For the 28th consecutive month, solar led every other energy source in new US generating capacity — a streak that began in September 2023 and ran unbroken through December 2025, according to FERC data reviewed by the SUN DAY Campaign. The streak survived a genuine policy hit and a measurable volume decline, and it is now being driven less by tax incentives than by two structural forces neither Congress nor the White House controls: soaring electricity demand from AI data centers, and a manufacturing bottleneck that has made new gas-fired power plants physically unavailable for years.
Utility-scale solar installations totaled 34.7 gigawatts (GW) in 2025, according to the Solar Energy Industries Association (SEIA) and Wood Mackenzie, a 16% decline from 2024’s record of 41.4 GW. The slowdown came after the One Big Beautiful Bill Act, signed July 4, 2025, ended the residential 30% solar Investment Tax Credit effective December 31, 2025. Projects that had been scheduled to come online in the fourth quarter were delayed as developers recalibrated to the new tax-credit deadlines. Yet solar still claimed more than 72% of all new US generating capacity for the full year — and the broader pipeline shows no signs of contracting.
The core reason solar keeps leading despite policy headwinds is a deployment speed advantage that no administrative decision can erase in the near term. Solar panels are modular, manufactured along parallel global supply chains, and installed without combustion infrastructure, steam turbines, cooling systems, or fuel delivery networks. From contract to electrons on the grid, a utility-scale solar-plus-storage project typically takes 12 to 18 months. NextEra Energy CEO John Ketchum said renewables and battery storage are the cheapest to build and can power the grid in that window, framing the point at his company’s Q1 2025 earnings call and at CERAWeek.
Gas-fired power plants now take five years or longer. According to the Rocky Mountain Institute, the lead time for new combined-cycle gas plants jumped from 3.5 years in 2023 to five years in 2025, with costs up 49% over the same period, per BloombergNEF data. The three companies that supply more than 75% of large gas turbines — GE Vernova, Siemens Energy, and Mitsubishi Power — have each announced delivery backlogs that push new turbine availability to late 2028 at the earliest. Mitsubishi states that turbines ordered today will not arrive until 2028 to 2030. Siemens reported a record order backlog of €131 billion. GE Vernova has committed nearly $600 million to expand US manufacturing capacity — but confirms no new units will ship before late 2028.
“Renewables and storage continue to be the fastest way to get new electrons on the grid until additional gas-fired generation can be built,” Ketchum said in comments cited by Reuters. NextEra added a company-record 4 GW of renewables and storage to its project backlog in the first quarter of 2026 alone, bringing its total pipeline to approximately 33 GW.
The electricity demand surge driven by AI data centers has become the single largest accelerant for US solar procurement. In 2024, Amazon, Microsoft, Meta, and Google collectively held 84 GW of large-scale corporate power purchase agreements (PPAs), a total that grew by more than 69% in 12 months. By 2026, those deals had accelerated further: TotalEnergies and Google signed a 1 GW solar PPA for Texas data centers covering 15 years, with a separate 1.2 GW Clearway deal complementing it. US technology companies collectively contracted for an estimated 48 GW of clean energy in 2025 alone.
These are not speculative investments. They are long-term, fixed-rate contracts that de-risk the capital stack for solar developers — tech companies accept stable power pricing in exchange for supply certainty, removing the project finance risk that has historically constrained utility-scale solar expansion. The Brookings Institution reports that hyperscalers like Google, Meta, and Amazon collectively accounted for 43% of all clean energy PPAs signed globally in 2024, making them the world’s largest corporate clean energy buyers. The International Energy Agency projects that renewables — primarily solar and wind — will supply roughly 50% of the growth in data center electricity demand globally between 2024 and 2030.
Read more: Effective Climate Tech Beyond AI Features Grid‑Scale Batteries, Smart Inverters, and HVDC Lines
Installed utility-scale solar capacity in the United States reached 164.5 GW at the end of 2025, representing 12.2% of total generating capacity, according to FERC. That figure has now surpassed the individual installed capacities of wind (161.1 GW), nuclear power (104.4 GW), and hydropower (102.1 GW). Wind and solar together account for 24.1% of total utility-scale capacity.
A critical distinction: installed capacity and actual electricity generation are not the same number. Solar plants operate at a capacity factor of roughly 23% — they produce power only when the sun shines. Nuclear plants, by contrast, run at capacity factors above 90%. This means nuclear still generates far more actual electricity than solar despite holding a smaller share of installed capacity. The comparison that matters for grid planning is how fast new capacity can be added, and on that dimension, solar has no peer in the near term.
FERC’s “high probability” forecast projects 86 GW of additional solar between January 2026 and December 2028. Should those additions materialize, solar will surpass coal — currently at roughly 173 GW — in installed capacity, likely before the end of 2026. By 2028, FERC’s projections place solar at approximately 17% of total US installed capacity, trailing only natural gas at around 40%. Meanwhile, the EIA’s January 2026 Short-Term Energy Outlook projects roughly 70 GW of new solar capacity online in 2026 and 2027 combined — a 49% increase in operating solar capacity within two years.
