NRC chairman Gregory Jaczko said the agency had been right to move from a licensing process of “build first and then licence” to a system under which applicants could “licence first and then build”.

“Unfortunately, applicants have not used this process as the commission had envisioned,” Mr Jaczko said.

“The agency always understood that the early submission of completed designs would be the key to success under the new process Almost no one is following that ideal process. Instead, we are now again in a situation where applicants submit incomplete designs and less than high-quality applications for review. This is certainly allowed under our regulations, but I do not believe it is the most efficient or predictable path forward,” he added.

Mr Jaczko said he believed everyone would be better served by “focusing on the lesson of those plants that never got built and concentrating on getting designs completed first to ensure that review process proceeds smoothly”.

Another lesson of the past that has not been learned is the failure to standardise around a limited number of reactor designs, Mr Jaczko added. “We have approximately five different designs in the pool of new reactor applications before the agency. At one point, that number was closer to three. I think that would be a more manageable number. It would provide a diversity of supply without providing an unnecessary and inefficient number of reactor designs to potentially license and regulate.

“But this is not ultimately up to the agency. It is a decision for vendors and utilities to make. Regardless of the number of designs that are submitted for review, the NRC always will always meet its mission of protecting public safety, security, and the environment in reviewing new applications,” he said.