Source Report

45U Nuclear Production Tax Credit: Preserved but with Strings Attached

The Inflation Reduction Act's Section 45U provides a zero-emission nuclear power production tax credit of up to 1.5 cents/kWh (adjusted for inflation) for existing facilities through 2032, creating a effective price floor of $40-44/MWh after grossing up for power sales, which has kept Constellation's nuclear fleet profitable amid volatile wholesale prices; however, the Trump administration's One Big Beautiful Bill Act (OBBBA) in 2025 imposed Foreign Entity of Concern (FEOC) restrictions starting 2026, limiting credits for projects with ties to China/Russia/Iran/North Korea, and accelerated phaseouts for related credits, though 45U survived largely intact with marginal changes like energy community bonuses.[1][2][3]
- OBBBA retained 45U for existing plants through 2032 but added FEOC rules (e.g., no credits if >25% foreign ownership or material assistance from prohibited entities), potentially complicating supply chains for nuclear components.[4]
- Joint Committee on Taxation estimates key IRA credits (including 45U) now worth ~$280B through 2034 vs. $490B pre-OBBBA, signaling fiscal tightening without full repeal.[2]
- Republicans like Rep. Dan Meuser called repealing 45U a "mistake" for baseload support, reducing outright repeal risk, but ongoing budget fights could cap extensions.[5]

For competitors or entrants, this means nuclear uprates/restarts remain viable only with domestic supply chains; monitor Treasury FEOC guidance for enforcement rigor, as violations could slash ~20-30% of fleet economics overnight.

Capacity Market Reforms: Tight Supplies Boost Near-Term Revenue but Reforms Cap Upside

PJM's capacity auctions hit caps ($329/MW-day for 2026/27, up 22% YoY) due to retirements, data center load growth outpacing new entry, and interconnection delays, driving ~$170B in transmission needs over 10 years and benefiting Constellation's ~90% nuclear fleet; however, FERC-approved reforms (e.g., price floors/caps at $175-333/MW-day, reliability backstops) and proposals like sub-annual markets or data center auctions could flood supply, eroding scarcity premiums that comprise 20-30% of revenue.[6][7][8]
- ERCOT (energy-only, no capacity market) saw 22% energy price rise to $33/MWh in 2025; post-Calpine, Constellation gains gas peakers but faces scarcity-driven volatility without formal capacity payments.[9]
- NYISO less detailed, but regional trends mirror PJM: 2025/26 auctions cleared at caps amid 5.4GW load growth vs. 2.7GW new entry.[7]
- Trump-era emergency auctions for data centers (15-year contracts) risk new gas/solar buildout, pressuring incumbents like Constellation to bid against their own customers.[10]

New entrants should hedge via bilateral PPAs over auctions; watch FERC dockets (e.g., ER24-99, ER25-682) for backstop costs allocated to utilities, potentially passing through to loads and compressing margins.

Calpine Integration: Antitrust Divestitures Resolved but Debt and Synergies at Risk

Constellation closed its $26.6B Calpine acquisition on Jan 7, 2026, creating the largest U.S. generator (~55GW, 10% clean share), but DOJ/ FERC required divesting ~6GW (e.g., $5B PJM gas plants to LS Power), raising ~$100M+/yr in consumer costs via reduced competition in ERCOT/PJM Mid-Atlantic; integration risks include $22% YoY op-ex jump to $5.48B in Q4 2025 and deleveraging focus post-deal, delaying growth.[11][12][13]
- DOJ suit alleged 12% ERCOT share enables withholding for higher prices; divestitures (York 2, Fusco, etc.) mitigate but signal scrutiny for future M&A.[14]
- Seeking Alpha "Sell": Debt/dilution, no 2026 guidance, prioritizes deleveraging over AI upside.[15]
- S&P affirms 'BBB+' but notes diversification reduces nuclear concentration risk.[16]

For peers, this sets precedent for DOJ structural remedies even <30% shares; track divestiture proceeds (~$5-7B) deployment—if debt paydown > growth capex, thesis weakens.

