Market — Nuclear & SMR Power for AI Data Centers

Verified claims and product-axis read for Nuclear & SMR Power for AI Data Centers. Every fact below is sourced; every product judgment traces back to underlying signals.

Verified facts

• Oklo announced plans to acquire Atomic Alchemy in late 2024 to expand into medical isotope production using its reactor heat. ↗ _(historical_event)_
• The U.S. nuclear fleet capacity factor averaged ~92.7% in 2023 per EIA, the highest of any generation type. ↗ _(historical_event)_
• ERCOT projects peak demand growth of ~80 GW by 2030 vs ~85 GW historical peak, driven significantly by data center and crypto loads. ↗ (other)
• Constellation reported its nuclear fleet operated at >94% capacity factor in 2023, among the highest globally. ↗ _(historical_event)_
• Holtec's SMR-300 (formerly SMR-160) is rated at 300 MWe and Holtec announced intent to deploy at Palisades and at Oyster Creek sites. ↗ _(historical_event)_
• Oklo's NRC pre-application engagement on the resubmitted Aurora design accelerated through 2024, with Oklo targeting a new combined license application in 2025. ↗ _(historical_event)_
• Kairos Power began construction on its Hermes demonstration reactor at Oak Ridge in July 2024 after receiving an NRC construction permit in late 2023. ↗ (other)
• Westinghouse's AP300 SMR, announced in May 2023 as a derivative of the AP1000, is rated at ~300 MWe and targets first deployment in the early 2030s. ↗ _(historical_event)_
• Vogtle Units 3 and 4 came online ~7 years late and ~$17B over budget vs initial estimates, costing ~$35B total for ~2.2 GW. ↗ (financial)
• TerraPower's Natrium received ~$2B in DOE ARDP funding (50/50 cost share) under the original program structure. ↗ (financial)

Cross-cutting opportunities (industry read)

• PCIe Gen5/Gen6 & CXL Retimer / Switch IC — Hyperscale AI racks need PCIe Gen5→Gen6 + CXL retimers/switches just to keep host-to-accelerator links alive at scale; this is Astera Labs' core ALAB-listed business.
• 800G / 1.6T Co-Packaged Optics & Silicon Photonics module — Scale-out beyond ~100K GPU clusters (verified Colossus class) is bandwidth-bound; pluggable optics are hitting reach/power limits, so the industry is racing to ship CPO at 800G→1.6T while still selling AEC for short reach.
• UALink AI Scale-Up Fabric (GPU-to-GPU) — NVLink lock-in is the single biggest non-NVIDIA accelerator pain point; the UALink consortium is the answer and Astera is the first merchant silicon making it ship-worthy.

See the Products and Strategy modules for the full product list and forward-looking judgment.