Global investors are following the electricity demand signal rather than the political one. Portuguese utility EDP announced in November 2025 that approximately 35% of its €12 billion three-year investment plan — roughly €4.2 billion, or about $4.6 billion — is directed toward the North American renewables market, with the US share of EDP Renewables’ underlying earnings projected to rise from 60% in 2025 to 68% by 2028. EDP’s integrated annual plan explicitly attributed the decision to AI-driven electricity demand and the competitive advantage of renewables paired with storage as the most scalable source of new generation.
Brookfield Asset Management, one of the world’s largest private equity investors in clean infrastructure, in September 2025 confirmed its 200-gigawatt pipeline of new clean energy and storage projects in the US, noting that renewable energy projects remain the most important energy solution for data centers despite federal opposition. The firm had just closed a $3 billion deal with Google to supply hydroelectric power to its data centers.
The result is a dynamic that has surprised analysts who expected federal policy to slow US renewable buildout: an administration working to unwind climate programs has presided over a solar-led capacity boom, while the European Union — the institutional champion of the green transition — has watched investment capital migrate toward the US market, drawn by AI electricity demand and the engineering economics of rapid deployment.
Utility-scale solar is now consistently the cheapest new generating technology in most US markets when measured by levelized cost of energy (LCOE) — the total lifetime cost of building and operating a plant divided by the electricity it produces. The absence of fuel costs, the predictable 25-to-30-year depreciation of photovoltaic panels, and the elimination of fuel price risk make solar’s cost floor structurally immune to monetary policy and commodity cycles in ways that gas-fired generation is not.
What locks the economics in place are the long-term PPAs. When a hyperscaler signs a 15-year fixed-rate solar contract, it is effectively converting a capital expenditure into a predictable operating expense — and providing the certainty a project finance bank needs to close a construction loan. This dynamic, now operating at gigawatt scale, has given the US solar industry a capital structure that does not depend on federal incentive certainty in the way the residential market does. The One Big Beautiful Bill Act’s elimination of the residential 30% solar credit hurt homeowners and small installers. It did not materially alter the economics of the utility-scale market, which operates on a different credit structure and which, under the 48E provision, retains access to federal support through the end of 2027.
The speed advantage that makes solar attractive to grid planners also introduces the constraint that shapes every planning decision: solar generates power only when the sun shines. A capacity figure of 164.5 GW of utility-scale solar translates to an average output of roughly 38 GW when capacity factor is applied — competitive with nuclear’s actual output but with a fundamentally different dispatch profile. Meeting baseload demand and serving the flat, 24-hour power requirements of AI data centers requires storage paired with solar, or complementary dispatchable sources.
Battery Energy Storage Systems (BESS) are scaling rapidly alongside solar. The EIA projects battery capacity in the Texas grid alone to grow from 15 GW in 2025 to 37 GW by the end of 2027, largely to smooth the dispatch profile of the growing solar fleet. The technical challenge for grid operators is not whether solar and storage can supply adequate energy over a day or week — it is whether the transmission infrastructure, interconnection queues, and ancillary services markets can handle the pace of deployment. Seven of 13 major US grid regions are projected to operate below critical safety margins by 2030 under current growth trajectories, per Brookings, as transmission infrastructure lags behind new supply.
Why is US solar energy growth continuing despite federal policy cuts?
Two forces independent of federal tax incentives are now the primary drivers of US solar deployment: AI data center electricity demand, which has generated gigawatt-scale, long-term power purchase agreements that fund utility-scale solar projects without requiring federal credit certainty, and a global gas turbine manufacturing bottleneck that has pushed delivery timelines to 2028–2030. Solar’s 12-to-18-month deployment window gives it a structural speed advantage over gas-fired alternatives that no policy change can close in the near term.
How much new solar capacity did the US add in 2025, and how does that compare to gas?
The utility-scale solar sector added 34.7 GW in 2025, according to SEIA and Wood Mackenzie — a 16% decline from 2024’s record 41.4 GW, but still more than eight times the 4.2 GW of net new natural gas capacity added in the same year per FERC data. Solar alone accounted for more than 72% of all new US generating capacity, while all renewables combined reached 88%.
When will US solar capacity surpass coal?
The SUN DAY Campaign’s analysis of FERC data projects that solar installed capacity will surpass coal — currently at approximately 173 GW — before the end of 2026. FERC’s three-year “high probability” forecast projects 86 GW of additional solar through December 2028, which would place solar at roughly 17% of total US installed capacity and well above coal’s shrinking share.
What is the gas turbine shortage, and why does it benefit solar energy?
GE Vernova, Siemens Energy, and Mitsubishi Power — the three companies supplying more than 75% of large gas turbines — currently have order backlogs stretching to 2028–2030. Utilities seeking to add power capacity in the next two to three years cannot get new gas plants built fast enough to meet demand. Solar-plus-storage projects, deployable in 12 to 18 months, are the only technology available at the required speed, which is why multiple utilities and grid planners have pivoted toward renewables not for environmental reasons but as the only practical option for near-term capacity additions.
ⓒ 2026 TECHTIMES.com All rights reserved. Do not reproduce without permission.

source