China Uses More Electricity Than Any Country: Two Grid Academicians Can’t Agree on What Bends First

A Wuxi conference transcript and a spring television interview put demand-side flexibility against green fuels over a power system that consumed 10,370 TWh in 2025

The scene that frames this piece is unremarkable on its face. Late October 2026, a hotel conference hall in Wuxi, the annual Industrial Comprehensive Energy Efficiency Conference. On the vendor floor, Longshine Technology demos its Jiugong AI energy model, pitched at turning load forecasts into trading decisions. In an interview room away from the stage, China Engineering Academy academician Jiang Yi takes questions from a 21st Century Business Herald reporter, and the answers are published as a cleaned-up transcript, with cuts noted at the bottom. What Jiang says in that room, and what a fellow academician said on television seven months earlier, amount to two different theories of how the world’s largest electricity system absorbs the world’s fastest renewable buildout.

Why this transcript matters now

The numbers behind the conversation are no longer debatable in scale. China’s total electricity consumption crossed 10,370 TWh in 2025, up 5.0 percent year on year, and installed generating capacity reached 3,890 GW by the end of 2025, of which non-fossil sources held 2,400 GW, a 61.7 percent share (China Electricity Council, 2025-2026 annual supply and demand report). Renewable capacity stood at 2,337 GW and renewable generation at 3,990 TWh, or 38.3 percent of output. For 2026, the council projects consumption of 10.9 to 11 trillion kilowatt-hours, with the non-fossil share of capacity reaching about 63 percent.

The supply question has changed shape. At an April 2026 press conference, Liu Mingyang, deputy director of the National Energy Administration’s power department, forecast summer peak load of 1,575 to 1,600 GW, an increase of 70 to 90 GW over the previous year’s extreme, and consumption growth of roughly 500 billion kilowatt-hours over 2025 (National Energy Administration, 2026-04-27, link 🔗). His regional breakdown was calm on paper: supply broadly balanced nationwide, with parts of East, Central and South China tight at peak hours, to be absorbed by cross-provincial mutual aid, while the Northwest runs with headroom. The caveat came in the same breath: under widespread extreme weather or primary-fuel shortage, individual provinces could still see tight hours, managed through what the agency calls “one province, one policy” work plans and daily monitoring of supply and demand. Building capacity is no longer the binding constraint. Organizing the hours is.

The core argument: load should follow source

Jiang Yi’s opening exchange states the thesis in one line of system history. “In the past it was mainly ‘generation follows load’: users use more, plants generate more,” he says. “As wind and solar share rises, we need to mobilize the demand side and do more of ‘load follows generation’: when power is abundant, consume and store more; when it is tight, draw less, and where possible release stored power back.” The quote is the intellectual core of the transcript, and its position on the map of Chinese energy debates is clear: Jiang stands for demand-side flexibility, run on price and carbon signals, rather than on more wires and more batteries owned by the system.

His mechanism is domestic and specific. Buildings own the grid connection; electric cars own the battery. “You can think of the building as the mouth and the battery as the stomach,” he says, meaning the car’s storage enters the grid through the building’s distribution panel. The arithmetic he offers: at a future 300 million private electric vehicles with 50 kWh batteries each, the fleet holds roughly 15 billion kilowatt-hours of distributable storage. Private cars sit parked most of the day near homes and workplaces, which is exactly when the grid needs the help.

The obstacles he names are mundane, and that is what makes them credible. Many residential compounds already lack distribution headroom for ordinary charging piles. His fix is behavioral and procedural: “bidirectional and orderly” charging, where a car arriving home at the evening cooking-and-air-conditioning peak delays charging or discharges to support the building, then charges after the local load falls. His policy slogan is “one parking spot, one pile”, so drivers form the habit of plugging in on arrival. A car that stays connected is a car that can be dispatched.

Where each speaker stands

The Wuxi transcript is not a lone voice. Seven months earlier, in an interview with Yicai’s Chief Comment program published on The Paper on 25 March 2026, China Engineering Academy academician Huang Zhen, founding dean of Shanghai Jiao Tong University’s carbon neutrality research institute, laid out the competing frame: electrify everything that can be electrified, then make the electricity itself zero-carbon, and where electrons cannot reach, convert green power into fuel (Huang Zhen interview, 2026-03-25, link 🔗).

