How much does it cost to build utility-scale solar, per watt?
Utility-scale solar, installed cost per watt · $/W · measured 2010–2024 · fitted live from 15 observations by the Many Minded cost-curve engine.
The last measured value is $1.28/W (2024, LBNL Utility-Scale Solar 2025 (median $/W-DC, real 2024$)). Over the fitted window the series falls 8.3% a year (80% interval −11.8% to −4.7%) — a halving every 8.0 years (80%: 5.5–14.3). Carried forward on that fit, 2029: $0.83/W (80%: $0.59/W–$1.16/W).
Last measured 2024. The source has not published a newer figure we have verified, so the 2 years since are projections too, not observations.
the curve
what this series measures
US utility-scale INSTALLED system cost, median $/W-DC in constant 2024 dollars — not module-only price, not residential; the 2021–24 plateau (supply chain + interconnection) is real.
the gate — what crossing this threshold changes
Open since 2019 — measured, not forecast. The threshold was crossed and has stayed crossed.
what the fit says
| quantity | value | how it is computed |
|---|---|---|
| fitted rate | −8.3%/yr (80%: −11.8% to −4.7%) | Farmer–Lafond drift over the trailing window · n=11 points, 10 years |
| halving time | 8.0 years (80%: 5.5–14.3) | implied by the fitted drift |
| curve check | decelerating (−14% → −6%/yr) | first half of the record versus the second, judged against the direction that helps this metric (falling) |
| regime break | ~2020 (−14% → −2%/yr, F=66.9) | Chow-style best single breakpoint, kept only above a conservative sup-F threshold |
Not shown for this curve, because the machinery returns nothing: Wright's law (no cumulative-deployment series for this technology).
the projection, with its interval
Log cost as a random walk with drift: the forecast variance grows with horizon (τ + τ²/m), which is why these bands widen instead of staying parallel. The middle column is the least useful number on this page; the interval is the claim.
| year | central fit | 80% interval | |
|---|---|---|---|
| 2025 | $1.17/W | $1.03/W to $1.33/W | near horizon |
| 2026 | $1.08/W | $0.89/W to $1.30/W | near horizon |
| 2027 | $0.99/W | $0.77/W to $1.25/W | near horizon |
| 2028 | $0.90/W | $0.68/W to $1.21/W | the band is already wide here |
| 2029 | $0.83/W | $0.59/W to $1.16/W | the band is already wide here |
| 2030 | $0.76/W | $0.52/W to $1.11/W | the band is already wide here |
| 2031 | $0.70/W | $0.46/W to $1.06/W | far horizon — read the interval, not the middle |
| 2032 | $0.64/W | $0.40/W to $1.01/W | far horizon — read the interval, not the middle |
| 2033 | $0.58/W | $0.35/W to $0.97/W | far horizon — read the interval, not the middle |
Projected to 2033, the horizon the whole engine uses.
questions this page answers
How much does it cost to build utility-scale solar, per watt?
$1.28/W as of 2024, the latest measured value in the series (LBNL Utility-Scale Solar 2025 (median $/W-DC, real 2024$)). The fitted trend has it falling 8.3% a year, with an 80% interval of −11.8% to −4.7%.
How fast is the installed cost of solar falling?
−8.3% a year over the fitted window, an 80% interval of −11.8% to −4.7% — a halving every 8.0 years (80%: 5.5 to 14.3 years). Fitted from 11 observations spanning 10 years.
What will the installed cost of solar be in 2029?
The central fit says $0.83/W, inside an 80% interval of $0.59/W to $1.16/W. The interval is the forecast; the middle number is only its midpoint. Bands widen with horizon because shocks accumulate — a constant-width band would be overconfident.
Has the installed cost of solar reached ≤ $1.50/W installed?
Yes — measured, since 2019. Crossing it is what the engine calls "solar cheap enough to live off-bill", and it fires: the hundred. That is an observation, not a forecast.
Is the decline in the installed cost of solar accelerating or slowing?
Splitting the record in half, the fitted rate went from −14% to −6% a year — decelerating. A Chow-style test finds a regime break around 2020 (−14% → −2% a year, F=66.9). Regime breaks, not window choice, are what dominate this method's errors — which is why every projection here carries an interval.
Where does this data come from?
