Huaneng Power International: Keeping the Lights On in the World's Largest Power Market

I. Introduction & Episode Roadmap

In January 2026, China's National Energy Administration published a number that had never appeared in any country's statistics before: in 2025, the People's Republic of China consumed 10.4 trillion kilowatt-hours of electricity, up 5% year on year.¹ That single figure is more than twice American consumption, and more than the European Union, Russia, India and Japan added together.¹ Electric vehicle charging demand alone grew 48.8%. Data-centre and IT-related power use grew 17%. The country that was rationing electricity to factories in the 1980s now runs an electrical system so large that its annual increment would rank among the world's larger national grids.

Somebody has to actually generate all of that. And one of the companies doing the most of it is a Beijing-headquartered enterprise that most Western investors have never traded and many can no longer easily buy: 华能国际电力股份有限公司 Huaneng Power International, listed in Shanghai as 600011 and in Hong Kong as 0902.

The scale is genuinely difficult to hold in your head. At the end of 2025, Huaneng Power International controlled 155,869 megawatts of generating capacity — roughly 156 gigawatts, or about one and a half times the entire installed capacity of the United Kingdom.² Its plants sold 437.6 billion kilowatt-hours into China's grids that year, and it booked RMB 229.3 billion of revenue, call it $32 billion.² It operates across 26 provinces, autonomous regions and municipalities, plus a wholly-owned power company in Singapore and an investment in Pakistan.³

Here is the puzzle that makes this a story rather than a fact sheet. In 2021, this enormous, state-backed, essential-infrastructure company lost RMB 10.6 billion — a swing of 547% from the prior year's profit.⁴ It lost money again in 2022.⁵ Then, over the following three years, it produced its best results in a decade: RMB 14.4 billion of profit attributable to shareholders in 2025 under PRC accounting standards, up 38%.² And it did that while revenue fell — down 6.6% in 2025, on 3.4% less electricity sold at a 3.5% lower average tariff.²

How does a company make dramatically more money on less revenue, less volume, and lower prices? That question is the spine of everything that follows. The answer involves a fuel cost that collapsed, a regulator that redesigned how coal plants get paid, and a business model where the company controls almost nothing about the price it receives and quite a lot about the cost it pays.

The road map: we start with why this company exists at all — a Deng-era experiment in raising foreign capital for Chinese power plants. Then the listing spree that put one operating company on three exchanges, and the coal-fired build-out that made it the largest of China's listed generators. Then the deepest section: how the business actually works, because in an administered-price market the mechanics are the investment case. Then the 2021–22 coal crisis that nearly broke the model, the 2023–24 policy rewiring that responded to it, the renewables build-out that is real but not yet the core, the people currently running it, the financial arc, and finally an honest bull-and-bear stress test of whether any of this constitutes a durable advantage or simply a well-positioned seat in a policy-driven system.

Let's start in a country that couldn't keep its factory lights on.

---

↑ Back to Top

II. Origins: A Power Shortage and a Reform Experiment (1985–1998)

Picture a textile mill in Jiangsu in 1984. The machines run three days a week — not because orders are short, but because the local grid cannot supply enough electricity to run them five. Across China in the early 1980s, this was ordinary. Rolling blackouts weren't an emergency; they were the baseline operating condition of an economy that Deng Xiaoping had just told to grow as fast as it possibly could. Power was the binding constraint. Every additional factory, every additional apartment block with a refrigerator in it, ran straight into a generation system built for a planned economy that had assumed demand would be, in effect, whatever the plan said it was.

The traditional fix would have been a bigger state plan and a bigger state budget. Beijing didn't have the budget. What it needed was capital it didn't own and management discipline the ministry system couldn't produce.

The answer, approved by the State Council in 1985, was 中国华能集团 China Huaneng Group — a state-owned enterprise created explicitly to break the mould, chartered to drive "reform, development and technology advancement of the power industry" and to serve as a model for other power companies on management standards and economic returns.⁶ The name itself carried the mission: 华 for China, 能 for energy. It was a state enterprise designed to behave like a company.

The structural innovation was the vehicle underneath it. Huaneng International Power Development Corporation — HIPDC — was set up to build and own power plants on a commercial basis, raising money offshore, importing equipment, and being held to a return on that capital. This mattered enormously in a system where power plants had historically been line items in a ministry budget. Somebody now had to care what a plant cost and how many hours it ran.

The people involved were not obscure. 李小鹏 Li Xiaopeng, son of Premier 李鹏 Li Peng — himself a career power engineer before he was a politician — spent most of the 1990s and 2000s inside Huaneng, rising to run the group before leaving in 2008 to become a provincial governor.⁷ That is worth stating plainly rather than winking at: Huaneng was, from the beginning, a commercially structured company operating inside a political system, staffed at the top by people with direct lines into it. That is not a scandal; it is the operating environment. It is also, as we will see in Section VIII, a permanent factor in how an outside shareholder should think about whose interests get optimised.

华能国际电力股份有限公司 Huaneng Power International, Inc. was incorporated in Beijing on 30 June 1994 as the listed flagship — the entity that would hold the good plants and face public markets.³

Then came the part that still surprises people. In October 1994, barely four months after incorporation, Huaneng Power International listed American Depositary Shares on the New York Stock Exchange.³ A mainland Chinese state enterprise, selling equity to American investors, five years after Tiananmen. The ADSs were a genuine signal — not that China was liberalising its power sector, which it was not, but that it was willing to submit a flagship state asset to US disclosure standards in exchange for hard currency and a validated share price. For the next twenty-eight years, HPI filed with the SEC, and those filings became one of the more detailed public windows into how a Chinese generator actually earned money.

A Hong Kong listing followed in January 1998, and a Shanghai A-share listing in November 2001.³ One operating company, three exchanges, three shareholder bases with three different sets of expectations — and, crucially, a structure that later gave the company the option to walk away from one of them without disturbing the others.

By the time the third listing closed, China's electricity problem had inverted. The question was no longer how to finance enough plants. It was how fast you could build them.

---

↑ Back to Top

III. Scaling Up: Asset Injections and the Coal Boom (1998–2015)

If the 1990s were about proving a Chinese generator could raise foreign equity, the 2000s and early 2010s were about something cruder and more consequential: pouring concrete.

China's growth model in that era ran on coal-fired electricity at a scale no country had attempted. Between the WTO accession year and the mid-2010s, China added generating capacity roughly equivalent to rebuilding the entire American power system, most of it burning coal. Huaneng Power International was in the middle of it, and it built at the technology frontier rather than the cheap end: the company put China's first 600 MW supercritical unit into service and, at its Yuhuan plant, China's first operational 1,000 MW ultra-supercritical coal unit.⁸ That distinction matters more than it sounds. Supercritical and ultra-supercritical units run hotter and at higher pressure, extracting more electricity per tonne of coal — the difference between a modern diesel engine and a 1970s one. In a business where fuel is the dominant cost and the selling price is set by the state, thermal efficiency is the margin.

But organic construction was only half the growth engine. The other half was a mechanism that Western investors often misread as a formality and Chinese investors watch closely: the parent-to-listco asset injection.

The template was on display in January 2015, when HPI completed the acquisition of ten generating entities from parent Huaneng Group and HIPDC — stakes in Hainan Power, Wuhan Power, Suzhou Thermal Power, Chaohu Power, Jingmen and Yingcheng Thermal, Ruijin and Anyuan Power, plus two hydro assets, Dalongtan and Hualiangting.⁹ The transaction added 7,787.5 MW of controlled capacity and 6,437.7 MW on an equity basis, and the company paid half the consideration on completion day.⁹

Understand what this pattern really is. The parent group develops and builds; the listed subsidiary, when the assets are mature and the price is negotiated between related parties, buys them. For minority shareholders this cuts both ways. On the good side, HPI gets to grow without carrying development risk, and the parent has an interest in the listco's share price because it is the currency for future injections. On the bad side, the buyer and the seller are the same ultimate controller, and a minority shareholder's protection rests on independent directors, connected-transaction rules and the Hong Kong listing regime rather than on genuine arm's-length negotiation. Injections are a growth channel and a governance exposure. That duality returns in Section VII, when we ask where the parent is choosing to house its renewable assets today.

