How China Became an Energy Superpower | Ep262: Professor Ning Li

ELI5/TLDR

China quietly turned itself into the first country that runs mostly on electricity instead of oil and gas, and built the factories to make the gear — solar panels, batteries, electric cars — cheaper than anyone else. When the Middle East blew up and oil routes got disrupted, most of Asia panicked while China barely flinched, because it had spent two decades weaning itself off imported fuel. The guest, a nuclear physicist who coined the term “small modular reactor,” argues the whole story comes down to one boring idea: technologies that can be mass-produced in a factory get cheap and good shockingly fast, and China bet on those. His current project is swapping the coal furnaces inside existing power plants for small nuclear reactors, reusing the building, the grid connection, and the workers.

The Full Story

The crisis that wasn’t

The conversation opens with a stress test. The Strait of Hormuz — the narrow shipping lane that roughly a fifth of the world’s oil and gas squeezes through — had been disrupted by Middle East conflict. The textbook prediction was that Asia, the biggest buyer, would get hammered. Japan, South Korea, Vietnam did: factories cut hours, offices sent people home to save power. China, the largest importer of all, shrugged.

“China has sort of started preparing for this kind of a problem or risk from quite a while back… it turned out to be not much.”

The reason isn’t a clever stockpile. It’s that China spent years reducing how much oil it actually needs. The trick was electrification — moving as much of the economy as possible off liquid fuel and onto the power grid. Cars, then industrial heat, then more. If your economy runs on electricity you make at home rather than fuel you ship in through a chokepoint, a closed strait becomes someone else’s problem.

Why factories beat construction sites

The intellectual spine of the whole episode is a single distinction: things you build on site versus things you build in a factory. Li spent a sabbatical at MIT around 2005 trying to figure out why nuclear power had stagnated while other technologies kept getting cheaper. His answer was modularity.

Think of it like the difference between building a house and building a phone. A house is assembled in place, one at a time, and each one is a bespoke project. A phone is stamped out by the million on a line, and every cycle the factory gets a little better and a little cheaper. Solar panels, batteries, and electric vehicles are phones. Traditional nuclear plants — and coal plants — are houses.

“The modular technologies just can scale faster.”

This is why Li, two decades ago, proposed making nuclear reactors small and identical enough to come off a production line — small modular reactors, or SMRs. Smaller also means safer: shrink a reactor enough and you can design out the possibility of a meltdown entirely, making Three Mile Island, Chernobyl, and Fukushima physically impossible rather than merely unlikely. Nobody picked up the idea then. Now, he notes dryly, it’s where everyone is rushing.

The phrase he keeps returning to is super-linear scaling: when a technology is modular, doubling the effort gets you more than double the output, because every unit teaches the factory something. China didn’t plan this as a doctrine. It just noticed which industries kept compounding and poured resources into those.

“Lots of things were done without a name or category. But once they find it’s effective and efficient, they started doing it very very quickly… it becomes a hindsight to study them rather than a foresight.”

The “new three”

The export numbers tell the story. China added over 260 gigawatts of solar and wind in a single year — more than most countries’ entire installed base. Electricity in some regions is so oversupplied that spot prices go negative for a few hundred hours a year. Western commentators have started calling China the first “electro-state.” China itself has a blunter label: the new three exports — solar panels, lithium batteries, and electric vehicles. Without those three, Chinese exports would have shrunk over the past several years.

When the Strait crisis hit, orders spiked. Chinese EV exports more than doubled in a month; solar and batteries nearly doubled. A blackout in Spain and Portugal the year before had already woken Europe up to the need for storage. The pattern Li describes is almost behavioural: a shock jolts buyers out of inertia, they discover the electric option is more reliable and more secure, and they don’t switch back. His framing of why it matters: energy security is the basis of national security. No energy security, no national security.

The deeper draw for the developing world is economic. Buy the panel or the EV once, and you have power or transport for decades, with no recurring fuel bill and no exposure to a chokepoint. That’s why electrification is accelerating faster in the global south than in rich countries — the poor countries have no legacy fossil system to protect.

The honest bit about dependency

The interviewer presses the obvious objection: aren’t we just swapping one dependency (fossil fuels) for another (China)? Li doesn’t dodge it, and his answer is more nuanced than a cheerleader’s. He explicitly tells the US and Europe not to go all-in the way China did. They have advanced combustion-engine industries and workers attached to them; rip that out too fast and you get stranded assets and unemployment, which is a social problem, not just an economic one.

But — and this is the load-bearing “but” — protect the old without subsidising it into permanence. Put the money into the new. His warning is generational:

“If you skew too heavily on the protection side, you’d be left out and then a generation or two later you’d be so far behind… You become the underdeveloped.”

He also turns the lesson on China: flooding neighbours with cheap goods and breaking their industries is bad statecraft. Both sides have something to learn. And concentration of an industry in one place is hardly unique — petroleum in the Middle East, cars once in America then Japan then Korea, large passenger jets now made almost only in the US and Europe. These things get negotiated.

