Ecology & Planetary Boundaries
Leading in the expansion of data centres, networks, and renewables. Compute hardware remains partly import-dependent; coal remains a bridging fuel.
Where does China stand, and where might it break?
In 2024, China generated approximately 12.6bn tonnes of CO₂ equivalent, accounting for roughly 30% of global energy-related emissions (IEA / Carbon Brief). That is more than the United States and the European Union combined. Per-capita emissions now exceed the EU average. China is also by far the largest builder of renewable energy: in 2024 the country installed approximately 277 gigawatts of solar capacity in a single year and reached its target of 1,200 gigawatts of combined wind and solar capacity six years ahead of schedule (NEA China / IEA). The share of renewable capacity overtook coal for the first time in 2024. Yet coal still accounts for roughly half to 60 per cent of electricity generation. China continues to approve new coal-fired power plants, which serve as reserve capacity and grid support. The official climate strategy — the "Dual Carbon" targets — envisages peak emissions before 2030 and carbon neutrality by 2060 (State Council of China). Analysts at Carbon Brief considered it possible that emissions could plateau as early as the mid-2020s, as the addition of clean energy capacity covered incremental electricity demand growth for the first time. Since 2021, China has also operated the world's largest emissions trading scheme, which initially covers the power sector and is being gradually expanded.
AI data centres are pushing electricity demand higher, precisely when emissions are supposed to be peaking. If the build-out of solar, wind, and grid infrastructure fails to meet the additional demand, coal remains the bridge — and the emissions peak is pushed further out. China faces a fundamental trade-off: compute sovereignty and economic growth require cheap, reliable electricity, while planetary boundaries and its own climate commitments demand a rapid exit from coal. The decoupling of economic growth from emissions is progressing slowly, because industry remains energy-intensive and grid expansion must keep pace with the variable build-out of wind and solar. Whether China becomes the first major emerging economy to decouple growth from emissions, or whether the AI boom cements its dependence on coal, is the defining ecological question through to 2040.
Paths for China
China is prioritising the continued build-out of solar, wind, storage and ultra-high-voltage grids to such a degree that renewables cover all incremental demand growth and push coal into a pure reserve role. This matters because China already commands the world's largest expansion dynamic and the costs of solar and batteries have fallen sharply. The price is the massive grid and storage capacity required to balance volatility, and the regional interests of coal provinces whose employment and tax revenues depend on the fossil model.
Trade-off: achieving climate targets versus energy security and regional interests
Feasibility: Succeeds only if storage and grids keep pace with capacity additions
China is expanding the emissions trading scheme launched in 2021 from the power sector to industry, steel, cement, and aluminium, and is tightening the volume of allowances in order to put a genuine price on emissions. This addresses the core of the fault line, as an effective carbon price directs investment away from coal and towards clean energy. The cost is the burden placed on energy-intensive industries competing internationally, and the risk that overly generous allocations dilute the steering effect — a problem that has characterised the scheme to date.
Trade-off: climate governance versus industrial competitiveness
Feasibility: What matters is the tightening of allowances, not the trading mechanism itself
China is requiring new AI data centres to source the majority of their electricity from renewables, and is directing them — via the "East Data, West Computing" programme — to regions with abundant wind, solar, and hydropower. This is attractive because it can decouple the AI boom from coal without constraining growth. The costs lie in transmission losses over long distances, the geographic constraint of remote locations, and the question of whether efficiency requirements can keep pace with the speed of AI deployment.
Trade-off: compute sovereignty versus environmental governance
Feasibility: Only viable with a strong grid connecting generation to computing load
Trigger-based phases
Trigger: Clean energy additions cover the entire incremental rise in electricity demand for the first time, and emissions reach a plateau.
Renewables overtake coal in installed capacity, yet coal remains dominant in generation. AI-driven demand rises; grid expansion becomes the bottleneck.
Trigger: The emissions trading scheme is extended to industry, while coal provinces and grid bottlenecks slow the phase-out.
China is wrestling with the balance between energy security and decarbonisation. Storage, grids and efficiency mandates for data centres are being scaled up; coal is being pushed incrementally into a reserve role.
Trigger: Emissions fall structurally despite continued economic growth — or the AI boom entrenches coal dependency.
The outcome will determine whether China permanently decouples growth from emissions and stays on track for climate neutrality by 2060. The result will shape the global achievability of climate targets.
Real-world pilot programmes
Status: Under development (gigawatt-scale projects in desert regions)
China is constructing vast solar and wind bases in deserts such as Kubuqi in Inner Mongolia, concentrating clean energy at record scale. These projects are a test of whether capacity additions can be scaled quickly enough to cover incremental demand growth, including AI compute demand.
Status: Operational (expansion to additional sectors under way)
Launched in 2021, the emissions trading scheme is the largest in the world by covered emissions. Initially limited to the power sector, it is being extended to steel, cement, and aluminium. It represents a test of whether a carbon price can exert a genuine steering effect in a state-directed economy.
Status: Being implemented (national data centre strategy)
Launched in 2022, the initiative redirects data centres away from densely populated coastal regions towards the energy-rich west, where wind, solar and hydropower are abundantly available. It represents an attempt to align AI's voracious energy appetite with clean generation and grid expansion.
What we are watching
The measure is whether growth and emissions are decoupling. The threshold matters because only a sustained decoupling makes the pathway to carbon neutrality by 2060 credible. If the trend turns downward, the emissions peak has been passed — a global inflection point, given that China alone accounts for roughly one third of global emissions.
This indicator shows whether the build-out of wind and solar is genuinely displacing coal or merely supplementing it. The threshold matters because emissions will only fall once coal's share of generation declines. If that share does not decrease, AI-driven demand will entrench the fossil baseline.
The measure is whether infrastructure keeps pace with generation. The threshold matters because without grids and storage, surplus wind and solar output is curtailed and coal remains necessary as backup. If the build-out falls behind, a portion of the clean capacity additions is rendered ineffective.