Around HPI, the industry consolidated into a shape that still defines it. Following the 2002 break-up of the old State Power Corporation, five national generating groups emerged: 中国华能集团 China Huaneng Group, 中国大唐集团 China Datang, 中国华电集团 China Huadian, 国家电力投资集团 State Power Investment Corporation (formed from China Power Investment and the State Nuclear Power Technology Corporation), and 国家能源投资集团 China Energy Investment Corporation (from the merger of Shenhua and Guodian). Collectively they controlled around 44% of China's total installed generating capacity as of the end of 2020, when the national fleet stood at roughly 2.2 terawatts.¹⁰

That "Big Five" structure is the most important thing to hold on to about the competitive landscape, and it is nothing like a Western utility market. These are not five companies fighting for customers. They are five state-owned instruments, each with a national footprint, each expected to build where the state wants capacity built, each dispatched by grid operators they do not control. Which raises the question that the next section exists to answer: if you can't set your price, can't choose your customer, and sell a product that is physically identical to your rivals' — what, exactly, is the business?

---

↑ Back to Top

IV. How the Core Business Actually Works — Industry Structure, Competitors, Economics

Start with the single fact that governs everything else, because almost every mistake outside investors make about Chinese power companies comes from forgetting it.

Huaneng Power International does not sell electricity to you. It sells electricity, overwhelmingly, to two entities: 国家电网 State Grid Corporation of China and 中国南方电网 China Southern Power Grid. These are the transmission and distribution monopolies that between them cover essentially the entire country. They are the near-monopsony buyer of wholesale power in China.

A monopsony is the mirror image of a monopoly — one buyer, many sellers. If you have ever wondered why a company with 156 GW of capacity and an essential product doesn't have pricing power, that is the answer. Scale gives you leverage over suppliers. It gives you nothing against a customer that is also, functionally, an arm of the state, purchasing at prices the state administers.

The tariff: a price you mostly don't set

For most of the modern era, coal-fired power in China was sold at a provincial "benchmark on-grid tariff" — a regulated price per megawatt-hour set by the 国家发展和改革委员会 National Development and Reform Commission, the NDRC, and adjusted infrequently. Over the 2010s a market layer was bolted on: an increasing share of volume was traded bilaterally or through provincial exchanges, but around that same benchmark, within a permitted band.

In October 2021 — in the middle of the crisis we come to next — the NDRC widened that band substantially. Coal-fired market prices were allowed to float up to 20% above and 15% below the benchmark, the cap was lifted entirely for energy-intensive users, and effectively all coal-fired output was pushed into market trading, with commercial and industrial users required to buy from the market rather than at an administered retail schedule.¹¹

Call this what it is: partial marketisation as crisis management. The state widened the band because generators were losing money so badly the lights were going out. It did not surrender price-setting; it enlarged the corridor. For an investor, the practical translation is that HPI's realised tariff moves within a policy-defined range, and the direction of that range is a political decision about how much of the cost of electricity households and industry should bear. In 2025, HPI's average on-grid settlement tariff was RMB 477.08 per megawatt-hour, down 3.48% year on year.² Nobody at Huaneng chose that number.

The fuel: the one thing you can actually fight about

If the price is administered and the volume is dispatched, the controllable variable is cost. And for a fleet where roughly 92 GW of the 156 GW is coal-fired, cost means coal.⁵

Here the numbers get vivid. In 2025, HPI's unit fuel cost was RMB 266.88 per megawatt-hour, down 11.13% year on year.² Set that against the tariff of RMB 477.08. Fuel alone consumes well over half of the revenue per unit of electricity — before depreciation on a hundred-billion-renminbi asset base, before interest, before staff, before maintenance. In 2021, at the peak of the crisis, unit fuel cost was RMB 316.36 per megawatt-hour.⁴ A RMB 50/MWh swing in fuel cost, applied across roughly 440 billion kilowatt-hours, is a swing of over RMB 20 billion in gross profit. That arithmetic is the entire earnings story of the past five years, and it is worth internalising before any discussion of strategy.

Which is why HPI does something most pure generators don't: it owns pieces of the coal supply chain. The company's business description includes ports, shipping and coal transport alongside generation, and its "all other" reporting segment is largely port and transportation operations.¹² Concretely, it has built and invested in dedicated coal terminals — a 60%-owned coal transit base at the Haimen terminal zone of Shantou Port, designed with 70,000 DWT unloading and 50,000 DWT loading berths and planned annual throughput of 22.7 million tonnes; a roughly RMB 2 billion coal pier in Jiangsu serving as a public transhipment, storage and distribution terminal for the Yangtze River Delta waterway network; and an interest in the Huaneng coal terminal at the Port of Tianjin on the Bohai Bay.¹³¹⁴

Why does a power company own a wharf? Because Chinese thermal coal is mined mostly in the north and northwest — Shanxi, Shaanxi, Inner Mongolia — and burned mostly in the coastal and central load centres. Between the mine and the boiler sits a chain of rail, port handling, coastal shipping and river barging. Every link is a place where a generator either pays a spread to somebody else or captures it. Owning the terminal means secured berth access when the market is tight, storage that lets you buy when prices dip, and a physical hedge against the exact freight squeeze that turns a coal price spike into a coal availability crisis.

Management put this to work explicitly. On the Q3 2025 call, they attributed an 11% year-on-year reduction in standard coal cost to strategic purchasing during the off-season and to buying in the international market when it was cheaper than domestic supply.¹⁵ That is a real, verifiable operating capability — not a slide-deck claim.

But be careful about how much of a moat this is. Datang, Huadian and China Energy Investment all run their own fuel companies and port interests; China Energy Investment, born from the Shenhua merger, owns the coal mines and one of the world's largest dedicated coal railways. On the spectrum of fuel integration, HPI is better than an unintegrated merchant generator and structurally worse than the mine-owning giant. It is a cost-position advantage of degree, not of kind.

A smaller but real second exposure: HPI runs gas-fired peaking plants concentrated in coastal cities — Shanghai, Fujian, Guangdong — where gas capacity totalled 17,700 MW at end-2025.⁵ These earn from flexibility and heat supply rather than baseload economics, and they import a different commodity risk: LNG pricing.

The competitors, and what actually separates them

The listed comparables are 华电国际电力 Huadian Power International (SSE 600027, HKEX 1071), 大唐国际发电 Datang International Power Generation (SSE 601991, HKEX 0991), and 华润电力 China Resources Power (HKEX 0836). The two unlisted giants — China Energy Investment and State Power Investment — are larger in aggregate capacity than any single listed peer but reach public markets only through subsidiaries.

Now run the war game. Electrons are the purest commodity in existence; a kilowatt-hour from a Huaneng plant is physically indistinguishable from a Datang one. There is no brand, no product differentiation, no customer relationship to defend. Rivalry, in the Western sense, barely exists — these companies do not undercut each other for customers, because the grid dispatches according to provincial allocation rules and market clearing, not according to who called the customer first.

So what separates winners from losers? Three things, in order: unit fuel cost, utilisation hours, and where your plants happen to sit. A plant in a province with tight supply and a favourable tariff earns more than an identical plant three provinces away. That is not strategy; that is geography and history.

Five Forces, honestly applied

Buyer power: extreme. Concentrated, price-administered, and simultaneously the regulator's instrument. This is the ceiling on the entire investment case.

Supplier power: high and volatile. Coal miners, increasingly consolidated and periodically subject to safety and import policy shocks, sell into a market where generators cannot pass through cost in real time.