Three Mile Island (Crane) Restart: Grid Delays Threaten Microsoft PPA Economics

Constellation's $1.6B restart of TMI-1 (renamed Crane Clean Energy Center, 835MW) ties to a 20-year Microsoft PPA for full offtake at ~$110-115/MWh, but PJM studies peg full grid deliverability to 2031 due to delayed 765kV/500kV lines (to WV), risking capacity forfeiture and PPA penalties unless FERC grants waivers for transferred rights.[17][18][19]
- NRC review on track for 2027 (safety/environmental), but PJM opposes waivers; market monitor flags reliability.[20]
- Stock fell 3-9% on 2031 news; history of overruns (e.g., transformers $75-100M) adds capex risk.[21]
- $1B DOE loan shifts taxpayer risk, but former regulators doubt full shutdown restarts.[22]

Entrants face similar queues; monitor FERC waiver (docket pending) and PJM RTEP ($11.8B upgrades)—2030+ delays break PPA value.

Analyst Bears, Shorts, and PPA Hurdles: Execution Over Hype

Bears cite 42x forward P/E demanding perfection (e.g., TMI on-time), Calpine debt, no new hyperscaler deals, and policy floods (e.g., new supply eroding prices); short interest ~2% float (down 20%), no Hindenburg/Muddy Waters reports, but downgrades (Mizuho/KeyBanc to $330) flag guidance gaps.[23][15][24]
- PJM monitor opposes TMI waivers; FERC reviews large-load PPAs for grid impacts.[20]
- Commodity exposure: Nuclear hedges/ZECs floor prices, but gas from Calpine volatile if forwards drop.[25]

Key Variables to Monitor for Thesis Breakdown

Track these for bull thesis erosion (e.g., stock re-rates on misses):
- PJM/FERC dockets: TMI waivers, capacity reforms, data center auctions—shortfalls >6GW signal supply flood.[7]
- 45U/OBBBA implementation: FEOC violations or phaseout acceleration; Treasury guidance by Q3 2026.[3]
- Calpine metrics: Q2 2026 synergies ($Xbn), net debt/EBITDA (<4x), divestiture closes.
- **TMI milestones**: NRC safety sign-off (Q1 2027), grid interconnection (PJM study updates), capex vs. $1.6B.
- **Power prices/contracts**: PJM West day-ahead ($50/MWh 2025), new hyperscaler PPAs (Meta-scale), hedge coverage.
- **Ops/CFO**: Nuclear capacity factor (>94%), outage costs, CFO volatility (negative = red flag).[26]

Misses here (e.g., TMI to 2030, op-ex +20%) could halve valuation from 42x P/E; beats sustain premium.

---

Recent Findings Supplement (April 2026)

Calpine Integration: DOJ Antitrust Divestitures Force $5B Asset Sale, Raising Execution and Debt Risks

Constellation completed its $26.6B acquisition of Calpine on January 7, 2026, creating a 55 GW fleet (largest U.S. wholesale power producer), but DOJ and FERC required divestitures of overlapping gas plants in PJM (4.4 GW sold to LS Power for $5B on March 18, 2026) and ERCOT (e.g., minority stake in Gregory plant) to mitigate market power concerns—mechanism: post-merger, Constellation's nuclear baseload would enable profitable withholding of Calpine's higher-cost gas peakers, raising ERCOT/PJM prices by $100M+ annually. This dilutes near-term revenue from high-margin gas assets while adding integration costs and $4.5B cash/debt from deal.[1][2]

• DOJ filed antitrust suit December 5, 2025, settling same day with divestitures beyond FERC's July 2025 PJM-only requirements (Hay Road, Edge Moor, Bethlehem, York 1/2).[3]
• FERC upheld merger approval February 19, 2026, rejecting rehearing on market power risks but mandating one-year notice for future retirements.[4]
• Calpine adds 27 GW (mostly gas/geothermal), but S&P affirmed ratings January 12, 2026, citing deleveraging priority over growth.[5]

Implications for competitors/entrants: New 55 GW scale strengthens hyperscaler PPA negotiations but exposes to FERC/DOJ scrutiny on withholding; smaller IPPs gain from divestitures but face higher barriers matching nuclear+gas diversity.

