Small Modular Reactors (SMRs)

Small Modular Reactors are nuclear reactors with output typically under 300 MWe (vs. 1,000–1,600 MWe for conventional), designed for factory fabrication and modular assembly — in principle allowing faster construction, lower upfront capital, and siting flexibility (including co-location with datacenters). They are the next-generation nuclear bet being made by hyperscalers and governments.

The case is compelling in theory: a 50–300 MW reactor could be purpose-built adjacent to a large datacenter campus, eliminating transmission losses and grid interconnection delays. Factory production (analogous to aircraft manufacturing) could eventually lower costs relative to one-off, on-site construction of large LWRs. And SMRs use diverse fuel cycles and coolant types — some are designed for HALEU (High-Assay Low-Enriched Uranium), enabling higher burnup and potentially lower fuel costs.

In practice, SMRs face significant near-term headwinds. No commercial SMR has operated in the US. The regulatory path (NRC construction permit → operating license) is multi-year. Construction permit decisions for TerraPower (Natrium) and X-energy (Xe-100) are expected in H1 2026 and late 2026, respectively — but first power for these demonstration projects is 2030 at the earliest. NuScale, the only NRC design-approved US SMR developer (77 MWe module), cancelled its first project (UAMPS, Utah) in 2023 due to cost escalation, and is now pursuing a TVA partnership for up to 6 GW.

Key SMR Approaches

Key Claims

• NRC expected to issue first commercial SMR construction permits in 2026 — TerraPower and X-energy are the lead candidates. Evidence: strong (American Bar Association, DOE)
• NuScale remains the only NRC design-approved US SMR, with a 77 MWe module (uprated from 50 MWe in May 2025), and a 6 GW TVA partnership non-binding agreement. Evidence: strong (DOE, NRC)
• TerraPower's Natrium has begun non-nuclear construction at Kemmerer, WY (retiring coal plant site); NRC environmental review complete; final safety evaluation December 2025. Evidence: strong (Yale Clean Energy Forum)
• First SMR power in the US is not before 2030 — the regulatory path and construction timeline make sub-2030 commercial operation extremely unlikely. Evidence: strong (consensus assessment)
• Google signed contracts with Kairos Power and X-energy — the hyperscaler SMR bet is on factory-built, purpose-adjacent units. Evidence: strong
• NuScale is the only SMR with full NRC design certification (standard design approval granted Jan 2023) — a regulatory moat; its module is 77 MWe, scaling to 924 MWe (12 modules) in a VOYGR plant. Evidence: moderate (SEC 8-K, summary-derived, nuscale-q1-2026-context)
• Oklo targets its first reactor ~2027, ahead of most peers; NuScale's NRC certification is the reference point against which Oklo and others are measured. Evidence: moderate (multi-source synthesis, hyperscaler-nuclear-procurement; nuscale-q1-2026-context)
• Hyperscaler demand is now underwriting advanced-reactor scale-up: Meta's 6.6 GW deal alone commits 1.2 GW to Oklo (first reactors as early as 2030) and 2.6 GW + 1.2 GW storage to TerraPower (two Natrium units online as soon as 2032). Neither developer holds an NRC construction permit yet; TerraPower completed its safety evaluation Dec 2025. Evidence: weak (single news item, Latitude Media)

Benchmarks & Data

• SMR output range: 15 MWe (Oklo Aurora microreactor) to 470 MWe (Rolls-Royce)
• NuScale 6-module plant: 462 MWe total; full 12-module VOYGR plant scales to 924 MWe (nuscale-q1-2026-context); cost estimate escalated from $5.3B to ~$9.3B (2023 cancellation driver)
• Natrium: 345 MWe base output, with molten-salt thermal energy storage allowing 500 MWe peak; co-located with Wyoming coal retirement
• HALEU enrichment: <20% U-235 (vs <5% for conventional LWR); only US commercial HALEU enricher is Centrus Energy (LEU)

Open Questions

• Will factory manufacturing actually reduce costs relative to large LWRs? (NuScale UAMPS cancellation suggests not yet)
• Can SMRs get construction timelines below 5 years? (Current estimates 7-10 years)
• Does the Google/Kairos KP-FHR fleet strategy produce meaningful data on factory-build economics?
• Who supplies HALEU at scale if the US fleet of advanced reactors scales? (Centrus alone is insufficient)
• Will any SMR achieve commercial power before 2030?

Related Concepts

• Conventional Nuclear — existing fleet context
• Uranium Fuel Cycle — fuel supply, HALEU challenge
• Power Purchase Agreements — hyperscaler SMR contracts

Changelog

• 2026-06-15 — Initial compilation from web research (Yale Clean Energy Forum, WWT, ABA, DOE, NRC)
• 2026-06-24 — Compiled new sources (meta-66gw-nuclear-deal, nuscale-q1-2026-context, hyperscaler-nuclear-procurement)