Barriers to entry: high, but not the capital. Anyone can raise money to build a power plant in China; the state-owned groups can raise it more cheaply. What is scarce is land, environmental permitting, provincial capacity quotas, and — above all — grid interconnection. In 7 Powers terms, that is close to a government-granted franchise: the barrier is administrative permission, not technology or capital.

Substitutes: rising, and this is the live one. Renewables plus storage are a genuine substitute for coal-fired energy, though not yet for coal-fired capacity, which is the distinction Section VI turns on.

Rivalry: structurally muted. Not because the companies are friendly, but because the mechanism through which they would compete — price to end customers — mostly doesn't exist.

The closest thing HPI has to a cornered resource is its coal logistics network plus its accumulated stock of long-dated grid interconnections and site permits in high-demand provinces. Both are real. Neither is exclusive.

Capacity share is not generation share is not profit share

One distinction to fix now, because it recurs in every discussion of Chinese utilities and is responsible for a great deal of sloppy analysis.

At end-2025, 41.01% of HPI's controlled capacity was classified as low-carbon and clean — 63,917 MW of gas, wind, solar, hydro and biomass.² That sounds like a company nearly half transitioned. It is not, and the reason is utilisation hours. HPI's overall average utilisation in 2025 was 3,111 hours, down 445 hours year on year.¹⁶ A coal unit in China typically runs several thousand hours more per year than a solar farm, which is limited by daylight to roughly 1,200–1,500 equivalent hours in most of the country. A gigawatt of coal and a gigawatt of solar are not the same asset.

So: capacity share overstates renewables. Generation share is lower. And profit share is a third number again, determined by which assets are earning above their cost of capital in a given year. Any analysis that treats the 41% headline as a measure of business transformation is measuring the wrong thing.

What makes the coal fleet still the profit engine is exactly what made it the source of catastrophe in 2021. Which is where we go next.

---

↑ Back to Top

V. The 2021–2022 Coal Price Crisis — the Inflection Point

In September and October 2021, something happened in China that had not happened in a generation: the lights went out on purpose.

Provincial governments in the northeast ordered rolling blackouts. Factories in Guangdong and Jiangsu were told to run on alternate days. Traffic lights failed in Shenyang. Apple and Tesla suppliers halted lines. The official explanations mentioned energy-intensity targets and dual-control policy, and those were part of it. But underneath sat a simpler, uglier mechanism, and it ran straight through companies like Huaneng Power International.

Here is how the trap closed. Post-Covid industrial demand rebounded hard, so electricity demand surged. Simultaneously, domestic coal supply was constrained by mine safety inspections, production discipline and import restrictions. Thermal coal prices went vertical. And the generators — who had to buy coal at whatever the market demanded — could sell electricity only at a benchmark tariff with a narrow permitted band.

They were, in the most literal sense, buying high and being told to sell low. So they did the only rational thing available: they stopped generating. Not out of protest, but because every additional kilowatt-hour destroyed value. A power shortage that looked like an energy crisis was, at the margin, a tariff design crisis.

The numbers at HPI are brutal and clarifying. Through the first three quarters of 2021, operating costs rose roughly 37% while tariffs rose under 5%, and the company's profit collapsed by around 91% year on year, including a loss of about RMB 3.5 billion in the third quarter alone.¹⁷

Full year, it got worse. Revenue actually grew 20.75%, to RMB 204.6 billion.⁴ Generation grew 13.2%, to 457.3 billion kilowatt-hours.⁴ Coal-fired utilisation hours rose 429 hours to 4,488.⁴ By every operating measure, 2021 was a boom year. And the company recorded a net loss attributable to equity holders of RMB 10.636 billion — a 547.27% reversal from 2020's RMB 2.378 billion profit.⁴¹⁸ The cause was one line: unit fuel cost up 51.32%, to RMB 316.36 per megawatt-hour.⁴

There is no more efficient illustration anywhere in global utilities of what it means to have no pricing power. HPI sold more electricity than ever before, at record utilisation, and lost more money than it ever had.

2022 was not the recovery. Northern port prices for 5,500 kcal thermal coal averaged RMB 1,296 a tonne, up another 24.2% year on year.¹⁹ Revenue rose again, to RMB 246.7 billion, and the company posted a second consecutive net loss attributable to shareholders — RMB 7.387 billion, a 26.17% narrowing but still deeply negative.¹⁹ Management's own explanation was blunt: coal prices remained high, the company's coal-fired share of capacity was large, and profits from renewable generation were not sufficient to cover the losses in coal power.¹⁹ Note the phrasing — it is an admission that the green fleet, at that point, was not big enough to matter to group earnings. That is a more honest statement than most managements would offer, and worth remembering when we assess credibility in Section VIII.

The quarterly shape tells you something else. The 2022 loss was not spread evenly: Q1 through Q4 losses ran RMB 0.956bn, 2.052bn, 0.934bn and 3.445bn.¹⁹ Nearly half the annual loss landed in the fourth quarter. Chinese generators typically clean house in Q4 — impairments on underperforming plants, provisions, true-ups. Anyone modelling HPI off nine-month run-rates without a Q4 haircut is going to be wrong in a predictable direction. That is a durable analytical point, not a one-off.

Two things about this crisis deserve emphasis. First, it was a national event, not a Huaneng failure — every Chinese thermal generator was hit, and China's power producers as a group saw profits tumble on record coal prices.¹⁷ Second, and more importantly, it exposed the structural flaw at the heart of the model: market-priced input, administered-price output. That mismatch had been tolerable for two decades because coal was cheap and stable. Once it wasn't, the whole design failed at once. Everything the state did to the sector in 2023 and 2024 is a response to this.

In the same window, HPI quietly closed a chapter. On 17 June 2022, it announced its intention to voluntarily delist its ADSs from the New York Stock Exchange, filing Form 25 around 27 June, with last trading around 7 July and a Form 15F to deregister the underlying H shares filed the same day.²⁰ The stated reasons were unremarkable — limited ADS trading volume relative to worldwide volume in the shares, and the administrative burden and cost of maintaining the NYSE listing and SEC compliance.²⁰

The temptation is to read geopolitics into this, and there is a modest amount to read: the timing coincided with the peak of Holding Foreign Companies Accountable Act pressure on US-listed Chinese issuers, and delisting removed an audit-inspection exposure. But don't overplay it. The ADSs genuinely were a small share of trading, the company remained fully listed and reporting in Hong Kong and Shanghai throughout, and Hong Kong disclosure is not a trivial regime.²⁰ The honest read is that a company in its second consecutive loss year removed a costly, low-utility listing at a moment when the cost-benefit had clearly flipped. What an investor loses is the SEC filing archive as a cross-check — a real, if modest, reduction in the disclosure surface.

By early 2023, the sector's economics were unsustainable and everybody in Beijing knew it. The fix, when it came, changed what a coal plant is for.

---

↑ Back to Top

VI. Rewiring How Coal Gets Paid: Carbon Market and the Capacity-Price Reform

Two policy moves, three years apart, did more to change the economics of Huaneng Power International than any acquisition in its history. Neither was announced by the company. Both were announced by the state.

The carbon market: a cost that is still small and definitionally rising

In 2021, China launched the world's largest emissions trading system by covered volume, and it did so with the power sector as the sole covered sector for the first several years. Today it regulates power generation, steel, cement and aluminium smelting across roughly 3,300 companies and an estimated 8 billion tonnes of CO₂.²¹

The design detail that matters is that allowances have been allocated free, on an output-based benchmark. Translate that into plain English: a plant is not given a fixed emissions budget. It is given allowances in proportion to how much electricity it generates, against a benchmark of emissions per unit of output. A plant that is more efficient than the benchmark ends up with surplus allowances it can sell. A plant that is worse must buy. It is, in effect, a subsidy paid by inefficient coal plants to efficient ones — which is precisely why HPI's early investment in supercritical and ultra-supercritical units has a second-order payoff nobody modelled in 2006.

