GOJO · CEG

CEG: The AI Power Compounder

Constellation just doubled in size, locked Microsoft into a 20-year nuclear PPA, and guided 23% EPS growth. The trade is the entry — where you scale equity, and how you structure LEAPs.

The Snapshot

  • Price: ~$275 (May 14, 2026)
  • Market Cap: ~$86B
  • 2026 EPS Guide: $11.00–$12.00 (+23% YoY at midpoint)
  • Q1 2026 EPS: $2.74 vs $2.14 prior year (+5% beat)
  • Q1 2026 Revenue: $11.12B vs $6.79B prior year (+64% YoY, +28% beat)
  • Generation Capacity: ~60 GW post-Calpine (largest in the US)
  • P/E (2026E mid): ~24x
  • Anchor Contracts: Microsoft 20-yr PPA (835 MW Crane Clean Energy Center), Meta, CyrusOne, US government

Why This Is The Setup

This is not a "find the mispricing" trade like GRPN. CEG is widely owned, well-known, and not cheap on screen. The edge isn't a hidden asset — it's duration. Three independent compounders just got bolted onto the same chassis, and the market is still pricing CEG like a utility.

1. Calpine Closed — The Engine Doubled

The $22B Calpine acquisition closed January 7, 2026. That move took CEG from a nuclear-pure-play to the largest US electricity producer at ~60 GW. Q1 revenue jumped 64% YoY to $11.12B in the first full quarter as the combined entity — and management reaffirmed the $11–12 EPS guide afterward. The deal isn't a thesis anymore; it's printing.

What you get: 23 GW of natural gas + geothermal added to a nuclear-heavy fleet, plus immediate expansion into Texas and California — the two highest-growth data center load zones in the country.

2. Microsoft Locked In For 20 Years

The Three Mile Island Unit 1 restart — now the Crane Clean Energy Center — is contracted to Microsoft for 20 years at 835 MW. Targeted online date: 2027, pulled forward a full year from the original schedule. That's $16B+ of contracted revenue from a single hyperscaler before any market-clearing risk.

Layer on Meta, CyrusOne, and the US government as additional long-term PPA counterparties. The contracted backlog is what changes the cash flow profile from "utility" to "infrastructure-grade compounder with multi-year revenue visibility."

3. The 45U Nuclear PTC Tailwind

The 45U production tax credit for nuclear is the structural reason CEG's earnings compound — it sets a floor on nuclear generation economics and accelerates IRR on every restart and uprate. Combined with the merchant power exposure into a power-constrained grid, the operating leverage on every incremental megawatt is asymmetric to the upside.

The Equity Entry

$275 is near highs. You don't go all-in here — you scale. Three zones to layer in, with a clean invalidation level below.

  • Starter (1/3): $265–280 — current zone. This is the "I want exposure if it never pulls back" tranche. Don't overthink it; momentum names rarely give you a clean entry once the thesis is paying.
  • Core (1/3): $240–255 — first meaningful pullback. Roughly a 10–13% drawdown, which the stock has given on every macro flush in the past 18 months without breaking trend.
  • Patience (1/3): $210–225 — a real correction. Reserve for an AI-capex panic, broad market drawdown, or hyperscaler-spend headline shock. This is where conviction buyers get paid disproportionately.
  • Invalidation: Below $195 — the post-Calpine close breakout zone. A break there says either the AI/nuclear narrative is structurally repriced or something specific to CEG is wrong.

The LEAP Entry

The methodology stays consistent: deep ITM, ~delta 0.75–0.85, minimal extrinsic value, paying for direction rather than time. CEG's $11–12 EPS guide and contracted PPA backlog give the rare combination LEAPs love — high implied earnings growth + visible cash flow + a hard floor on the downside narrative.

  • Expiration: Jan 2027 (primary) or Jan 2028 if liquid at your strikes. Jan 2027 gives ~20 months of runway through the Crane online date and at least one more full earnings cycle on the Calpine integration.
  • Strike (deep ITM, ~0.80 delta): $200–220 strikes. Maximum delta, minimum theta, behaves like stock with built-in leverage. Best risk-adjusted exposure.
  • Strike (moderate ITM, ~0.65–0.70 delta): $230–245 strikes. More leverage per dollar, more sensitivity to multi-month chop, more decay if the stock sits.
  • When to buy: Wait for the $240–255 equity zone. Buying LEAPs in the $265–280 range works but you're paying peak premium. Patience matters more on the option than on the stock.
  • Stop-out trigger: Close below $210 on the underlying or a 40% drawdown on the contract premium — whichever comes first.

The Risks

  • AI capex narrative cracks. If Microsoft, Meta, or the broader hyperscaler cohort signals a meaningful slowdown, CEG re-rates with them. Watch hyperscaler capex guides quarterly.
  • Multiple compression. 24x forward is reasonable for the growth, but utility-sector P/E compression from a rate spike or sentiment shift could clip 15–20% with no fundamental change.
  • Calpine integration cost overruns. $22B deals don't integrate smoothly. Synergy targets miss, one-time charges land. Q2/Q3 prints are the tells.
  • Regulatory risk on nuclear. Low-probability but high-impact. Any nuclear incident — even at a competitor's facility — re-prices the entire sector overnight.
  • Beta to broader market. CEG got crushed in the early 2025 AI-stock drawdown along with everything else in the trade. It moves with the AI complex on the way down, not just on the way up.

The Trade

  • Setup: Multi-year compounder. Contracted revenue backlog + Calpine engine + nuclear PTC + AI demand curve.
  • Equity Sizing: Core position. Scale-in over the three zones above. Treat as a 3–6% portfolio holding, not a swing trade.
  • LEAP Sizing: Speculative-to-medium sleeve on top of equity. Jan 2027 $200–220 strikes for cleanest exposure. 1–2% of portfolio in premium, sized so a total LEAP loss is survivable.
  • Near-term Target: $310–340 (27–29x 2026 EPS, AI premium intact, no narrative change).
  • Bull Case (12–18 months): $380–420 on 2027 EPS scaling toward $13–14 + PPA ramp on schedule + Crane online.
  • Invalidation: $195 close on equity. AI capex slowdown headline + multiple compression simultaneously.

The Verdict

CEG isn't a value trade and it isn't a squeeze trade. It's the cleanest publicly-traded expression of "AI demand needs electrons" — and after the Calpine close, it's the largest one. The contracted PPA backlog turns what used to be a merchant power gamble into a multi-year revenue visibility story.

The entry discipline is what matters. Don't chase the $275 zone with size — scale in across the three levels, anchor LEAPs at the $200–220 strike when equity pulls back to $240–255, and respect the $195 invalidation. The thesis is right; the execution is everything.