TerraPower in deal with Meta for eight Natrium 345 MW nuclear plants

Reactor Design & Technology

  • Natrium is described as a sodium-cooled fast reactor with molten salt used for energy storage, not as a full molten-salt-fuel reactor.
  • Some participants conflate sodium and “molten salt”; others clarify sodium is a liquid metal coolant while heat is stored in molten salt “thermos” tanks.
  • Comparisons are made to prior troubled sodium fast reactors (Superphénix, Monju), with concerns about sodium’s reactivity, corrosion, and long-term maintainability.
  • Several commenters argue thorium/molten-salt (LFTR/MSR) designs are inherently safer and more scalable, but acknowledge they remain mostly experimental.

Regulation, Timelines, and Feasibility

  • Many doubt the claim that a demo reactor in 2031 can be followed by 8 commercial units by ~2032, citing historic nuclear delays.
  • One side argues the new combined construction/operating license should prevent “built but never operated” plants; others note this is unproven and previous “streamlined” processes also lagged.
  • NRC approval so far is only for the demonstration construction permit; further operating licenses and safety checks remain before fuel loading.

Economics, Alternatives, and Energy Mix

  • Some see this as effectively “Meta invests in a nuclear startup,” possibly more PR than deliverable capacity.
  • Others argue even if it flops, it’s a better use of Meta’s money than VR/Metaverse, and any push for abundant clean power is positive.
  • Multiple comments stress that solar, wind, and storage are already cheaper and improving rapidly; they doubt solid-fuel SMRs will stay competitive.
  • A few suggest fusion might arrive before these reactors are online, but this is treated skeptically (“20 more years”).

Meta’s AI, Power Demand, and Financing

  • Questioning how Meta can afford huge AI and power buildouts given limited visible AI products.
  • Replies: Meta monetizes AI via ads, recommendations, internal tooling, and has consumer models (Llama, Meta AI) even if not dominant.
  • Some suggest Meta will sell datacenter/AI capacity to others, exploit tax incentives, and possibly use special-purpose vehicles to raise debt off-balance-sheet (details remain unclear).

Risk, Ownership, and Public Interest

  • Strong NIMBY reaction: privatized profits vs socialized nuclear risk; demands that such plants be “in the billionaire’s backyard.”
  • Others counter that modern reactor designs are very hard to catastrophically fail and that coal/gas harm is far larger.
  • Concerns raised about private control, potential diversion of byproducts toward weapons, and desire for strictly regulated, utility-like ownership.
  • Another angle: most nuclear and much of the broader nuclear/arms ecosystem is already heavily privatized; Meta is just one more actor.

Performance, Reliability, and Water Use

  • Slogan-like “24/7/365” claims are challenged; critics note no operating history and point out new designs rarely hit mature uptime initially.
  • Proponents cite existing nuclear fleets achieving >90% capacity factors with planned refueling outages; skeptics say this isn’t guaranteed for a new sodium fast design.
  • Brief worries about increased freshwater usage; no detailed discussion of Natrium’s specific water impact.

Policy and Geopolitical Asides

  • Side debate about Europe’s perceived lack of tech investment; others say this is exaggerated and influenced by anti-EU propaganda.
  • Mention of large DOE funding for enriched uranium (LEU/HALEU) to support a “nuclear renaissance,” with some concern over the scale of public subsidy.
  • Comparison to loosely regulated off-grid gas plants being fast-tracked for datacenters, contrasted with nuclear’s tight oversight.