Prices have been modest but structurally directional. The average secondary market price in 2025 was about RMB 70.78 per tonne of CO₂; by March 2026 the closing price sat around RMB 86, having eased about 3.8% in the month.²¹²² For a company burning coal at HPI's scale, that is currently a manageable line item, not a crisis.

The trajectory is what matters. In August 2025, guidance outlined a transition from intensity-based allocation toward absolute emission caps and the gradual introduction of auctioning, though without a fixed timetable.²¹ Those two changes — a hard cap, and paying for allowances instead of receiving them — would convert carbon from a rounding error into a genuine cost of production. This is the clearest example in the story of an uncapped tail risk that is currently invisible in the P&L. It is not a reason to panic in 2026. It is a reason not to extrapolate today's coal-fleet margins indefinitely.

The capacity payment: the most important thing that has happened to this company in a decade

On 10 November 2023, the NDRC established a coal-power capacity price mechanism, effective 1 January 2024.²³ The structure is a two-part tariff: a fixed payment per kilowatt of available capacity, layered on top of the market price for energy actually generated.

The initial levels were RMB 100 per kilowatt per year in most provinces, and RMB 165 per kilowatt in seven provinces with higher renewable penetration — Henan, Hunan, Chongqing, Sichuan, Qinghai, Yunnan and Guangxi.²³ These were calibrated to recover either 30% or 50% of a benchmark coal plant's total fixed costs, which the NDRC assessed at RMB 330 per kilowatt.²³ From 2026, the floor rises: all provinces move to at least RMB 165 per kilowatt, recovering at least 50% of fixed costs nationwide.²³

Here is why this is a genuine inflection rather than a technical footnote.

Before 2024, a Chinese coal plant earned money only by generating. Its fixed costs — depreciation, interest, staff, maintenance — had to be recovered entirely from the spread between the energy tariff and fuel cost. When that spread inverted in 2021, there was nothing underneath. The plant's revenue went to zero at the same moment its costs did not.

After 2024, a qualifying coal plant earns a payment simply for being available, whether or not it runs. Roughly a third to a half of its fixed cost base is now underwritten by a payment that is indifferent to coal prices, dispatch, and merit order.

That does two things simultaneously. It puts a floor under the coal fleet's economics that did not exist during the crisis — the direct policy answer to Section V. And it formally redefines what coal power is for. The state is no longer paying coal plants primarily to produce energy. It is paying them to stand ready as the backup and balancing layer for an electricity system in which wind and solar supply the energy and cannot be relied upon to supply it at 7pm on a still winter evening. Coal is being reclassified, by price signal, from primary generator to insurance policy.

For a company with roughly 92 GW of coal capacity, the arithmetic is material: independent analysis found the national payout in the scheme's first year was around RMB 107 billion, boosting coal plant revenues by approximately 5–8%, with individual cases reaching 7.9%.²⁴

Now the skeptical read, which is essential.

First, this is policy-granted, not earned. What the NDRC established, the NDRC can revise. There is no contract, no long-dated PPA, no regulatory compact of the kind that protects a US rate-regulated utility. A future administration facing pressure to cut electricity costs for industry has an obvious lever.

Second, the policy's own designers' intent is not clearly being achieved. The Regulatory Assistance Project's critique is pointed: the mechanism is available exclusively to coal, excluding storage, demand response and other flexibility resources that could do the same job; it permits newly built coal capacity to qualify, which invites overinvestment in exactly the asset class the transition is supposed to shrink; it lacks a rigorous process for identifying how much capacity is actually needed; and it sets fixed percentage allocations rather than using competitive auctions to discover the price of availability.²⁵

Third — and this is the finding that should give a bull pause — Carbon Brief's analysis of the first year found no clear evidence that the scheme reduced the hours coal plants actually operate, and documented loose application of eligibility criteria, with units dating back to 1994 and apparently captive facilities receiving payments.²⁴ If the policy's purpose was to enable coal to step back into a supporting role, the early evidence is that it has instead subsidised coal to keep doing what it was doing.

For an investor in HPI, that ambiguity is uncomfortable but not immediately harmful: a policy that overpays coal is, in the near term, good for a company that owns a lot of coal. The risk is asymmetric and deferred. If Beijing concludes the scheme is being gamed, the correction lands on the earnings floor that currently makes the coal fleet investable.

Which brings us to the other side of the same policy coin: the assets that are supposed to make coal's supporting role necessary in the first place.

---

↑ Back to Top

VII. The Green Build-Out — Real and Growing, Not Yet the Core

Walk the seawall at one of Huaneng's offshore wind sites in the Bohai or off the Jiangsu coast and the transition looks total. Turbine towers to the horizon, blades the length of a football pitch, service vessels shuttling technicians. At end-2025 the company held 5,880 MW of offshore wind within a total wind fleet of 20,618 MW, alongside 25,069 MW of solar.⁵ Add gas at 17,700 MW, a small hydro position of 370 MW and 160 MW of biomass, and you reach that 63,917 MW low-carbon figure — 41.01% of controlled capacity, up 5.19 percentage points in a single year.²⁵

That is not greenwashing. Adding roughly five percentage points of clean-energy share in twelve months, at 156 GW scale, is an enormous physical build programme. Through the first three quarters of 2025 alone, HPI commissioned 10.3 GW of new units, of which 6.83 GW was renewable, taking combined wind and solar to 44.66 GW.¹⁵

Three things need to be said clearly about it, and only one of them is flattering.

First: the sibling problem, and a correction worth making

There is persistent confusion — including in secondary research — about the relationship between HPI and 华能新能源 Huaneng Renewables Corporation. Let's settle it.

Huaneng Renewables was a separate group-level renewables company, listed in Hong Kong, and it is no longer listed. China Huaneng Group took it private in a voluntary conditional offer for all its H shares announced in October 2019, valued at roughly HK$15.9 billion; the offer went unconditional on 5 February 2020 and the H shares were delisted from the Hong Kong Stock Exchange on 24 February 2020.²⁶ It was the largest privatisation of a Chinese state-owned enterprise from the Hong Kong exchange at the time. The rationale reported at the time was that the market persistently undervalued the business, partly on concerns about its dependence on renewable subsidies that were being wound down.²⁷

So anyone still citing a listed "0958.HK" as Huaneng's renewables arm is working from stale information. It is a wholly group-controlled entity, and its results are not HPI's results.

That entity has kept growing outside the listed company. By end-2023 its total installed capacity had reached 28,455.77 MW, up 134% from September 2019, with solar capacity alone at 11,868.72 MW — an increase of 623.7% over 2021.²⁸ And in December 2024, China Huaneng Group signed an RMB 15 billion capital increase for Huaneng Renewables with a consortium of state investors — 中国国新 China Reform Holdings, 中国邮政 China Post's insurance arm (RMB 4 billion), 中国太平 China Taiping's Taiping Life, the National Green Development Fund, and an investment vehicle of China Southern Power Grid.²⁸ It was the largest private-equity style equity raise in China's renewables sector that year, roughly $2.1 billion.²⁹

This is the structural question a minority HPI shareholder should be asking, and it does not have a comfortable answer. The parent has two vehicles for renewable capital: the listed one, where outside shareholders participate in the returns, and the unlisted one, where they do not. In December 2024 it chose to route RMB 15 billion of fresh third-party equity into the unlisted one. Historically, mature assets have eventually been injected into HPI — that is the 2015 template. But "eventually, at a related-party price" is a materially worse deal than "developed inside the listco from the start." The 2015 pattern is a reason for hope, not a commitment.

Second: the returns are not yet proven superior to coal

This is where the story gets genuinely awkward for the transition narrative, and where the Q3 2025 earnings call is more revealing than any presentation.