Table 1: The two academicians’ positions as stated in their own 2026 interviews

Question Jiang Yi (Wuxi transcript, October 2026) Huang Zhen (Yicai interview, March 2026)
Primary lever Demand-side flexibility: storage and dispatch at the user edge Supply-side decarbonization: electrification plus zero-carbon power and fuels
Key device Bidirectional charging piles, building-linked EV batteries Green methanol and hydrogen as long-duration storage
Control signal A carbon-emission-responsibility price computed every 15 minutes A completed electricity market: spot, capacity, ancillary services
Who moves first Users, once connected and rewarded Generators and fuel producers, once carbon pricing spreads

Source: positions compiled from the 21st Century Business Herald transcript and the Yicai interview as published

Jiang’s control layer is the most concrete claim in either interview. “Since the end of 2022, we have worked with State Grid’s North China branch to compute, every 15 minutes, the carbon-emission responsibility carried by one kilowatt-hour at that moment,” he says, and he stresses the word responsibility, meaning each unit of consumption is priced by what was actually on the margin, not by an annual average. Feed that signal to devices with the right algorithms, he argues, and chargers, distributed storage and data-center backup batteries begin scheduling themselves. At sufficient scale he is willing to claim the label: users who discharge during high-carbon hours and displace coal can reach “zero-carbon electricity use”, and beyond it “negative-carbon electricity use”.

His incentive design is equally blunt. “The car owner has supplied a battery resource and should be paid for it,” he says, floating free charging for vehicles that stay connected more than 6,000 hours a year. The owner gets a benefit, the system gets a balancing resource, and in his framing the meaning of a charging pile changes accordingly: not a refueling nozzle but a grid node.

There is a second half of his argument that the demand-side framing usually hides. Asked about storage, Jiang switches timescales and talks about heat. Electricity storage solves hours; heat storage solves seasons. Summer photovoltaic output can cover the evening, but waste heat from data centers and continuous heat from waste incineration arrives when nobody needs warming, and winter heating demand arrives when the sun is weak. Bridging that gap, in his telling, requires heat-storage facilities plus conversion equipment for different temperature levels, and he is emphatic about one device: “I am very optimistic about heat pumps.” Air and ground heat, lifted to usable temperature by a heat pump, count in his ledger as part of the same flexibility problem. The industry list he leaves for the transition years is deliberately unglamorous: distributed electricity storage, heat storage, heat pumps and the associated conversion equipment. For an academician associated with building physics, the ending note is characteristically material: energy matters when it arrives at the moment it is needed. Huang, asked about electricity demand peaking, rejects the premise from the other direction. “There is no such concept as a peak in electricity use,” he says, because living standards keep rising; the state will control carbon, not kilowatt-hours. He points to 4 trillion kilowatt-hours of the 2025 total already coming from renewables, and to wind and solar capacity growing from 530 GW in 2020 to 1,840 GW by the end of 2025, past coal in 2024 and past all thermal in 2025.

His own arithmetic is worth stating in full, because it supplies the numbers the Wuxi transcript lacks. The 2026 government work report, he notes, targets a carbon-intensity cut of about 3.8 percent for the year and a cumulative 17 percent across the 15th Five-Year Plan period, which he calls the decisive window for the 2030 peaking pledge. Electrification, in his version, has two layers. The first is direct: replace coal, oil and gas with electricity in boilers, kilns, building heat and transport, “wherever it can be electrified, electrify it”. The second is indirect, and it is where fuel enters: use green power to synthesize fuel, which he treats as a third energy form that solves both renewables absorption and fuel decarbonization at once. His policy shopping list is long and specific: a national low-carbon transition fund, shipping brought into the national carbon market so that a carrier switching from diesel to green methanol earns carbon allowances, and a carbon tax or “green premium” on conventional fuels to close the loop.

What each could say against the other

Read against each other, the two interviews hand each side ammunition. Huang’s frame implies Jiang’s flexing fleet is a rounding error next to the storage problem that actually matters: his green methanol route stores summer solar for winter heating and ships Jilin’s surplus down the coast to Shanghai and Guangzhou, a scale of time-shifting a car battery cannot touch. He also notes the demand driver for green fuel is currently European, not domestic: EU carbon rules on shipping pull Maersk and CMA CGM to buy Chinese green methanol, while domestic green methanol still costs roughly twice coal-based methanol.

Jiang’s transcript cuts back at Huang’s cost problem from the pricing side. A green fuel is only as cheap as the marginal green power feeding it, and Jiang’s 15-minute carbon-responsibility signal is precisely an instrument for finding the hours when that power is nearly free. Huang himself supplies the supporting fact: green power generation costs in western China have reportedly fallen below 0.15 yuan per kilowatt-hour, and he argues a green-power price around 0.2 yuan would let green methanol compete with diesel. His proposed demand-side fix, folding shipping into the national carbon market plus a carbon tax on conventional fuels, is a pricing mechanism, which is Jiang’s home turf. The scale context sits in the framing paragraph of Huang’s own interview: by the end of the 14th Five-Year Plan period, China’s new-type storage installations had passed 130 GW, non-fossil generating capacity had overtaken fossil, and non-fossil energy carried 21.7 percent of consumption. Grid-scale batteries of that size handle hours of balancing; the seasonal gap both men gesture at remains unfilled, which is exactly where the car fleet and the fuel route are being asked to compete.

