VST — VST

Quote unavailable this cycle.

Investment Read as of 2026-06-06

The Read

VST is a SOLID investment. Its core value—owning nuclear power for AI—is supported by strong demand signals and a discount to peers, keeping the conviction high at P=0.74.

Bull case

• The combination of nuclear power and a retail customer base provides a stable revenue floor, hedging against swings in AI power demand [Core thesis].
• Major hyperscalers like Meta and Amazon have signed long-term Power Purchase Agreements (PPAs) for significant capacity, validating demand [Pillars 2].
• The current valuation multiple is low compared to peers, offering a clear margin of safety for investors [Pillars 3].

Bear case / what breaks it

• A cancellation or delay of a signed PPA would directly undermine the revenue visibility and raise doubts about near-term demand [Thesis-breaking triggers].
• If the company's debt-to-equity ratio climbs above 3.5x, it signals excessive leverage risk, especially if growth stalls [Thesis-breaking triggers].
• A sudden, major breakthrough in AI efficiency that drastically lowers power needs could reduce the required capacity build-out [Thesis-breaking triggers].

What the latest signal says

The recent signals confirm massive global build-out interest in data centers (e.g., AirTrunk's $30B commitment in India, Grupo Fotones' campus in Spain), validating the underlying AI power demand thesis. However, the specific regulatory signals (e.g., mandatory hearings) suggest ongoing, complex regulatory hurdles that remain a key operational risk.

Posterior history

Thesis detail

Core thesis

The cheapest way to own nuclear-for-AI: 6.4 GW nuclear inside a 41 GW fleet, with

an integrated retail book (~5M customers) that hedges margins. Signing 20-year

hyperscaler PPAs (Meta 2,600+ MW, Amazon 1,200 MW) at scarcity prices while

trading at a discount to peers (~19x).

Pillars (with priors)

1. Integrated nuclear + retail = firm AI power with a margin hedge · P = 0.78

2. Meta (2,600+ MW) & Amazon (1,200 MW, Comanche Peak) PPAs deliver · P = 0.70

3. Cheapest operator multiple = margin of safety · P = 0.72

4. Uprates add low-cost capacity (Perry/Davis-Besse/Beaver Valley) · P = 0.55

Expected news (the prior)

• PPA delivery + uprate milestones; EBITDA tracking ~$6.8–7.6B guide; capacity prices

Residual = a PPA delay/cancellation, a 4th straight revenue miss, or a debt downgrade.

Thesis-breaking triggers (→ set P near 0)

• ☐ Revenue misses in 4 of 5 quarters (already 3 of 4)
• ☐ Debt/equity rises above ~3.5x
• ☐ A signed PPA slips or is cancelled
• ☐ Durable AI inference-efficiency step-change flattens power demand

Leading vs lagging indicators

• Leading: PPA signings/delivery, uprate approvals, capacity-auction prices
• Lagging: EBITDA, debt ratios, share price

Key metrics

• PPA MW delivered vs. scheduled · EBITDA vs. guide · D/E (~2.8x) · uprate MW

Valuation anchor

Cheapest of the operators (~18–19x fwd P/E); targets ~$217–256 (mean ~$233).

Leverage ($15.8B debt) is the bear's lever.

Cross-arena sensors

B2 (PPAs/uprates/capacity), B1 (AI demand + efficiency risk), B7 (policy).

Posterior log

• 2026-05-26 · P 0.69→0.74 ↑ · UP · · https://www.navalnews.com/naval-news/2026/05/south-korea-unveils-historic-plan-to-build-first-nuclear-powered-submarine/
• {{date}} · created · — · AI-Energy-Thesis-Scaffold