Management disclosed that average tariffs for both wind and solar fell year on year and quarter on quarter, reflecting market oversupply and policy change; wind tariffs were down about 14% and solar about 9%.¹⁵ More pointedly, net profit in the wind segment fell 12.27%.¹⁵ Solar performed better, with unit profits up 5.11% sequentially on improved utilisation and operating efficiency.¹⁵

Read that again. In the year HPI posted record group profits, its wind business — the segment representing the future — saw earnings decline. The renewables build is adding capacity faster than it is adding profit.

The cause is not execution. It is market design. Under NDRC Document 136, issued in early 2025, all new wind and solar projects must in principle participate in power market trading, with on-grid prices formed through market mechanisms rather than fixed feed-in tariffs; projects commissioned after June 2025 have prices set through competitive bidding.³⁰³¹ Provinces are meant to run a contract-for-difference-style mechanism price, awarded by auction to a limited annual quota of capacity, providing a revenue buffer outside direct market transactions.³⁰ Implementation has been uneven — by late 2025 only about half of provinces had finalised their Document 136 rules.³⁰

For HPI, this means the renewable assets it is building today face a fundamentally less certain revenue profile than the ones it built five years ago. The subsidy era is over. The market era has arrived, and it is arriving into a market with a supply problem.

Third: curtailment and the negative price problem

Which brings us to the most concrete threat to renewable unit economics in China, and it is not a forecast — it is already happening.

Solar generates in the middle of the day. Everywhere. Simultaneously. In provinces that have built solar fastest, midday supply now routinely exceeds what the grid can absorb, because a large share of demand is already locked up in off-market long-term contracts, mostly with coal plants, and because there is not enough storage or flexible demand to soak up the surplus.³² The result is curtailment — output the grid simply refuses — and, in spot markets, negative prices.

Shandong has become the canonical case. In April 2025, solar cleared on Shandong's spot market at an average of about RMB 0.02 per kilowatt-hour, against a coal benchmark of RMB 0.35–0.45.³² Over the Labour Day holiday in May, Shandong spot prices stayed negative for 22 consecutive hours.³² Negative prices have appeared in Zhejiang and Sichuan too, in periods when wind and solar supplied only 20–25% of demand.³²³³

A negative price means a generator pays to produce. It is the market's way of saying that at this moment, this electricity has negative value. And it lands hardest on precisely the assets whose output is not dispatchable.

The capital allocation tension follows directly. HPI is funding heavy renewable capex — while carrying substantial leverage and a legacy coal fleet that still requires maintenance capital — into a segment whose realised prices are falling, whose regulatory support has been converted from a guaranteed tariff into an auctioned mechanism price, and whose marginal output is increasingly curtailed in the very provinces where build-out is fastest.

None of that means the build is wrong. Chinese renewable capacity crossed 60% of the national fleet by end-2025, and a generator that refuses to participate in that shift is choosing to become a stranded-asset story.³⁴ But "strategically necessary" and "value-accretive at current prices" are different claims, and only the first is currently supported by HPI's own segment disclosure.

Who is making these calls, and do they have a track record worth trusting?

---

↑ Back to Top

VIII. Current Management, Ownership, and Capital Allocation Credibility

On 16 June 2026, at Huaneng Power International's annual general meeting in Beijing, all resolutions passed by poll. About 60.9% of voting share capital participated, across 1,991 attending shareholders, predominantly A-share investors. Ten of the eighteen directors attended, alongside the chief accountant and board secretary. And on one resolution, connected shareholders holding roughly 46.23% of the issued shares abstained, as Hong Kong rules require.³⁵

That last detail is the whole governance story in a single line.

Who actually controls this company

Huaneng Group and HIPDC together hold approximately 46–47% of Huaneng Power International — the level of connected-party abstention at the 2026 AGM confirms the order of magnitude.³⁵ HIPDC has historically been the single largest direct holder, with China Huaneng Group holding the balance directly and through subsidiaries.³⁶

Note what this is not. It is not a majority stake in the arithmetic sense. But in a company where the free float is split across A shares and H shares, where institutional coordination among minorities is minimal, and where the controller is a central state-owned enterprise whose chairman is appointed through the state cadre system, a coherent 47% is control in every sense that matters.

An investor should price this honestly rather than either ignoring it or treating it as disqualifying. The controlling shareholder's objective function includes grid security, provincial employment, national energy policy, carbon targets and industrial strategy. Shareholder return is in that function. It is not the whole of it, and in a conflict it does not automatically win. When Beijing needed generators to keep burning expensive coal in 2021 rather than shutting down and protecting their P&L, the state-owned generators were the ones that kept burning it. That is a cost minority shareholders bore, and it is a cost they may bear again.

The corollary is the activist question, and it can be answered in one sentence: there is no realistic activist path here. A skeptical long/short investor can express a view on the shares. It cannot force an asset sale, a spin-off of the renewable business, a change of board composition, or a capital return policy. Any thesis premised on "unlocking value" through pressure on management runs into a controlling shareholder that does not need minority votes to do anything. That is a structural ceiling on the multiple, and it is a permanent feature, not a temporary discount.

The people at the top

Wang Kui has served as chairman since 2023, and Liu Ancang became president in 2025, succeeding Huang Lixin, who resigned; Liu was subsequently appointed an executive director of the eleventh board and a member of the strategy and nomination committees, taking effect from the conclusion of a shareholders' meeting.³⁷³⁸

Two observations. First, the transition is recent enough that judging Liu Ancang's strategic imprint would be premature — the operating results through mid-2026 largely reflect decisions and coal procurement contracts set before he arrived. Second, the pattern of executive movement in a central SOE is not the pattern of a Western public company. Senior appointments flow through group and Party channels; presidents move between the Big Five, between listcos and their parents, and into and out of government. This has an underrated implication for analysis: management turnover here carries less information about performance than it would elsewhere. A president leaving is not necessarily a signal about results. Investors used to reading CEO changes as verdicts should recalibrate.

A smaller governance note worth flagging as an accounting-and-disclosure signal rather than a scandal: in March 2026 the company appointed chief accountant and board secretary Wen Minggang, 56, as joint company secretary alongside Jiang Xiao, a Fangda Partners partner with a Columbia Law JD admitted in both New York and Hong Kong. The Stock Exchange granted a three-year waiver from strict compliance with Listing Rules 3.28 and 8.17 — the qualification requirements — conditional on Ms. Jiang's continued assistance.³⁹ This is a routine arrangement for mainland issuers. It is also, factually, a company secretary who does not meet the exchange's standard qualification bar, supported by an external adviser. Worth knowing; not worth alarm.

Capital allocation: the record is actually clean

This is where HPI's behaviour deserves credit, and where the evidence contradicts a common assumption about Chinese SOEs.

Through the crisis, the company did not pay a dividend it could not afford. Losing RMB 10.264 billion in 2021 and RMB 7.387 billion in 2022 on a PRC-standards basis, it made no distribution in either year.⁴⁰ For a state-controlled company whose parent presumably would have welcomed cash, cutting the dividend to zero through two loss years is discipline, not weakness.

Then the recovery, and the reinstatement, in careful steps. For 2023, with attributable profit back at RMB 8.446 billion, the dividend was RMB 0.20 per share.⁴⁰ For 2024, with profit up 20.01% to RMB 10.135 billion, the payout rose to RMB 0.27 per share — RMB 4.238 billion in total, a payout ratio of roughly 42%.⁴¹ For 2025, with attributable profit at RMB 14.410 billion under PRC standards, the board proposed RMB 0.40 per share, totalling RMB 6.279 billion and a payout ratio of 53.96%.²

Read the arc rather than the individual numbers: zero, zero, 0.20, 0.27, 0.40, with the payout ratio rising from roughly 42% to roughly 54% even as the earnings base grew. That is a management team increasing distributions faster than profits during a recovery — the opposite of hoarding, and a meaningful signal about their own confidence in the sustainability of the capacity-payment-supported earnings floor.