LBNL Utility-Scale Solar 2025 (median $/W-DC, real 2024$). 15 observations spanning 2010–2024. Curated benchmark history, extended by a weekly authoritative fetch and by news figures fact-checked against their source before they may touch a fit. Both the observation ledger and the fitting code are public, and the engine publishes its own calibration score and its misses.
the raw numbers
Every observation behind the fit, unrounded by us and unsmoothed. This table is here on purpose: graphs make people underestimate exponential change, and the raw series beside the curve is the one correction shown to work.
| year | $/W | change |
|---|---|---|
| 2010 | $5.52/W | — |
| 2011 | $4.59/W | −16.8% |
| 2012 | $4.18/W | −8.9% |
| 2013 | $3.60/W | −13.9% |
| 2014 | $3.06/W | −15.0% |
| 2015 | $2.67/W | −12.7% |
| 2016 | $2.23/W | −16.5% |
| 2017 | $2.04/W | −8.5% |
| 2018 | $1.61/W | −21.1% |
| 2019 | $1.48/W | −8.1% |
| 2020 | $1.44/W | −2.7% |
| 2021 | $1.24/W | −13.9% |
| 2022 | $1.27/W | +2.4% |
| 2023 | $1.27/W | +0.0% |
| 2024 | $1.28/W | +0.8% |
where this comes from
- Source: LBNL Utility-Scale Solar 2025 (median $/W-DC, real 2024$).
- Basis: measured. Curated benchmark history, maintained by hand against the source, extended by figures our daily editor extracts from the news — each of which is fact-checked against its source article before it may touch a fit.
- Observations: n=15, spanning 2010–2024. Ledger last updated 2026-09-07.
- Fit: the Farmer–Lafond stochastic model — log cost as a random walk with drift over a 10-year trailing window, Student-t tails, and (for series under 15 points) a volatility floor at their drift-volatility prior. The code is readable source; the observation ledger is public JSON.
- Reuse: this compiled series and the numbers on this page are published under CC BY 4.0 — take them, cite the page. The underlying source keeps its own terms.
- Accuracy: the engine grades its own intervals in public — every feasible hindcast re-fitted on data as it stood, scored for whether the actual landed inside the 80% band. The calibration score is on the engine page, misses included.
the rest of the graph
Other ENERGY curves:
- Lithium-ion battery pack price $/kWh
- Lithium-ion battery cell price $/kWh
- Solar module price per watt $/W
- Onshore wind, levelized cost of electricity $/kWh
- Solar photovoltaic, levelized cost of electricity $/kWh
- Offshore wind, levelized cost of electricity $/kWh
Recent briefs from ENERGY, the daily record of what moved:
- Eugene joins four other Oregon cities in requiring energy scores on home listings 2026-10-30Eugene's council voted 6-1 to require a one-to-ten energy score when homes are listed for sale, bringing such rules to nearly 1 in 4 Oregonians.
- Finnish team pushes electrolyser voltage to 1,000 V to trim plant costs 2026-10-30LUT University's record 1,000 V alkaline test could cut hydrogen plant investment by 10–15%, by the researchers' estimate. It is still at the development stage.
- Senate permitting bill would let FERC push utilities to build power lines 2026-10-30A bipartisan Senate bill would end incumbent utilities' automatic right to build regional lines and give FERC penalties to enforce timelines.
- Wind approvals stay stuck at the Pentagon despite a reported deal pledge 2026-10-29Only repowerings of existing wind sites have moved, trade groups say, while about 30 projects still wait on Pentagon sign-off.
- EU's patchy battery policy leaves storage tied to imported cells 2026-10-29Uneven results from EU battery supply chain policy are slowing local cell production and keeping grid storage costs tied to imports.
- China's factories now set the pace for electrolyzer supply and price 2026-10-29China holds nearly 70% of electrolyzer manufacturing capacity, and its stacks cost about 60% less, but the cost lever is smaller than it sounds.
Every tracked curve: disk storage · human genome sequencing · transistors per chip · fastest supercomputer · DRAM memory · solid-state storage · space launch, to LEO · wheat yield · AI task horizon (METR) · frontier AI training compute · modular housing, turnkey · frontier-grade intelligence · AI video, finished minute · humanoid robot, capable unit · cheapest humanoid, list price · people still offline · people outside a mobile-broadband signal · farm labor share · healthcare spend per person
← the whole engine: build clock, cascade, calibration · the archive · how we know