The 2015 dip, the only prior year fossil output contracted, came with demand growth of just 0.5%. The 2025 numbers indicate that China's clean generation base has grown large enough to simultaneously fuel economic expansion and begin retiring coal output at the margin. For grid operators and capital allocators watching the world's largest power market, the math has flipped from capacity addition to displacement.
Capacity versus displacement
The headline installation figures remain staggering but increasingly misleading as a guide to returns. China added roughly 1,700 GW of wind and solar over the past decade — more than its entire 2015 power system, which stood at approximately 1,450 GW. Going forward, project economics are migrating away from gigawatts-installed toward a narrower set of variables: siting, wholesale price capture, grid availability, storage co-location, and capacity factors. The next tranche of renewable capex will compete directly with existing coal on dispatch economics, not just serve incremental demand.
Storage economics finally move
Battery storage overtook pumped hydro in installed capacity by the end of 2024 and expanded another 84% in 2025, giving China roughly 60% of global additions. The deeper signal sits in utilisation: standalone batteries climbed from 146 equivalent full cycles in 2022 to 299 in 2025, while renewable co-located systems rose from 80 to 199 cycles. The chemistry mix is diversifying in parallel — compressed-air capacity expanded more than sixfold between 2023 and 2025, flow batteries more than tenfold, and molten-salt thermal nearly tripled. A 600 MW / 2.4 GWh compressed-air facility connected in early 2026 is now the largest of its kind worldwide. Storage is no longer a balancing accessory for variable renewable output; it is the asset class underwriting dispatchability.
Commercial viability, soberly assessed
For investors and grid planners outside China, the relevant question is no longer whether the country can build renewables at scale — that thesis is settled. It is whether China's storage and grid economics will propagate through global supply chains at prices that bend the cost curve elsewhere. The Ember data suggests yes, with a caveat: storage utilisation improvements of this magnitude only persist if wholesale price spreads widen enough to justify deep cycling. If they do, the capex case for global battery and long-duration storage shifts from speculative to arithmetic. If they don't, today's cycle counts revert to 2022 levels, and a sizeable share of installed capacity becomes stranded balance-sheet weight. That is the variable worth tracking through 2026, not the gigawatt totals.