Does the cash actually cover it? Third-party analyses have periodically flagged dividend coverage and operating-cash-flow-to-debt as weak, and the fair test is the cash flow statement rather than the headline. On the evidence available: operating cash flow in the first nine months of 2025 was RMB 52.773 billion, up 22.75% year on year, and in the first half of 2025 alone RMB 30.75 billion, up 30.27%.⁴²⁴³ Against an annual dividend of RMB 6.279 billion, the distribution is covered many times over by operating cash flow.

The honest qualifier is that operating cash flow is not free cash flow. HPI is simultaneously building 10 GW a year of new capacity, and capital expenditure at that scale consumes the great majority of operating cash. The company's asset-liability ratio stood at 64.33% at the end of Q1 2025 — improved by 3.25 percentage points year on year, but still a heavily levered balance sheet.⁴² So: the dividend is comfortably covered by operating cash flow, and the combination of dividend plus growth capex is not self-funding. Growth is being part-financed by debt, at a leverage level that is falling but high. That is a legitimate concern, and it is a different concern from "the dividend is unsafe."

Guidance discipline and narrative consistency

Compare how management has explained itself across the cycle.

In 2021, the loss was attributed to coal price increases negatively affecting the domestic power business.⁴ That is external, and it is also true — an industry-wide shock, verified by every peer's results. In 2022, the framing added something less comfortable and more revealing: high coal prices, a high proportion of coal-fired capacity in the company's own fleet, and renewable profits insufficient to cover coal power losses.¹⁹ That second clause is a portfolio admission, not a blame-shift.

By the Q3 2025 call, the tone had changed in a way that is analytically useful. Management gave concrete procurement mechanics — off-season purchasing, international sourcing when arbitrage allowed — tied to a specific outcome, an 11.01% reduction in standard coal cost.¹⁵ They also volunteered that wind and solar tariffs were falling and that wind segment profit had declined, and they told analysts to expect coal prices to rebound in 2026 as economic conditions stabilised, while noting capacity payments would improve cost recovery in certain regions.¹⁵

That last point is the credibility test, and they passed it. Guiding toward a coal price rebound at the moment you are reporting record profits driven by falling coal prices is not what a promotional management does. And the guidance proved directionally right almost immediately: Qinhuangdao 5,500 kcal thermal coal fell from about RMB 770 a tonne at the start of 2025 to RMB 617 by end-June, then rebounded roughly 35% to about RMB 830 by November.⁴⁴

Where the disclosure is thinner: management commentary on renewable returns has been less forthcoming than on coal costs, and headline explanations of 2025's performance leaned on language about "enhanced safety, lean management, green development and governance initiatives" alongside marketing and pricing strategies.¹⁶ That is boilerplate. The actual driver was an 11.13% fall in unit fuel cost against a 3.48% fall in tariff — a fuel windfall, not a management triumph.² Investors should weight the operating disclosure heavily and the self-assessment lightly.

Which is the right frame for looking at the numbers as a whole.

---

↑ Back to Top

IX. The Financial Arc, 2018–2025: From Squeeze to Recovery

Lay out eight years of Huaneng Power International's income statement and you get a shape that looks, at first glance, like an error.

Revenue rose from roughly RMB 170 billion in 2020 to RMB 204.6 billion in 2021, RMB 246.7 billion in 2022, and a peak of RMB 254.4 billion in 2023 — a 50% increase in three years, driven by capacity growth and the higher tariffs that the October 2021 reform permitted.⁴¹⁹⁴⁵ Then it went into reverse: RMB 245.6 billion in 2024, RMB 229.3 billion in 2025.⁴⁵²

Now overlay profit. Losses of RMB 10.6 billion and RMB 7.4 billion in the two years revenue was climbing fastest. Then RMB 8.4 billion, RMB 10.1 billion and RMB 14.4 billion of attributable profit in the three years revenue declined.⁴⁰⁴¹² The lines run precisely opposite to each other.

The explanation is contained entirely in one relationship: the gap between the tariff HPI receives and the fuel cost it pays.

When coal spiked, revenue had to rise — tariffs were pushed up by policy and the company was generating flat out — but costs rose faster. When coal fell, revenue had to fall, because tariffs came back down and because the company was generating less as renewables displaced coal in the dispatch order. But costs fell faster. Revenue is a poor proxy for the health of this business. It is closer to a coal price index with a volume adjustment.

The 2025 figures make the mechanism explicit. Electricity sold from domestic plants fell 3.39% to 437.56 billion kilowatt-hours; the average settlement tariff fell 3.48% to RMB 477.08 per MWh; utilisation hours fell 445 to 3,111.²¹⁶ Every top-line indicator deteriorated. And unit fuel cost fell 11.13% to RMB 266.88 per MWh.² The spread widened by roughly RMB 17 per megawatt-hour against a tariff decline of about RMB 17 — but because the fuel decline applies to a much smaller base moving by a much larger percentage, the arithmetic came out strongly positive. Attributable profit rose 38.24% under PRC standards and 42.73% on the IFRS basis, with earnings per share of RMB 0.75.²¹⁶ Net assets per share rose 11.60% to RMB 4.52.¹⁶

Underneath that, the capacity payment has been quietly doing work no revenue line reveals. A payment for availability rather than output does not show up as higher volume or a higher tariff. It shows up as the coal fleet being profitable at utilisation levels that would previously have been marginal — which is exactly what a fall of 445 utilisation hours alongside a 38% profit increase implies.

The interim data through 2026 shows the mechanism running in reverse, and it is the clearest confirmation that this reading is right. In H1 2025, revenue fell 5.70% to RMB 112.03 billion while attributable profit rose 24.26% to RMB 9.26 billion, return on equity improved from 10.75% to 12.71%, and profit before tax rose 31.93%.⁴³ Through nine months, profit reached RMB 14.841 billion, up 42.52%, on revenue down 6.19%.¹⁵⁴⁶

Then coal rebounded. In Q1 2026, revenue fell 5.89% to RMB 56.78 billion and attributable profit fell 9.83% to RMB 4.48 billion — the first year-on-year profit decline of the recovery.⁴⁷ Domestic electricity sold dropped 4.82% to 101.489 billion kilowatt-hours and the average settlement tariff fell 5.63% to RMB 460.73 per MWh.⁴⁸ Operating cash flow declined sharply while assets stayed roughly stable.⁴⁷

Volume down, price down, and now fuel cost no longer falling fast enough to compensate. That is the whole model in one quarter, and it should be the base case for how investors read this company: the coal spread is the earnings driver, and it is not a management-controlled variable. The capacity payment raises the floor. It does not change what determines the level.

That tension — a genuinely improved floor sitting under a fundamentally uncontrollable spread — is what the bull and bear cases are actually arguing about.

---

↑ Back to Top

X. Bull vs. Bear — The Investment Case, Stress-Tested

Set aside the narrative for a moment and try to state, in the plainest possible terms, why someone would own Huaneng Power International from here, and what would break the case.

Why it may win

The demand backdrop is not in question. China's electricity consumption grew 5% in 2025 to a level no country has reached, and the growth drivers are structural rather than cyclical: EV charging up 48.8%, IT and data-centre load up 17%, residential demand up more than 6% on urbanisation, air conditioning and electrification of heating.¹ Whatever happens to Chinese GDP growth, electricity intensity is rising. A generator with 156 GW of capacity has a claim on that.

The earnings floor is genuinely higher than it was. The capacity mechanism structurally de-risks the coal fleet in a way nothing did in 2021, and the payment level steps up from 2026 to at least RMB 165/kW nationally, recovering at least half of benchmark fixed costs.²³ A generator that previously earned nothing when it didn't run now earns something. That is a real change in the shape of the downside.

