Why is it going to take so long to get them online? We need more clean power now.
Licensing, first-of-a-kind construction risk, and industrial supply chain gaps stack together — going from a design on paper to electrons on the grid takes a decade even when everything goes right, and none of these companies has built a reactor before.
Why it matters: Big Tech signed contracts promising nuclear power by 2030, but the realistic delivery window for new SMR designs sits well past that, leaving a gap that will almost certainly be filled by gas plants.
- The NRC's new Part 53 licensing framework — effective April 29, 2026 — aims to cut design review to 18 months, but that's just step one; construction permits, site preparation, and actual building add years on top, and no US SMR construction license has ever been issued.
- Hadron filed its very first design document with the NRC in April 2026, meaning it is at the opening move of a multi-year regulatory marathon before a shovel touches dirt.
- First-of-a-kind nuclear projects historically run 2.5x over their original schedules; NuScale's flagship Idaho project — the furthest-along US SMR — imploded entirely in 2023 when costs surged past $9 billion.
- The closest thing to a benchmark: Ontario Power Generation's BWRX-300 at Darlington, which started earlier than any US entrant, is targeting 2029 as the first grid-scale SMR in North America — and that's considered aggressive.
- SMR optimists — including the NRC itself — argue that the new Part 53 rules, combined with factory fabrication and pre-fueling before transport, can compress timelines dramatically and make a 2030 target achievable for the most advanced applicants like X-energy and GE Hitachi.
- Skeptics point to the consistent pattern of cost and schedule overruns in nuclear (NuScale's collapse being the sharpest recent example) and note that regulatory reform alone does nothing for supply chain bottlenecks and the US shortage of nuclear-trained workers.
