There Is a Significant Gap Between Announced PPA Capacity and Contracted Capacity, with Nuclear Operators Offering Greater Contract Certainty
AI summary card
There Is a Significant Gap Between Announced PPA Capacity and Contracted Capacity, with Nuclear Operators Offering Greater Contract Certainty
Bernstein tracks AI- and data center-related PPAs using a five-tier classification, finding that only 19,495 MW of the 59,039 MW total disclosed capacity is binding, resulting in a portfolio Ghost Ratio of 3.03x.
- All tracked transactions for CEG and VST are executed PPAs, with a binding ratio of 100%, reflecting the monetization advantage of existing nuclear baseload assets.
- Headline capacity for NEE, BE, and FRVO includes a substantial proportion of framework agreements, alternative arrangements, or non-binding commitments, warranting caution over the gap between capacity narratives and deliverable revenue.
- ORA's signed projects remain contingent on approval by the Nevada Public Utilities Commission, making regulatory approval a key variable.
- Global corporate PPA contracting totaled 31.7 GW in the first seven months of 2026, annualizing to approximately 54 GW, which may be below 2025 levels; U.S. year-to-date PPA volume stands at 15.1 GW.
Report interpretation
Overview
This report establishes a PPA deal tracker covering CEG, VST, NEE, BE, FRVO, ORA, and certain pure-play nuclear/SMR developers, focusing on publicly disclosed transactions serving data centers and AI infrastructure. Its core objective is to separate total GW capacity cited in press releases into signed, legally binding contracts and projects still in development, negotiation, or the letter-of-intent stage.
Core views
AI and data centers are driving demand for stable, dispatchable power, but announced capacity should not be equated directly with near-term construction starts or recognizable revenue. CEG and VST have greater contracted certainty through existing nuclear assets, while NEE, BE, and FRVO have larger pools of non-binding projects whose subsequent value depends on conversion of framework agreements into formal PPAs. Due to faster construction timelines, natural gas remains an important pathway for meeting incremental data center load; while nuclear demand is rising, its scale remains below that of solar, wind, and storage combinations.
Analysis framework
The report aggregates publicly announced PPAs and related power-supply transactions since January 2024, tiers them by contractual bindingness, and uses the binding ratio and Ghost Ratio to measure the difference between contracted capacity and total announced capacity for each company; deal status and classifications will be updated periodically.
Methodology notes
Distinguishes executed PPAs, development agreements, framework agreements, options, MOUs, and LOIs by contractual bindingness.
Tier 1 represents binding formal agreements, while Tier 5 represents non-binding arrangements such as MOUs and LOIs, preventing long-dated project pipelines from being mistaken for near-term deliverable capacity.
Tier 1 binding capacity divided by total disclosed capacity.
A ratio of 100% indicates that all disclosed transactions in the tracker are executed, binding PPAs.
Total disclosed capacity divided by Tier 1 binding capacity.
A higher value indicates more non-binding, option-based, or conditional capacity behind each unit of contracted capacity, implying greater capacity-inflation risk.
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- Constellation Energy Corporation (CEG)Beneficiary of nuclear PPAs and data center power supply
- Strengths
- 100% binding ratio; existing nuclear baseload assets support large-scale, executable contracts; the additional 744 MW of PPAs with undisclosed counterparties may indicate an underestimated contracted backlog.
- Weaknesses
- The pace of new nuclear capacity expansion is limited.
- Comparison
- Higher contractual certainty for disclosed capacity than NEE, BE, and FRVO.
- Risks
- Contract pricing for open capacity after tax-credit expiration, delays in new nuclear capacity additions, and weaker PJM capacity prices.
- Vistra Corporation (VST)Beneficiary of nuclear power and electricity supply
- Strengths
- 100% binding ratio, with all disclosed transactions being executed PPAs.
- Weaknesses
- Sensitive to regional data center load and power market conditions.
- Comparison
- Similar to CEG and superior to development-oriented peers relying primarily on framework agreements.
- Risks
- Data center demand in ERCOT and PJM falling short of expectations, higher natural gas prices, and unfavorable hedging outcomes.
- NextEra Energy Inc (NEE)Developer of natural gas, renewable energy, and energy storage
- Strengths
- Has a multi-GW pipeline of data center-related projects spanning natural gas, renewable energy, and storage.