Coal as a shock absorber, not an engine

One of the most counter-intuitive sections is about coal. China keeps building coal plants, which fuels the “two new coal plants a day” headlines. Li reframes it. China’s electricity used to be over 80% coal; it’s now around 50%. A coal plant is designed to run about 5,500 hours a year; the average Chinese plant now runs under 4,000, drifting toward 3,500 — less than half the time. Coal consumption is falling even as plant count rises.

Why keep them at all? Because fossil fuel does something renewables struggle with: it stores energy inside the fuel itself, cheaply and at scale, and the spinning turbines add physical inertia that keeps the grid stable.

“Before you dismantle the old you have to build the new first so that you don’t have a disruption in your system.”

The cautionary example is Germany, which retired coal and nuclear before storage could carry the load. So China is converting coal plants into stabilisers — dispatchable backup for the weeks when the sun and wind don’t show up — and adding nuclear for steady baseload. The coal plant becomes a fire extinguisher you keep on the wall, not the furnace you cook on.

Many plants are going further and switching from making power to making heat — steam piped to nearby factories, often more profitable than selling electricity, because steam can’t be shipped across a national grid the way power can. Roughly 50–60% of Chinese coal plants now have a heat component.

Heat pumps and the “one buy, three free” idea

This leads to the heat-pump tangent, where Li lands the cleanest mental model of the episode. Burn electricity to make heat directly and you get, at best, 100% of it back as heat. Use that same electricity to run a heat pump, and it moves existing heat instead of creating it — typically three to five units of heat per unit of electricity. The interviewer’s pitch: “buy one, get three free.” It’s not an appliance, it’s the single biggest efficiency upgrade most buildings can make.

The bigger point underneath is structural versus incremental efficiency. Adding loft insulation buys you a few percent. Switching the whole technology — coal heat to heat pump, petrol car to electric — buys you a step change of several-fold. An electric car from renewable power to wheels is near 90% efficient; a petrol car from well to wheels is around 20%.

“Why would people fight against these kind of dramatic improvements when you fight for like a one or two percentage point of improvement… Sticking with incrementalism is a losing proposition.”

The BYD–Octopus deal, and the kill-switch fear

The interviewer offers a concrete win-win: BYD and the UK’s Octopus Energy. An EV battery holds roughly a week of a household’s power and mostly sits idle. Turn on vehicle-to-grid (V2G) — let the car feed power back when the grid is short — lease the battery to Octopus for flexible storage, and the customer drives effectively for free, because driving uses only a sliver of the pack. BYD wins, Octopus wins, UK wind gets used instead of wasted, and nobody has to build separate storage.

The “what about kill switches?” fear — that Chinese cars or inverters could be remotely sabotaged — Li answers practically: a car’s comms are less sophisticated than the phone in your pocket, which is also made in China. The fix is the same as for phones — security testing and procurement rules — not blanket bans. His broader read is that treating everything as a threat works against your own interest, because it locks you out of the better technology. Taken to its logical end, total self-sufficiency means a massive cut in consumer choice, and consumers have voted for choice for decades.

Designed in America, built in China

The reactor story is the mirror image of the EV anxiety. Westinghouse brought its new AP1000 reactor design to China. China didn’t refuse out of fear of American kill switches; it paid for the technology transfer and worked the bugs out together. The design wasn’t finished when the deal was signed — thousands of engineers shuttled back and forth, CEOs camped in each other’s countries for months. China spent millions and ran through 23–24 iterations to cast a single complex pipe without welds. Before the deal, only Japan Steel Works could forge the big reactor vessels; after it, China had two suppliers. The result is one of the safest large reactors in the world — invented in the US, made real in China.

Li’s lament about America is pointed but not gloating: the US industrialised first, then let financial engineering take over and de-industrialised itself. When Trump visited China recently, the things on offer were soybeans, corn, and large aircraft whose design hasn’t changed in 40 years. He thinks it can be corrected — AI buildout, SMRs, reshoring — but the West shouldn’t dismiss China’s lessons as mere copying, because a lot of innovation comes from doing the thing, not just thinking about it.

Nuclear’s awkward truth, and the repower idea

Even Li, nuclear’s advocate, is honest about its position. China has the world’s largest, fastest nuclear program — and it’s a rounding error next to renewables. Nuclear adds 5–10 GW a year; solar and wind add 200–300 GW. The lesson he learned applies to his own field: factory-made beats site-built, and big reactors are site-built. China has a small modular reactor (the ACP100, about a tenth the size of a large reactor) finishing now, and once it learns to mass-produce, it’ll move fast.

His current project, “Coal-to-Nuclear repower,” is the synthesis of everything in the episode. Over 60% of energy’s end use is heat; in industry it’s 60–70%. China has thousands of coal plants sitting on good grid connections with trained workforces. So: pull out the coal boiler, drop in a modular reactor, keep the building, the grid, and the jobs. Clean heat into the existing system, minimal stranded assets, minimal local resistance — the same logic as TerraPower’s project in Wyoming, where a coal town welcomes a reactor because it saves the grid and the jobs.

