New gas plants are widely proposed for AI demand, but prime turbine slots are heavily reserved into 2031.
65%likely
Well sourcedNext check
Watching
Why it is different
The measured bet is that solar generator MW plus battery power-capacity MW outbuild gas generator MW through 2028. It is a capacity comparison, not a claim that batteries generate electricity.
Why it matters
It decides what the US grid is actually built out of for the rest of this decade.
Next check · 31 Jan 2027
Does GE Vernova report a gas-turbine order book of 125 GW or more at the end of 2026, with factory capacity still around 20 GW a year?
How we scored it ·
from how often this kind of thing happens
If right, how big4/5How overlooked3/5Upside vs downside2/5Source trust4/5Independent sources4/5
The exact wording we score
In each calendar year 2026, 2027, and 2028, the sum of US solar-PV generator net-summer MW additions and battery-storage net-summer MW additions exceeds natural-gas generator net-summer MW additions in EIA's final generator inventory. Battery MW is explicitly a power-capacity deployment measure, not electricity generation, and co-located solar and battery units are counted separately by technology as EIA records them. Resolves YES only if all three annual comparisons hold; NO on the first year gas is equal to or greater than the solar-plus-battery sum. Final adjudication occurs by 31 Mar 2029. Turbine backlog, energy supplied, and LCOE are mechanisms only.
PROOF
Evidence and provenance
The links below are the evidence recorded when this forecast was scored. Each note states source quality or bias; a citation is not an endorsement.
Recorded sources · 4 citations / 4 independent origins
2026-07-28 | sell-side: gas-bridge demand and equipment scarcity are both recognised; mainstream press: announced gas projects dominate the narrative; priced market: turbine-OEM rerating reflects part of the scarcity; specialist: backlog and deployment data show solar-plus-storage is what can arrive on the required clock | E3
Revision history4 entries
Probabilities and interpretations are never silently overwritten. This is the append-only record of what changed and why.
Domain review revision: P 75→65. The claim no longer says batteries “supply” generating capacity. It now names EIA's technology-specific net-summer MW measure and states that battery MW is storage power capacity, not energy generated. claimtype is corrected to conjunction and the final probability follows the documented 0.97 × 0.86 × 0.78 arithmetic.
(2nd pass) — Shared-kill-vector checkpoint added: behind-the-meter gas. 'Behind-the-meter gas at data centers' appears in the bear case of THREE signals (this one, datacenter-grid-gear, and the former china-electron-gap) in near-identical language, yet no signal in the book tracked it — every indicator measured its own mechanism and none measured the one event that weakens all of them. That was the largest monitoring gap in the base. Also flagged: this signal's GE Vernova backlog checkpoint resolves off the same filing as grid-gear's. Open tension recorded, not resolved: grid-gear says interconnection queues jam the pipeline; this signal calls the solar+storage pipeline 'buildable' at P=98 — solar+storage needs the same transformers, switchgear and queue as gas. Both cannot be fully right about how binding the interconnect is. corrgroup → ai-capex-physical.
v2 rescore. P 70→75; the backlog figure this file cites is badly stale. GE Vernova's Q2-2026 call (2026-07-22) put the gas-turbine backlog at 116 GW, not the ~80 GW cited here — up from 100 GW in Q1. GEV is taking reservations for 2031 deliveries, expects to be more than half-contracted for 2031 by end-2026, guides to 125 GW of backlog-plus-reservations at year-end, and still produces ~20 GW/yr with 30 GW/yr not arriving until 2030. Independently, the July-2026 STEO has US gas-fired capacity reaching only 508 GW by end-2027, +3% vs 2025 — roughly 7-8 GW/yr of gas additions against ~86 GW of total 2026 additions.
created from research sweep; re-tagged horizon near (resolves by 2028). Edge trimmed to 3.
Method challenge
Adversarial review record
Admitted after revision
Structured internal role review tied to this frozen claim. It is not independent human peer review. Independent specialist review not yet performed
systems or deployment specialist
measurement and data reviewer
policy or geopolitical reviewer
forecasting-method reviewer
Strongest specialist challenge
Battery MW is storage power capacity rather than electricity generation, and summing it with solar MW can double-count co-located project capability in ordinary-language claims about power supplied. The original title and claim therefore overstated the physical interpretation.
Outside view
EIA's 2026 development pipeline gives solar plus battery MW a large lead over gas, but the forecast requires three consecutive annual wins. Policy change and project attrition make 2028 materially harder than the nearly banked first year.
Causal rival
Behind-the-meter gas can expand outside the compared inventory, turbine production can ramp, solar projects can miss interconnection, or tax changes can cancel the 2028 queue. Any of those can invalidate the narrative before the EIA comparison flips.
Measurement risk
The revised claim explicitly compares technology-specific EIA net-summer MW additions and labels battery MW as storage capacity. Co-located units are counted separately exactly as EIA records them; energy generated or firm capacity cannot be substituted.
Residual risk
Final generator inventories can revise prior-year additions, and the shared transformer or interconnection constraint also affects solar. The three-year conjunction must use one frozen EIA vintage at final adjudication and disclose any revisions.
COLDTELL