AECC Aviation Power: China's Bet on the Engine That Took Fifteen Years

I. Introduction & Episode Roadmap

There is a stretch of industrial Xi'an, out past the old Tang-dynasty walls in the Weiyang district, where the street grid still follows the logic of a 1950s Soviet planning manual. Wide boulevards, low workshop halls, a rail spur, and the kind of perimeter fencing that tells you the buildings behind it were once assigned a number rather than a name. The number was 430. Today the sign says 中国航发动力股份有限公司 AECC Aviation Power Co., Ltd., and inside those halls sit some of the most closely guarded machine tools in China: five-axis mills, directional-solidification furnaces, and the vacuum casting equipment used to grow single-crystal turbine blades.

Those blades are the reason this story exists. And the company that makes them employed 30,428 people at the end of 2025, generated RMB 46.33 billion of revenue for the year — and earned a net profit of RMB 634 million on it.¹²

Read that again. Roughly one and a third percent of revenue fell to the bottom line. On a business carrying RMB 123.1 billion of total assets, the weighted return on equity was 1.58%.¹ Strip out one-off gains — chiefly a subsidiary selling down shares in a sister listed company — and profit was RMB 294 million, a 62.81% collapse from the prior year.¹ Meanwhile the market capitalises this enterprise at roughly RMB 98.7 billion, some 154 times trailing earnings.²

No sane investor buys AECC Aviation Power for its margins. They buy it for a thesis that has almost nothing to do with the income statement: that China has finally solved what Chinese engineers themselves call 航空发动机 the aero engine — nicknamed in domestic commentary "the chip of the sky," the last conspicuous hole in the country's military-industrial self-sufficiency. For thirty years China built world-class airframes and then hung Russian engines under them. If that dependency is now genuinely over, the listed vehicle that manufactures essentially every fighter, transport and helicopter engine in the People's Liberation Army order of battle becomes a different kind of asset.

And yet the stock has round-tripped. It touched RMB 63.78 within the past fifty-two weeks and traded at RMB 37.01 on August 18, 2026 — a decline of more than forty percent from the high, and a fall of 6.2% over the trailing year.² The market, in other words, is not serenely confident. Something in the 2025 numbers spooked it, and the argument over whether that something was noise or signal is the live investment question.

Here is the route this story takes. First, the origin: a Cold War engine plant that spent four decades copying other people's designs, and why that formative habit still shapes the company. Second, the transaction that actually created the AECC Aviation Power of today — not a product launch but a state-directed corporate restructuring between 2016 and 2021 that turned a single Xi'an factory into a national monopoly inside one listed shell. Third, the economics almost nobody models correctly: a defence pricing formula that caps profitability by policy rather than by competition. Fourth, the WS-15 saga — fifteen years, repeated public skepticism, and a payoff that arrived in December 2025. Fifth, what the company actually looks like on the numbers today, including a working-capital picture that deserves more attention than the engine headlines. Then management, industry structure against the global engine oligopoly, the civil-aviation option, and a stress-tested bull and bear case.

The through-line is a tension: a company with an unassailable strategic position and a profit-and-loss statement that looks like a job shop. Which one is the real company?

---

↑ Back to Top

II. Origins: A Cold-War Engine Plant (1958–1990s)

In 1958, China was in the first flush of the Great Leap Forward, and the industrial map of the country was being redrawn by Soviet advisers with a strong preference for putting strategic factories a long way inland. Xi'an — ancient capital, geographically central, ringed by mountains — got a cluster of them. One was designated the Xi'an Red Flag Machinery Factory, known in the internal numbering system as Factory 430. Its assignment was aero engines.²³

For its first three decades, the plant did not design engines. It reproduced them. The early output was the WP-5 and WP-7 family of turbojets — Chinese productions of Soviet designs that powered the J-6 and J-7 fighters, themselves licensed derivatives of the MiG-19 and MiG-21. The relationship was straightforward: Moscow supplied the drawings, the metallurgy, and the tolerances; Xi'an supplied the labour and the discipline to hit them.