The cost position is real, if not exclusive. Owning coal terminals, buying counter-seasonally, and sourcing internationally produced a measurable 11% reduction in standard coal cost in a year when peers faced the same market.¹⁵ Combined with a fleet skewed toward high-efficiency supercritical units, this is a defensible position at the low end of the industry cost curve.⁸

Capital access is cheap and reliable. A central SOE with a controlling state parent does not have a funding problem, which matters when you are running 64% leverage and building 10 GW a year.⁴²

And the transition is being funded, not deferred. Adding 5.19 percentage points of clean-energy capacity share in one year is not a company managing decline.²

Why it may not

The upside is capped by design. Everything above happens inside a system where the price is administered and the buyer is a monopsony. When costs fall, the benefit does not accrue permanently to the generator — tariffs get adjusted down, as they did by 3.48% in 2025 and 5.63% in Q1 2026.²⁴⁸ This is the difference between a utility with a regulated return and a generator with a regulated price. The former is compensated for its capital; the latter is squeezed between two variables it doesn't control. Investors modelling HPI on a Western regulated-utility template will systematically overestimate the durability of good years.

The growth segment's returns are unproven and currently deteriorating. Wind segment net profit fell 12.27% in a record year for the group, with wind tariffs down roughly 14% and solar down 9%.¹⁵ Document 136 has removed the fixed-tariff safety net for new projects.³⁰ Curtailment and negative midday pricing are already destroying value in the fastest-building provinces.³² The company is deploying its largest capital block into the part of the business with the least visible returns. Bulls should be required to explain why that changes.

Leverage plus capex plus a rising payout is a three-way claim on the same cash. Operating cash flow covers the dividend comfortably; it does not cover the dividend and the build programme.⁴² The gap is debt. At a 64% asset-liability ratio, the balance sheet is serviceable and improving, but there is limited slack for a bad coal year, a rate shock, or a capex overrun.⁴²

Carbon is a growing, uncapped tail cost. Free output-based allocation is a transitional design. The stated direction is absolute caps and auctioning.²¹ When that arrives, coal-fleet margins compress in a way that no amount of procurement skill offsets.

Governance ceilings the multiple. A 47% controlling bloc, an objective function including grid security and industrial policy, related-party asset injections as the growth channel, and no realistic mechanism for minority shareholders to force anything.³⁵ The renewables sibling receiving RMB 15 billion of outside equity in December 2024 while HPI shareholders watched is not a hypothetical concern about capital allocation — it is a completed transaction that routed growth capital away from the listed vehicle.²⁸

Testing the claimed advantages against evidence

Take each purported edge and ask what would falsify it.

Scale. HPI is the largest listed Chinese generator by capacity, but China Energy Investment and SPIC are larger at parent level, and scale in a monopsony market does not translate into negotiating power with the buyer. Verdict: real, but does not produce pricing power.

Fuel logistics. Verified by a specific, disclosed cost outcome. But Datang, Huadian and China Energy Investment run comparable fuel and transport operations, and China Energy Investment owns the mines outright. Verdict: a cost-curve position, not a moat. It narrows if peers invest, and it disappears entirely against a vertically integrated coal producer.

Capacity payments. The strongest element of the bull case and the most fragile in origin — granted by policy, revisable by policy, and already criticised by the analysts closest to it for failing to achieve its stated purpose.²⁴²⁵ Verdict: a genuine earnings floor with a political half-life.

Renewables growth. Physically impressive, financially unproven. Verdict: strategically necessary, not yet demonstrably value-creating on a like-for-like basis versus the coal fleet.

Government-granted franchise. The most durable of the Powers on offer — grid interconnection rights, provincial quotas and permits genuinely block entry. But note who grants it. A franchise from the state is held at the state's pleasure, and the same authority that blocks your competitors sets your price. Verdict: real, and double-edged.

The synthesis: this is a company whose downside has been meaningfully improved by policy and whose upside is structurally limited by the same policy apparatus. That is a defensible thing to own for cash generation. It is not a compounder, and the evidence does not support treating it as one.

---

↑ Back to Top

XI. Durable Lessons, KPIs, and What to Watch

Strip away the specifics of Chinese electricity policy and a few transferable lessons survive.

In an administered-price market, obsess over the controllables. HPI cannot set its tariff, choose its customer, or determine its dispatch. What it can do is buy coal well, run efficient units, manage its capital structure and choose its capacity mix. The entire difference between the 2021 disaster and the 2025 record was fuel cost — a controllable — moving in the opposite direction. Volume growth, the metric most companies are judged on, was actively negative in the good years and strongly positive in the terrible one. In this industry, growth is not a virtue and revenue is not a signal.

A single policy change can outweigh any deal in a company's history. No acquisition Huaneng Power International has ever made altered its economics as much as the November 2023 capacity price mechanism. In regulated and quasi-regulated industries, the highest-return analytical work is not modelling the company — it is understanding the regulator's incentives and reading the consultation documents before they become policy.

State-controlled utilities can be durable cash generators without ever re-rating like free-market compounders. The controlling structure that guarantees cheap capital and political support also guarantees that minority interests will sometimes be subordinated and that no activist can ever force the issue. Those are two sides of one coin, and pricing only one side is the classic error.

And distinguish capacity from generation from profit — always. The gap between those three numbers is where most utility misjudgements live.

The three KPIs that matter

If you track nothing else about this company, track these. Do not accept a summary; find them in the results announcement each quarter.

1. Average on-grid settlement tariff, by fuel type. The headline blended tariff hides everything. What matters is coal-fired tariff separately from wind and solar tariffs — because the coal number tells you what the state is allowing the fleet to earn, and the renewable numbers tell you whether the growth investment is being paid for. In 2025 the blended figure was RMB 477.08/MWh; by Q1 2026 it was RMB 460.73/MWh.²⁴⁸ Wind and solar tariffs were falling faster than the blend.¹⁵

2. Unit fuel cost per megawatt-hour, against the benchmark thermal coal price. Not just the absolute number — the relationship. If HPI's unit fuel cost falls faster than Qinhuangdao 5,500 kcal prices, the procurement and logistics advantage is working. If it tracks the index, the advantage is rhetorical. The 2025 datapoint: unit fuel cost RMB 266.88/MWh, down 11.13%, against a northern-port spot average down about 18.7% for the January–November period.²⁴⁴ That comparison is not flattering, and it is exactly the kind of check that separates a claimed edge from a demonstrated one.

3. Renewable capacity share versus renewable generation share. The gap between them is the curtailment signal. If clean capacity share climbs while clean generation share lags, the company is building assets whose output the grid is refusing. In an environment of negative midday prices in Shandong and elsewhere, this is the single most direct read on whether the green capex is earning its cost of capital.³²

The risk radar, mechanism by mechanism

Coal price volatility. The dominant P&L driver, with no hedge available at the scale required and no real-time pass-through. Already re-emerging in Q1 2026.⁴⁷

Tariff and regulatory risk. Upside capped by policy; and the capacity mechanism that supports the floor can be revised.²⁵

Carbon cost escalation. Currently modest, structurally directional, and dependent on decisions about caps and auctioning that have been signalled but not scheduled.²¹

Refinancing and cost of capital. A 64% asset-liability ratio funding a 10 GW-a-year build.⁴² SOE status mitigates but does not eliminate this.

Renewable curtailment and negative pricing. The direct threat to the returns on new capital.³²

Execution risk on a two-speed transition. Growing renewables fast enough to matter without letting the coal fleet — which still pays for everything — degrade. Both fleets need capital, and the company is currently funding both plus a rising dividend.

