Skip to content
TE

TerraPower — AI supply-chain exposure

TerraPower · Private· Energy· United States
The quick read

The model reads TerraPower primarily as a r&d in Energy. Its most binding exposure is Grid capacity (system bottleneck #6), which it consumes rather than makes — a price-taking dependency. Its largest modeled sensitivity is a shock at Grid capacity (constraint β 70).

31
Chain weight /100
1
Parts exposed
1
Layers spanned
70
Constraint β
TerraPower across the stack
Energy

The structural read · model-generated

The model reads TerraPower primarily as a r&d in Energy. Its most binding exposure is Grid capacity (system bottleneck #6), which it consumes rather than makes — a price-taking dependency. Its largest modeled sensitivity is a shock at Grid capacity (constraint β 70).

In depth · editorial + model · written 2026-07-13

TerraPower is an advanced-nuclear developer, not yet an operator — which is why the model tags its role as R&D. It is building the Natrium design, a sodium-cooled fast reactor paired with gigawatt-scale heat storage so output can be dialled up to follow demand. The pitch is aimed squarely at the AI chain's tightest physical bottleneck: firm, carbon-free power that a data centre can lean on around the clock.

Its structural hook is optionality on the scarcest input in the whole build-out. Compute is increasingly capped not by chips but by electricity, and firm baseload that regulators and buyers will accept is in short supply — which is why AI operators have begun contracting directly for nuclear. Against that, TerraPower carries first-of-a-kind timelines, regulatory risk and no operating fleet. The model keeps its weight modest: real strategic relevance, but as a bet on future grid capacity rather than power available today.

Chain footprint by layer

Energy
100%

How it participates

R&D
100%

Every part TerraPower touches

Critical materials it leans on

High-voltage cable & XLPE insulationEnriched Uranium (HALEU)Grain-oriented electrical steel (GOES)SF6 insulating gas (sulfur hexafluoride)Uranium enrichment capacity (SWU)

Geographic concentration

Texas — ERCOT GridCentral Ohio (New Albany / Columbus)Northern Virginia (Ashburn / Loudoun)Saskatchewan (Athabasca Basin)Ireland — Dublin Hyperscale Cluster

Frequently asked

What is TerraPower's role in the AI supply chain?

The model reads TerraPower primarily as a r&d in Energy. Its most binding exposure is Grid capacity (system bottleneck #6), which it consumes rather than makes — a price-taking dependency. Its largest modeled sensitivity is a shock at Grid capacity (constraint β 70).

Which parts of the AI value chain is TerraPower exposed to?

TerraPower is mapped to 1 part of the AI value chain, most strongly Grid capacity. It sits primarily in the Energy layer as a r&d.

Does TerraPower own an AI bottleneck?

Not in the current model — TerraPower is exposed to constrained parts but sits downstream of them rather than producing them.

What is TerraPower's biggest AI supply-chain risk?

Its largest modeled sensitivity is a shock at Grid capacity (constraint β 70). 3 nodes depend on it; pressure 88/100

Who are TerraPower's closest peers by AI-chain position?

By shared chain dependencies: NuScale Power, X-energy, Kairos Power, Westinghouse Electric Company.

Go live on TerraPower

  • The interactive dependency graph and full company Nexus
  • The analyst bull / bear thesis and valuation lens
  • Live signals, today’s movers and the read-through
  • Track it in your Portfolio Cockpit — positions, P&L, valuation, thesis

model v0.7.0 · research, not advice

Chain analytics are illustrative, order-of-magnitude estimates from our model of the AI value chain — not investment advice.

as of 2026-07-17Medium confidence model v0.7.0
All companies