Growth, not burden

The closing note is philosophical. Li argues that whenever climate action is framed as a cost or a sacrifice — burden-sharing, degrowth, preservation — it fails to move people, because people want development. China’s shift happened the moment it reframed solar, wind, and EVs not as a climate obligation but as an economic engine.

“It’s not the burden… Creating more opportunities, more growth, sharing the benefits.”

He thinks China is already the de facto leader on climate through actions rather than words, but learning to lead on the world stage takes time, made harder by geopolitics turning a practical problem into an ideological one.

Key Takeaways

Modular beats site-built. Technologies assembled in a factory (solar, batteries, EVs, phones) get cheaper and better far faster than things built on-site (large reactors, coal plants), because every production cycle teaches the factory. This is the single idea the whole episode rests on.
Super-linear scaling is the payoff of modularity: double the effort, get more than double the output, because each unit improves the line.
China’s resilience to the Hormuz disruption came not from stockpiles but from needing less oil — years of electrifying transport and industry off imported fuel.
China’s grid went from 80%+ coal to ~50% coal; coal consumption is dropping even as plant count rises, because plants now run under 4,000 hours/year (vs. 5,500 designed) as backup, not baseload.
Build the new before dismantling the old. Germany’s mistake was retiring coal and nuclear before storage could carry the load.
Fossil fuel’s hidden advantage: energy is stored in the fuel itself (cheap, large-scale), and spinning turbines add grid-stabilising inertia. That’s why coal can’t just be switched off.
Heat pump COP of 3–5: one unit of electricity moves 3–5 units of heat (it relocates heat rather than creating it). “Buy one, get three free.” The biggest single efficiency upgrade for most buildings.
Structural vs. incremental efficiency: switching the whole technology (petrol→EV, coal heat→heat pump) yields several-fold gains; tinkering with the old one yields a few percent. EV well-to-wheel ~90% efficient vs. ~20% for petrol.
V2G (vehicle-to-grid): an EV battery holds ~a week of household power and mostly sits idle; let it feed the grid back and it doubles as cheap mobile storage (the BYD–Octopus model).
The “new three exports” — solar, lithium batteries, EVs — are what kept Chinese exports from shrinking; the global south adopts faster because it has no legacy fossil system to protect.
The kill-switch fear is answerable with security testing and procurement rules (the same as for phones), not bans — Li notes a car’s comms are less sophisticated than a smartphone.
Coal-to-Nuclear repower: swap coal boilers for modular reactors inside existing plants, reusing the building, grid, and workforce — minimising stranded assets and local resistance.
Nuclear is tiny next to renewables (5–10 GW/yr vs. 200–300 GW), because big reactors are site-built; SMRs aim to make nuclear factory-made and therefore scalable.
Smaller reactors can be designed so a meltdown is physically impossible, not merely unlikely.
Framing matters: climate action framed as cost/burden/degrowth fails; framed as a growth engine, it succeeds. China’s turn came when renewables stopped being an obligation and became an industry.

Claude’s Take

This is a genuinely good interview, and the reason is the guest. Ning Li is a nuclear physicist who studied complexity science and has actually built reactors with Westinghouse, so when he talks about modularity and scaling he’s not reaching for a metaphor — he’s describing the thing he spent a career inside. The “factory beats construction site” frame is the kind of idea that’s obvious in retrospect and quietly reorganises how you read every energy headline afterward. That’s the mark of a useful watch.

The BS filter does need to come on for the China-flattering bits. This is a podcast called Cleaning Up, hosted by Bryony Worthington (a climate peer who helped write the UK’s Climate Change Act), interviewing a professor at a Chinese university, and the tone is admiring throughout. The harder edges of the China story — coerced technology transfer dressed up as “let’s work on it together,” the industrial overcapacity that’s hollowing out competitors’ factories, the labour conditions, the fact that “negative power prices” also signal colossal misallocation — get either softened or skipped. Li is more honest than most about some of this (he openly warns against China’s flood-the-market behaviour and tells the West not to copy China wholesale), which is why it doesn’t tip into propaganda. But you’re hearing the optimistic, engineer’s-eye version, and the geopolitics is sanded smooth.

I’d also flag one factual wobble worth not absorbing uncritically: the transcript repeatedly garbles the Middle East timeline (it says the conflict “broke out” at different points and mangles dates), so treat the specific event-sequencing loosely. The conceptual content — modularity, COP, the coal-as-backup logic, repowering — is solid and is the part worth keeping.

Eight out of ten. Docked from higher for the one-sidedness on geopolitics and the transcript’s date confusion; held at eight because the core mental models are excellent, durable, and cleanly explained, and Li is a rare guest who can speak with authority across physics, manufacturing economics, and policy at once.