot understand nuclear timelines without understanding the institution that governs them, so let me describe the Nuclear Regulatory Commission plainly and without either the industry's impatience or the anti-nuclear movement's suspicion. The NRC is the independent federal agency that licenses and regulates civilian nuclear power in the United States. Its mandate is safety, and its institutional temperament is caution, because the cost of a mistake in its domain is catastrophic and permanent in the public mind. When people complain that the NRC is slow, they are describing a feature the agency was designed to have. Speed and certainty trade off, and the NRC was built to prioritize certainty.

Historically, there have been two routes to authorizing a plant. The older two-step process required a construction permit before building and then a separate operating license before running, meaning the regulatory risk hung over the project through the entire construction period, which is financially terrifying. The newer combined license, or COL, merges these so a developer can get approval to both build and operate in one proceeding, ideally resolving the big regulatory questions before the concrete is poured. Alongside these, a reactor vendor can seek design certification, an NRC blessing of a standardized reactor design independent of any particular site, so that the design questions are settled once and can be referenced by multiple projects. NuScale's achievement was reaching this design-approval milestone; it is a genuine regulatory asset even though its first project fell apart on economics.

Why does all this take so long? Because for a novel design the NRC must satisfy itself about an enormous range of questions with no prior template. How does the reactor behave in every conceivable accident scenario? What happens on loss of coolant, on loss of power, on operator error, on earthquake? For a light-water reactor the agency has decades of accumulated understanding to draw on, which is a large part of why light-water SMRs like NuScale's move faster through review than exotic designs. But for a sodium-cooled fast reactor, or a salt-cooled high-temperature reactor, or a gas-cooled TRISO-fueled design, the agency is in less familiar territory and must build its confidence more slowly. The novelty that makes an SMR technically interesting is the very thing that makes it slow to license. This is an unavoidable tension, and any developer who waves it away is selling you something.

The restart case is the mirror image and explains the sector's sequencing. When Constellation seeks to return Three Mile Island Unit 1 to service, the NRC is not licensing a new or unproven design. It is reviewing whether a specific, well-documented plant that it regulated for decades can safely resume operation after a period of shutdown. That is a serious review with real requirements around aged equipment and requalification, but it is a review of known ground. The regulator is not asking whether this kind of reactor can be safe; it already knows the answer. It is asking whether this particular unit is fit to run again. That narrower question is answerable in a shorter time, which is the regulatory reason, layered on top of the economic reasons we covered earlier, that restarts lead the sector.

There is also the matter of the fuel supply chain, which the licensing conversation often overlooks. Several advanced designs depend on a more enriched form of uranium fuel known as HALEU, high-assay low-enriched uranium, which currently has a thin and largely non-domestic supply base. A reactor design can clear every regulatory hurdle and still be stuck if it cannot secure fuel. The empiricist watches the fuel supply chain as closely as the licensing docket, because a bottleneck there can delay a fleet regardless of how the design review goes. It is one of those unglamorous constraints that does not make headlines until it is suddenly the thing holding everything up.

So when you assess any nuclear timeline, build your judgment on the regulatory posture. Ask what exactly the NRC must approve. Ask whether the technology is familiar light-water ground or novel territory. Ask whether the developer is restarting a known plant or pioneering a new design. Ask whether the fuel exists. The answers to those questions will tell you far more about when power actually arrives than any announcement date, because the announcement reflects the developer's hopes and the regulatory docket reflects reality. In this sector, reality has a filing number.