The AI angle, which Huang volunteers unprompted, sharpens the competition between the two theories. Data centers and AI computing took about 1.6 percent of China’s electricity consumption in 2024, and his optimistic reading puts that at 5.3 percent by 2030. New computing hubs, he notes, are required to source at least 80 percent of their power from green electricity, which makes cheap western green power an industrial policy weapon rather than a grid-balancing nuisance. If the cheapest hours flow to computing hubs under long contracts, the flexible-hour arbitrage that Jiang’s connected cars depend on competes for the same electrons, and the price of flexibility rises accordingly. Neither interview asks this question, so the reader is left to.

The strangest part is where they already agree. Huang, the supply-side man, says China’s spot, capacity and ancillary-service markets are “still incomplete and in need of push”, and lists virtual power plants and vehicle-to-grid among the demand-side tools he endorses. Jiang, the demand-side man, presupposes a functioning market price to reward connected cars. Both theories fail at the same junction: without market instruments that pay for flexibility, the mouth-and-stomach fleet stays parked, and without cheap green power at scale, the fuel route stays a European subsidy story.

What went unanswered

The Wuxi Q&A leaves money on the table, unclaimed. Nothing in the transcript says who pays for upgrading distribution capacity in the neighborhoods that already lack it, before a single bidirectional pile earns its keep. The 6,000-hour free-charging idea comes without arithmetic: who funds the free electrons, and what a kilowatt-hour of flexibility is worth compared with the alternatives. The 15-minute carbon signal has run since late 2022 inside one State Grid regional branch, and the transcript does not say when or whether it becomes a product any device vendor can settle against.

The March interview has its own silences. Huang’s 5.3 percent projection for data-center and AI electricity demand in 2030, up from 1.6 percent in 2024, arrives without a named source or scenario. His pricing comparison, Chinese industrial power at just over 0.6 yuan against the United States 20 to 30 percent higher and Germany at roughly ten times, is offered as an advantage without a base year. And neither academician confronts the unglamorous actor in the middle: distribution companies, whose revenues and obligations decide whether “one parking spot, one pile” is a slogan or a construction plan.

What the pair of interviews actually exposes

Put the two documents side by side and the exposed seam is not disagreement but sequencing. China has solved the record-breaking part: 10,370 TWh consumed in 2025, 3,890 GW installed, renewables at 38.3 percent of generation. The flexibility layer that both men consider decisive is still at pilot scale: one regional carbon signal running since 2022, green methanol priced by European rules, vehicle-to-grid waiting on settlement rules that neither interview describes. The system’s physical growth and its institutional plumbing are on different clocks.

That gap is visible in the language each man reaches for. Jiang speaks in mechanisms, minutes and parking spots. Huang speaks in national KPIs, a 17 percent cut in carbon intensity for the 15th Five-Year Plan period and the 2030 peaking deadline, with 36 billion kilowatts of wind and solar targeted by 2035. Mechanisms get tested in neighborhoods; KPIs get tested in 2030. Between the two test dates sit the markets neither man can design alone, and the users nobody has yet paid to participate.

Sourcing note

Data as of November 2026. The two primary documents are published interview transcripts, edited by their outlets: the Jiang Yi Q&A from the Industrial Comprehensive Energy Efficiency Conference in Wuxi, conducted by 21st Century Business Herald reporter Deng Hao and published 30 October 2026 with cuts noted by the paper; and the Huang Zhen interview for Yicai’s Chief Comment program, published on The Paper on 25 March 2026. Direct quotes are taken from the published Chinese text and translated here for this article. Positions in the table are reconstructed from the stated content of each transcript, not from the speakers’ other writings. Capacity, consumption and forecast figures come from the China Electricity Council’s 2025-2026 supply and demand report and the National Energy Administration press conference of 27 April 2026. No figures were adjusted or re-estimated; where a speaker’s claim lacked a source (data-center demand, price comparisons), it is reported as a claim.

Limitations

  • Both transcripts are edited (“with cuts”), so completeness of each speaker’s argument cannot be assumed.
  • Speaker claims about costs and 2030 projections are unverified by third parties.
  • The carbon-responsibility pilot’s scope is described only by its proponent.

Sources

This article was prepared by the expert team at Global Infinity Labs, drawing on insights and data provided through its MCP data services.

About Global Infinity Labs

Global Infinity Labs is an AI innovation company specializing in MCP data services.

Email: support@globalinfinitylabs.com