Three Mile Island (Crane CEC) Restart: PJM Grid Delays and Monitor Opposition Threaten 2027 Timeline, Microsoft PPA

Constellation seeks FERC waivers to transfer 760 MW Capacity Interconnection Rights (CIRs) from retiring Eddystone gas plant to Crane (ex-TMI Unit 1, 835 MW restart for Microsoft 20-year PPA), as PJM transmission upgrades (765/500-kV lines) slip to 2030/2031—mechanism: without waivers, Crane can't bid full capacity into June 2026 PJM auction despite 2027 readiness, risking PPA breach and $1.6B investment. PJM monitor opposes April 2026, citing harm to third parties; FERC decision by June 1 critical.[6]

• March 31 FERC filing: Delays from PJM feedback could push grid tie to 2031; seeks June 1 approval for 2028/29 auction bid.[7]
• Monitor (Monitoring Analytics) April 21: Fails FERC waiver criteria (no good-faith error, harms markets); PJM neutral but notes future studies needed.[8]
• Water permit sought April 2026 for 73M gallons/day from Susquehanna River.[9]

Implications: Delays erode nuclear restart economics (no capacity revenue 2028+); entrants face similar PJM queue bottlenecks, favoring incumbents with waiver leverage.

PJM Capacity Markets: Price Caps ($325/MW-day) Curb Revenue Upside Amid Record Auctions

PJM 2027/28 Base Residual Auction cleared December 17, 2025, at $333.44/MW-day cap (RTO-wide record, 14.8% reserve vs. 20% target; 6.6 GW short), yielding Constellation ~$2.2B from 17,950 MW cleared—but FERC/PJM seek extension to 2028/29-2029/30 auctions ($325 cap proposed, floor $175), mechanism: caps prevent scarcity pricing ($530+ uncapped) from data centers/AI, protecting consumers after 2025/26 hikes (PA households +$800 over 4 years via Shapiro caps). Caps compress uncontracted nuclear margins.[6][10]

• 2026/27 BRA: $329.17/MW-day cap; prices up from $28.92 (2024/25) on retirements/delays.[11]
• PA Gov. Shapiro/FERC approved cap extension April 2026, saving $800/household; governors' January 2026 principles push emergency auction by September 2026 (hyperscalers bid 15-year PPAs).[12]

Implications: Caps floor revenue at ~$325 vs. scarcity potential, hitting merchant exposure; new entrants need hyperscaler funding to compete.

45U PTC: Preserved in OBBBA but FEOC Rules Add Compliance Burden

One Big Beautiful Bill Act (OBBBA, July 2025) shortened 45U phaseout to 2035 (vs. emissions-based), preserved vs. repeals for solar/wind/hydrogen, but added Prohibited Foreign Entity (PFE/FEOC) restrictions (e.g., China/Russia-linked ownership >25% disqualifies from 2026)—mechanism: Treasury Notice 2026-15 safe harbors require supply chain audits, raising costs for nuclear ops like Constellation's 22 GW fleet (~$0.015/kWh base +5x PWA adder). No Trump repeal; bipartisan support continues.[13][14]

• IRS February 2026 FEOC guidance: Applies to 45U/45Y/48E; projects pre-July 4, 2026, exempt full value.[15]
• OBBBA: Nuclear untouched amid IRA cuts; extends to 2032 base + phaseout.[16]

Implications: Compliance hikes capex 5-10%; foreign-free nuclear moat favors U.S. incumbents.

Bearish Analyst Views and Stock Pressure: Valuation/Execution Fears Mount

CEG down 25% from 52-week high ($413 Oct 2025) to ~$297 (April 30, 2026); Mizuho/others cut PTs (e.g., $300 from $330 April 1) on no new hyperscaler deals, regulatory fog, Crane delays—P/E 49x vs. peers 22x; 2026 EPS guide $11-12 midpoint misses $11.60 consensus. No short reports, but Seeking Alpha "Sell"/"Hold" cite debt/CapEx/free cash erosion post-Calpine.[17][18]

• Q4 2025: Revenue flat, FCF -$5B on CapEx; stock -14% YTD March.[19]
• Trefis: -21% avg drawdown in crises, -47% on geopolitics.[20]

Implications: High multiple vulnerable to misses; watch for PPA announcements to re-rate.

Key Monitoring Variables for Thesis Breakdown

Track quarterly: Crane waiver/FERC rulings (June 2026), PJM 2028/29 auction results/caps (post-July), new hyperscaler MW contracted (aim 5GW+), Calpine synergies/debt metrics (net debt/EBITDA <3x), 45U FEOC compliance costs. Breakdown if Crane >2028, caps hold <$300, no PPAs by Q3, EPS misses $11. Threshold: Capacity revenue <20% prior peaks signals merchant weakness.