- Weaknesses
- A large share of capacity remains at the MOU, framework, or pre-PPA stage, with a Ghost Ratio of 5.0x.
- Comparison
- Its project pool is sizable, but binding certainty is lower than for CEG and VST.
- Risks
- Slowing demand growth, pressure on NEER generation growth, and tariff-related supply-chain risks.
- Bloom Energy Corporation (BE)Provider of on-site fuel-cell power solutions
- Strengths
- Can provide rapidly deployable on-site power solutions for data centers and has relationships with customers such as Oracle.
- Weaknesses
- There is a substantial gap between binding capacity and headline capacity, with a Ghost Ratio of 2.6x.
- Comparison
- Contract certainty is weaker than for nuclear operators and more exposed to customer deployment and project conversion.
- Risks
- Grid upgrades reducing demand for on-site power and data center demand falling short of expectations; stronger-than-expected demand and contracting activity would provide upside.
- Fervo Energy Company (FRVO)Developer of enhanced geothermal and clean firm power
- Strengths
- Approximately 658 MW of operational Tier 1 PPAs; geothermal aligns with long-term demand for clean firm power.
- Weaknesses
- Google's geothermal framework agreement for up to 3 GW has not yet converted into formal PPAs, with a Ghost Ratio of 5.6x.
- Comparison
- Like NEE and BE, it has substantial announced capacity but significant conversion risk.
- Risks
- Power demand or long-term PPA prices falling short of expectations, substitution by competing technologies, permitting and policy changes, and project operating and capital-expenditure overruns.
- Ormat Technologies, Inc. (ORA)Developer of geothermal, solar, and energy storage
- Strengths
- Has exposure to geothermal and energy storage businesses.
- Weaknesses
- Certain contracted projects will not become legally effective until approved by the Nevada PUC.
- Comparison
- Contract status is affected by regulatory conditions, providing less certainty than fully executed PPAs.
- Risks
- Approval delays or denials, intensifying enhanced-geothermal competition, and energy-storage margins falling short of expectations.
Key data
- Portfolio Ghost Ratio3.03x59,039 MW of total disclosed capacity divided by 19,495 MW of binding capacity.
- CEG and VST Binding Ratio100%All transactions disclosed by both companies in the tracker are executed PPAs.
- FRVO Tier 1 PPA CapacityApproximately 658 MWGoogle's geothermal framework agreement for up to 3 GW through 2033 has not yet converted into binding PPAs.
- Global Corporate PPA Contracting Volume31.7 GW in the first seven months of 2026Annualizes to approximately 54 GW, below 66.0 GW in 2025.
- U.S. PPA Contracting Volume15.1 GW year-to-date in 2026The full-year total for 2025 was 39.1 GW.
- Nuclear PPA Capacity2.79 GW in the first seven months of 2026Annualizes to approximately 4.8 GW, below 5.71 GW in 2025.
Impact & implications
Investment analysis should prioritize contracted capacity, counterparties, regulatory conditions, interconnection arrangements, and actual commercial operation dates rather than relying solely on announced GW figures. A declining Ghost Ratio generally indicates conversion of the project pipeline into contracted revenue; if headline capacity continues to grow without improvement in this metric, it may indicate repetitive promotion or capacity expansion lacking executability.
Risks
- Non-binding framework agreements, MOUs, or options may fail to convert into formal PPAs.
- Utility rates, co-location, and interconnection arrangements may be restricted or re-examined by FERC or state regulators.
- Nuclear, SMR, and geothermal projects face permitting, construction-schedule, technology, and capital-expenditure risks.
- Changes in data center load growth, natural gas prices, and regional power prices may alter project economics.
- Growth in headline capacity may include repeatedly disclosed or non-deliverable projects, causing demand visibility to be overstated.
What to watch
- Quarterly changes in each company's Ghost Ratio and binding ratio.
- Whether FRVO's Google geothermal framework agreement converts into binding PPAs.
- Nevada PUC approval progress for ORA/NV Energy projects, where the report expects progress in the second half of 2026.
- Further information on CEG's 744 MW nuclear PPAs with undisclosed counterparties.
- Contracting and commercial-operation progress for NEE's Google, Comstock, ExxonMobil, and Basin Electric-related projects.
- Whether global and U.S. corporate PPA contracting volumes continue the slowdown seen since the start of 2026.