Related-party and structural risk. Growth via injections priced between affiliated parties; a parallel unlisted renewables vehicle absorbing outside equity.²⁸

---

↑ Back to Top

XII. Epilogue

In mid-2026, Huaneng Power International is a company caught precisely mid-transition, and unusually honest about it in its numbers if not always in its language.

The coal fleet — still roughly 92 GW and still the source of most generation and most profit — now sits on a policy floor that did not exist when it lost RMB 10.6 billion in a single year, and that floor steps higher in 2026 as capacity payments rise to at least RMB 165 per kilowatt nationwide.²³ The renewable fleet is being built at a pace few utilities anywhere can match, and is currently earning less per unit each year it grows.¹⁵ Leadership changed hands in 2025 and has not yet been tested through a full cycle.³⁸ And the first quarter of 2026 delivered the reminder that the recovery was never fully a management achievement: coal went back up, and profit went back down.⁴⁷

The next two or three annual results cycles will settle whether the 2023–24 reforms were a durable redesign or a bridge. The tell will be whether the coal fleet stays profitable through a coal price upcycle — because that, not the falling-coal years, is what a genuine floor is supposed to survive.

Three things would materially change the reading. A genuine move toward market-based wholesale pricing — real price discovery rather than a widened administered band — would transform the upside case and simultaneously remove the protection. Evidence that curtailment is being solved, through grid investment, storage build-out or demand flexibility, would turn the renewable capex from a strategic necessity into an economic one. And a shift in how much capital the parent group routes through the listed company versus the unlisted renewables vehicle would tell minority shareholders, more clearly than any statement, where they actually sit in the queue.

For now, what exists is a company that generates an enormous quantity of an essential product, in the largest and fastest-growing power market on earth, with almost no control over what it is paid for it. Everything else is a footnote to that sentence.

---

References

1. China's power consumption hits 10-trln-kWh milestone in 2025 — The State Council of the People's Republic of China, 2026-01-17 ↩↩↩

2. 华能国际电力股份有限公司2025年年度报告摘要 — 上海证券报, 2026-03-25 ↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩

3. Company Profile — Huaneng Power International, Inc. ↩↩↩↩

4. Net Loss Attributable to Equity Holders of the Company in 2021 Amounted to RMB 10.636 Billion, Year-on-Year Decrease 547.27% — PR Newswire, 2022-03 ↩↩↩↩↩↩↩↩↩

5. 华能国际电力股份有限公司2025年年度报告 — 2026-03 ↩↩↩↩↩

6. China Huaneng Group (CHNG) — China Daily, 2019-04-22 ↩

7. China's Li Xiaopeng, son of former premier Li Peng, signals end to closely watched career — South China Morning Post ↩

8. Huaneng Power International — company data profile, Global Energy Monitor ↩↩

9. Huaneng Power International, Inc. Announces Completion of Generating Assets Acquisition from Huaneng Group and HIPDC — PR Newswire, 2015-01 ↩↩

10. China's Big 5 power producers face uphill battle in meeting peak emissions targets — S&P Global Commodity Insights, 2021-06-07 ↩

11. Power Sector Reform — Guide to Chinese Climate Policy, Oxford Institute for Energy Studies ↩

12. Huaneng Power International, Inc. — Company Profile, MarketScreener ↩

13. Huaneng Power International, Inc. Obtains Approval on Shantou Port Haimen Terminal Zone Huaneng Coal Transit Base Project — PR Newswire ↩

14. Tianjin Port Huaneng Coal Terminal — Global Energy Monitor ↩

15. Huaneng Power International Inc (HUNGF) Q3 2025 Earnings Call Highlights: Record Net Profit Amid Revenue Challenges — Yahoo Finance / GuruFocus ↩↩↩↩↩↩↩↩↩↩↩↩

16. Huaneng Power International Boosts 2025 Profit Despite Revenue and Volume Declines — The Globe and Mail ↩↩↩↩↩

17. China power generators' profits tumble on record coal prices — Asia Financial, 2021-10 ↩↩

18. Net Profit Attributable to Equity Holders of the Company in 2020 Amounted to RMB 2.378 Billion, Year-on-Year Increase 210.28% — PR Newswire, 2021-03 ↩

19. 华能国际：新能源发电盈利未能覆盖煤电亏损等，2022年亏损73.87亿元 — 界面新闻, 2023-03-22 ↩↩↩↩↩↩

20. Huaneng Power International, Inc. Announcement of Intention to Delist American Depositary Shares from the New York Stock Exchange — PR Newswire, 2022-06-17 ↩↩↩

21. China National ETS — International Carbon Action Partnership ↩↩↩↩↩

22. China Carbon Prices Rise as Metals and Cement Enter the National Trading Scheme — Carbon Credits ↩

23. China Coal Capacity Payment Mechanism — Global Energy Monitor ↩↩↩↩↩↩

24. Guest post: China's "capacity payments" boosted coal-plant revenue by up to 8% — Carbon Brief ↩↩↩

25. China's coal power capacity payment policy: What it means and what's next — Regulatory Assistance Project ↩↩↩

26. Herbert Smith Freehills advises Huaneng Renewables Corp. Ltd on its privatisation from Hong Kong's Stock Exchange — Herbert Smith Freehills Kramer, 2020-02 ↩

27. Sun Sets for Huaneng New-Energy Unit on Hong Kong Stock Exchange — Caixin Global, 2020-02-11 ↩

28. 中信建投证券助力华能新能源150亿元增资项目圆满完成 — 新浪财经, 2024-12-18 ↩↩↩↩

29. China Huaneng secures $2.1 billion equity boost for renewables expansion — pv magazine, 2024-12-20 ↩

30. Analysis: Only half of Chinese provinces finalise key 'Document 136' renewable rules — Carbon Brief ↩↩↩↩

31. China to switch from FITs to market-oriented renewables pricing — pv magazine, 2025-02-12 ↩

32. China's overuse of coal is causing negative power prices — Eco-Business ↩↩↩↩↩↩↩

33. Negative electricity prices in Shandong put spotlight on China's energy transition challenges — S&P Global, 2023-05-23 ↩

34. Renewables account for over 60 pct of China's power capacity in 2025 — The State Council of the People's Republic of China, 2026-01-30 ↩

35. Huaneng Power International Shareholders Approve All AGM Resolutions and 2025 Final Dividend — The Globe and Mail ↩↩↩

36. Shareholding Structure — Huaneng Power International, Inc. ↩

37. Huaneng Power International, Inc. Announces Change of President — MarketScreener ↩

38. Huaneng Power International, Inc. Announces Appointment of Liu Ancang as Executive Director, Member of the Strategy Committee and the Nomination Committee — MarketScreener ↩↩

39. Appointment of Joint Company Secretaries — Huaneng Power International, HKEXnews, 2026-03-25 ↩

40. 华能国际时隔两年扭亏，拿出30多亿分红 — 界面新闻 ↩↩↩

41. 华能国际新旧能源齐发力全年赚101亿 负债率再下降拟派现42亿分红率42% — 新浪财经, 2025-03-27 ↩↩

42. 华能国际2025年一季报：净利润同比增长8.19%，行业领先表现引关注 — 搜狐财经, 2025-04 ↩↩↩↩↩↩

43. 2025 Interim Results Announcement — Huaneng Power International, HKEXnews, 2025-07-29 ↩↩

44. 2026 China Coal Market Trend Analysis — SunSirs ↩↩

45. Huaneng Power International (HKG:0902) Financials — StockAnalysis, data to 2026-03-31 ↩↩

46. 华能国际：2025年前三季度净利润148.41亿元 同比增长42.52% — 东方财富网, 2025-10-28 ↩

47. Huaneng Power International Q1 Profit Falls 9.8% as Declining Power Output and Prices Hit Revenue — BigGo Finance, 2026-04 ↩↩↩↩

48. Huaneng Power's Q1 2026 Electricity Sales and Tariffs Decline Amid Renewables Shift — TipRanks ↩↩↩