China clears data center grid connections faster than US as renewable resistance grows
Chinese developers are securing power access ahead of US rivals, but state grid operators are pushing back on renewable-heavy AI loads.
Wood Mackenzie found in June (2026-06-15) that data center developers across Asia-Pacific are increasingly required to provide grid services support in exchange for power access — a condition raising project costs and extending development timelines that US developers have so far avoided at the same scale.4 That shift in bargaining power between utilities and data center operators is rewriting the terms on which AI infrastructure gets built.
Chinese developers are clearing interconnection hurdles faster than their US counterparts, where lengthy queue times remain the primary constraint on new capacity, project data and regulatory filings show.5 But the bottleneck in China has moved. Grid operators there now resist expanding renewable capacity dedicated to AI loads, concerned that peak demand at data centers is too difficult to forecast and would raise financial risks for power firms absorbing variable generation.5
The scale of what is being resisted is significant. More than 32 GW of planned data center capacity spread across over 1,150 projects in Asia-Pacific is already reshaping how grids accommodate large electricity users, with regulators introducing new requirements on reliability, flexibility and clean energy procurement, according to an Eco-Business report published in mid-June (2026-06-14).7 Chinese state grid companies, risk-averse by design, are effectively pumping the brakes on central government clean power ambitions for the sector.
In the US, the queue problem is compounded by supply chain exposure. Battery storage firms are seeing surging interest from AI data centers but face lengthy grid connection queues and a supply chain heavily dependent on China, Reuters reported in May (2026-05-19).1 The US added a record 57.6 GWh of new battery energy storage capacity in 2025, bringing total deployed capacity to 166.1 GWh, according to the Solar Energy Industries Association.1 The group projects annual deployments reaching 110 GWh by 2030, with a significant share driven by data center demand.1
Power demand from US data centers could reach 9% to 17% of total electricity supply by 2030, or up to 790 terawatt-hours, compared with around 4% at the time of the Electric Power Research Institute's May (2026-05-19) projection.1 For context, data centers consumed nearly 40% of global AI-related electricity demand in 2025, with the US accounting for more than any other country, including China, according to Canary Media's Chart of the Week series published on 3 July (2026-07-03).6
Storage is emerging as a workaround on both sides. Fluence is engaged in over 30 GWh of data center-related projects globally, with a meaningful portion in the US, CEO Julian Nebreda said.1 Tesla booked $430 million in revenue last year selling storage systems to Elon Musk's xAI, and Calibrant Energy agreed to provide a 31 MW/62 MWh battery system at an Aligned data center campus in the Pacific Northwest.1 US hyperscalers are pairing storage with contracted renewables to bypass interconnection delays, though that approach carries its own development costs.
Neither market has resolved the core problem: predictable, dispatchable power for loads that swing with model training cycles. India may offer an alternative for South and Southeast Asia, holding the world's third-largest renewable installed capacity alongside expanding transmission infrastructure, Moody's Ratings said in late May (2026-05-27).2 Installed IT load capacity there stands at around 1.2 GW, with a broadly similar pipeline under construction or announced.2
The policy gap on the US side runs deeper than interconnection queues alone. Policymakers have repeatedly shifted focus from one data component to the next — from training data to model weights to inference outputs — without building a framework for securing the full spectrum of the AI supply chain, the Atlantic Council noted in June (2026-06-01).3 The same piecemeal pattern applies to power: interconnection reform moves utility by utility, with no federal coordination mechanism that matches the scale of projected demand.
Chinese grid operators' reluctance to absorb renewable variability could ultimately cap the country's build-out even as developers clear connection queues faster than US rivals.5 The demand is there; the committed power is not. Whether Beijing forces state grid companies to accept the forecasting risk on variable renewables, or whether developers shift toward gas-fired backup generation, will shape which market narrows its power gap first — and at what carbon cost.