Science1 distinct publisher3 min readPublished
The margin was three gigawatts. Capacity is the cheapest of the relevant numbers to move. Solar delivered 802.4 terawatt hours over January to July, about 13 percent of Chinese electricity consumption.
The Scientist · Science desk
Compiled by The ScientistSomething wrong?How this is made
Three gigawatts, set against a coal fleet of 1,285 GW, is a margin of roughly 0.23 percent [1]. It is a real crossing, and it will widen, but nothing about the operating behaviour of either fleet changed when the lines met.
The arithmetic that does change the read is the capacity factor. There are about 5,090 hours between January 1 and July 31. A fleet of 1,288 GW running flat out through all of them would produce roughly 6,550 TWh; the actual figure was 802.4 TWh, or about 12 percent of nameplate [3]. That 12 percent is a lower bound: the fleet was smaller in January than in July, so the average installed base sat below the end-July figure, and the true rate was likely closer to the low teens. Coal plants, as Matthew Davies of Swansea University put it, "can run for much longer periods" [4]. A gigawatt of solar and a gigawatt of coal are different products, and the crossover measures what China has bought rather than what dispatches at 7pm.
The generation number comes with its own denominator. If 802.4 TWh was around 13 percent of consumption, total consumption over those seven months was on the order of 6,170 TWh [2], which is the figure any future solar share has to grow against. What the output number does not tell you is how much solar was generated and then not used. The energy administration figures reported here cover installed capacity and delivered output; curtailment is absent from them [5]. Davies's list of what has to keep pace, grids and transmission and storage and demand flexibility [8], is where that missing number lives, and it is why the next gigawatt of panels may buy less usable electricity than the last one did.
Davies is careful about the emissions question, and correctly so. He calls this "a structural turning point, not the point at which solar has replaced coal" [6], and identifies the test as whether renewable generation grows fast enough both to meet rising demand and to push coal generation down in absolute terms [7]. Those are separable outcomes. Demand growth can absorb every new solar terawatt hour while coal output holds flat, and the share numbers would still look like progress. For a country responsible for around 30 percent of global planet-warming emissions [2], the absolute line is the one that matters to the atmosphere.
One more thing worth holding on to: polysilicon production is energy hungry and draws on coal-fired power, giving every panel an up-front carbon cost that its own output later repays [10]. The milestone stands regardless; it just means a build-out this fast raises coal burn before it lowers it, which is another reason to read capacity charts as a leading indicator of manufacturing scale and a lagging, noisy one for emissions.
Ranked by verification strength, evidence, and original report placement.
China had about 1,288 GW of installed solar capacity at the end of July, compared with 1,285 GW of coal-fired capacity, according to figures from the country's energy administration reported by the South China Morning Post. It is the first time solar capacity has overtaken coal.
China is the world's largest annual greenhouse gas emitter, accounting for around 30% of global planet-warming emissions.
Matthew Davies, UNESCO chair in sustainable energy technologies and a professor of chemical engineering at Swansea University in the UK, told Live Science that China reaching around 1,288 GW of installed solar capacity, marginally overtaking coal, is a major milestone.
Davies said solar only generates when sunlight is available, whereas coal plants can run for much longer periods, so solar still supplies a much smaller share of China's electricity than coal.
Between January and July this year, Chinese solar power output rose 15.5% compared with the same period in 2025, reaching 802.4 TWh, which is around 13% of the country's total electricity consumption. The reported figures cover installed capacity and output; no curtailment figure is given.
Davies said: "I would see this as a structural turning point, not the point at which solar has replaced coal."
Distinct publishers with included, body-backed reporting in this cluster.
Follow any of these and your For You feed starts watching them — no settings page required.
product
Half of severe alopecia patients on upadacitinib hit 80% scalp coverage by week 241 distinct publisher
leadership
Six months of on-off war has turned Gulf interruption into a line item, not a scenario1 distinct publisher
invest
China's crude imports fell to a 2016 low, and the self-sufficiency bill looks cheaper1 distinct publisher
science
A drug-insensitive enzyme in the mouse gut strips 80% of metformin's glucose-lowering effect1 distinct publisher
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
One government tally, twice relayed
Three numbers hold this story up — 1,288 GW, 1,285 GW, 802.4 TWh — and all three travel the same path: China's energy administration to the South China Morning Post to Live Science. No one in our coverage has read the primary release, and with a 3 GW margin the conclusion sits inside the width of a normal statistical revision. The derived readings are solid because they are just arithmetic on those same figures, which is also their limit.
Deployment is the one thing not in doubt
Whatever the reporting chain, the panels are on the ground: better than a terawatt of installed solar and 802.4 TWh delivered in seven months, growing 15.5% year on year. This is not a pilot or a pledge. The caveat is on the other side of the meter — 13% of consumption and a capacity factor near 12% describe a fleet that is vast and lightly utilised at once, and nothing here tells us how much of that output the grid actually absorbed.
The verb 'overtakes' is doing more work than the data
'Solar passes coal' invites a reader to picture a power system that has changed hands. What changed hands is a nameplate ranking by 0.23%, while coal keeps supplying the bulk of the electrons. The overstatement is real but small and largely lives in framing, because Live Science itself puts the capacity-versus-generation correction high in the piece and its expert refuses the replacement narrative outright. The unhedged claim is elsewhere: falling global panel prices, asserted with no figures at all.
Self-reported numbers, a sympathetic interpreter
The capacity tally originates with the agency of a government that gains reputationally from being seen to lead the energy transition, and it is not audited anywhere in this chain. The only interpretation offered comes from a UNESCO chair in sustainable energy technologies — a credible read, and one whose institutional standing rises with the technology's milestones. Neither is disqualifying; both are reasons a reader should want the primary release and a second voice.
Confident about the arithmetic, less about the source chain
We can be firm on what follows from the published numbers — the margin, the implied consumption total, the capacity factor floor — because that is division. Confidence drops on everything upstream: one outlet, one relay, one expert, no primary document, and no curtailment data to close the gap between generated and used. If the agency revises either capacity figure by a fraction of a percent, the headline inverts.