Autoresearch: Nuclear for AI data centers — May 2026 update

Generated by /autoresearch on 2026-05-26. 2 rounds of search (early exit). All content from search snippets (WebFetch 403). Treat as raw material. Context: vault/projects/stock-market

Summary

The nuclear-for-AI-data-center thesis has structurally strengthened since May 20. The cumulative hyperscaler nuclear commitment has reached 9.8 GW across 13 announced projects. A new deal not previously in the wiki: Google signed with NextEra Energy to restart Iowa's 615 MW Duane Arnold nuclear plant for AI data centers, with power delivery expected by 2029. Cameco CEO Tim Gitzel (May 5, 2026) confirmed data centers are "bringing dead reactors to life" at unprecedented speed; Cameco signed a $2.6B uranium supply deal with India. SMR timelines remain 2028–2030+. The nuclear-baseload-for-ai-data-centers concept (CCJ, CEG, BWXT) remains high conviction with no new contradicting evidence.

Findings

Cumulative hyperscaler nuclear landscape: 9.8 GW committed

As of May 2026, all four major hyperscalers have signed nuclear deals totaling >9.8 GW across 13 projects (smrintel.com, 2026):

Company	Deal	Capacity	Status
Microsoft	Three Mile Island Unit 1 restart (Constellation)	835 MW	$16B 20-year PPA; expected online 2027
Meta	TerraPower Natrium + Oklo Aurora + Vistra + Constellation	Up to 6.6 GW	Purchases begin late 2026; capacity scaling through 2034
Amazon	Susquehanna nuclear plant (Talen Energy, PA)	1,920 MW	17-year PPA through 2042
Google	Kairos Power Hermes (Oak Ridge, TN)	50 MW → 500 MW	SMR PPA by 2030
Google	Clearway Energy Group	1,200 MW	Wind/nuclear mix
Google	Duane Arnold restart (NextEra, Iowa)	615 MW	PPA signed; power by 2029

New since May 20: Google-NextEra Duane Arnold restart

Google has signed a deal with NextEra Energy to reopen the Duane Arnold Energy Center (615 MW, Iowa), shut since 2020 due to storm damage. Expected to deliver power to Google data centers by 2029. (Data Center Dynamics, 2026)

This is the second major nuclear plant restart signed by a hyperscaler (after Microsoft's TMI Unit 1), confirming a pattern: existing brownfield nuclear restarts are faster to market (2027-2029) than new SMRs (2028-2030+), and hyperscalers prefer guaranteed delivery timelines.

Implication: NextEra Energy (NEE) benefits from the restart contract — not currently in the nuclear-baseload-for-ai-data-centers suggested-tickers (which shows CCJ, CEG, BWXT). NextEra is a potential ticker to add.

Cameco CEO (May 5, 2026): uranium demand unprecedented

Cameco CEO Tim Gitzel: "The AI, the data centers, the hyperscalers. It's demand increase like we haven't seen." Data centers are "bringing dead reactors to life." (BNN Bloomberg, May 5, 2026)

Cameco $2.6B India deal: 22 million pounds of uranium supply to India (stocktitan.net, May 2026). CCJ Q1 net profit: $131M on $845M revenue.

Global uranium context: 78 GW of nuclear capacity under construction globally. 38 nations at the 2026 Paris summit pledged to triple nuclear capacity by 2050 — sovereign fuel commitments locking in long-term uranium demand.

SMR status: still 2028–2030+ timelines

SMR Developer	Status	Timeline
TerraPower Natrium	NRC environmental review complete (Oct 2025); final safety evaluation (Dec 2025); NRC permit decision expected H1 2026	Wyoming commercial: targeted 2028, but HALEU fuel scarcity may push to 2030+
Kairos Power Hermes	DOE HALEU TRISO fuel contract signed; Hermes 1 startup fuel secured	50 MW online 2030 (Google PPA)
X-energy Xe-100	Dow Chemical Texas site selected; NRC safety evaluation complete November 2026	Commercial: 2028–2030
Oklo Aurora	Meta deal signed; backed by OpenAI's Altman	2027–2028 target; regulatory path partly uncharted

Note: HALEU (high-assay low-enriched uranium) fuel supply is the binding constraint for most SMR timelines. TerraPower explicitly cited fuel scarcity as the reason their 2028 target may slip to 2030+.

New hypothesis: data center co-location premium near nuclear sites

One source noted data center land within 20 miles of operating or restarting nuclear plants commands meaningfully higher real estate valuations due to grid interconnection advantages. This is a picks-and-shovels angle: the nuclear-adjacent real estate market is a 2nd-order beneficiary of hyperscaler nuclear procurement rush. Not in the current wiki.

Contradictions and open questions

• Duane Arnold restart was new since May 20: Not in prior wiki entries. Update nuclear-baseload-for-ai-data-centers to include the Google-NextEra deal.
• NextEra Energy (NEE): Benefits from Duane Arnold restart contract but is not in suggested-tickers. CEG (Constellation) and CCJ are current. NEE is the cleaner pure-play on plant restarts if Google-style deals multiply.
• HALEU fuel scarcity: The binding constraint for SMR deployment timelines. If HALEU supply doesn't scale, the 2028 SMR delivery commitments slip significantly. Watch DOE HALEU commercial deployment program.
• SMR economics unproven at scale: All SMR deals are PPAs for capacity that doesn't exist yet. The nuclear baseload thesis is on firmer ground with existing plant restarts (TMI, Duane Arnold) than with SMR promises.

Provenance

Rounds run: 2 of 3 (early exit — sufficient data; no new announcement specifically post-May 20 other than the Duane Arnold deal confirmation)

Sub-questions by round:

Round 1 (broad survey):

1. New hyperscaler nuclear PPAs since May 20 2026?
2. SMR deployment status — Kairos, X-energy, TerraPower?
3. Uranium demand signals from nuclear plant restarts?

Round 2 (drill-down):

1. Nuclear plant restart deals — any new since May 20? — targeted gap: recency confirmation
2. CCJ/Cameco uranium demand from data centers — targeted gap: quantified demand signal

Anchor source: No Grokipedia entry (fast-moving corporate topic).

URLs fetched: 0 (all HTTP 403). Content from search snippets only.

Tools used: WebSearch (4 queries). WebFetch: all 403. Generated: 2026-05-26