Then the Soviet advisers left. The Sino-Soviet split stranded China's aviation industry mid-sentence, with production lines for engines whose upgrade paths ran through a supplier that had become an adversary. What followed was a long and instructive improvisation. In the mid-1970s, in one of the more unusual technology transfers of the Cold War, China licensed the Rolls-Royce Spey MK202 — a British military turbofan — and set Xi'an to work reproducing it. The result, eventually designated the WS-9, took the better part of two decades to reach a state Chinese industry considered domestically reproducible, and it powered the H-6 bomber line and the JH-7 fighter-bomber. Xi'an also participated in work on the CFM56 commercial turbofan, and in the long-running WS-10 and, later, WS-15 programmes.³

Why the copying mattered

It is easy to be dismissive about reverse engineering. It is more useful to understand what it teaches and what it does not. Reproducing a Spey teaches you enormously about manufacturing: how to hold tolerances on a compressor disc, how to inspect a weld you cannot see, how to build a supply chain for exotic alloys. What it does not teach you is design — the accumulated empirical judgment about why a particular blade profile stalls at a particular angle of attack, knowledge that Rolls-Royce and Pratt & Whitney acquired by breaking thousands of engines on test stands over fifty years and writing down what happened.

This is the single most important thing to understand about jet engines as a business. An engine is not a difficult product in the sense that a semiconductor is difficult — there is no equivalent of an EUV lithography machine that you can be denied. It is difficult in the sense that a Stradivarius is difficult: the physics is known, the materials are purchasable, and the gap between a working engine and a reliable one is a body of empirical, hard-won, largely undocumented knowledge about what breaks and when. A turbine blade in a modern fighter engine operates in gas hotter than the melting point of the alloy it is made from, kept solid only by internal cooling passages and a ceramic coating a fraction of a millimetre thick, while spinning fast enough to pull tens of thousands of times its own weight in centrifugal load. Getting that to work once is engineering. Getting it to work for a thousand hours, ten thousand times over, is institutional memory.

China had bought the first kind of knowledge. It had to grow the second.

The listed shell

In 1996, a slice of the Xi'an operation was carved out and floated on the Shanghai Stock Exchange as a listed company under the umbrella of 中国航空工业集团 AVIC, China's sprawling state aviation conglomerate. The plant itself was reorganised in March 1998 as Xi'an Aeroengine (Group) Limited Company.³ This is the entity that, through several name changes, became stock code 600893.

The structure is worth pausing on, because it explains a great deal about the governance questions that arrive later in this story. What listed in 1996 was not a company in the Western sense — an independent enterprise raising capital from the public to pursue its own strategy. It was a manufacturing asset, held inside a much larger state group, with a minority of its equity made tradable. Strategy came from above. Customers came from above. Prices, as the next sections show, came from a government pricing manual. The public shareholders were passengers on a vehicle whose steering wheel was elsewhere.

And through the 1990s and 2000s, that vehicle was not going anywhere especially fast, because China's air force was solving its engine problem by importing. The J-10 and the J-11 — China's most capable fighters of the era — flew on Russian AL-31F turbofans bought from Saturn and its successors. The bomber fleet ran on Spey derivatives. China had, by the 2000s, become genuinely good at airframes, radars, and missiles, and remained dependent on a foreign supplier for the one component without which none of it flies.

That dependency was not merely commercial. It shaped doctrine. An air force that cannot guarantee engine resupply flies its aircraft differently, trains less aggressively, and plans campaigns around a sustainment risk it does not control. Closing that gap was the strategic problem that everything from here — the 2016 corporate carve-out, the WS-15 programme, the multi-decade tolerance for terrible returns on capital — was designed to solve.

The first move was not technical. It was organisational.

---

↑ Back to Top

III. The Inflection: AECC Is Born, and the 2019–2021 Roll-Up (2016–2021)

On May 31, 2016, a new company was quietly registered with the Beijing Administration for Industry and Commerce. There was no ceremony, no press conference, and for a fortnight essentially no coverage. Only later did Chinese financial media notice what had appeared in the corporate registry: 中国航空发动机集团有限公司 Aero Engine Corporation of China, with registered capital of RMB 50 billion, headquartered in Haidian district, and four legal-person shareholders — the State-owned Assets Supervision and Administration Commission, the Beijing State-owned Capital Operation and Management Center holding RMB 10 billion or 20%, AVIC, and 中国商飞 COMAC.⁴ Its stated business scope covered propulsion for military and civil aircraft, auxiliary power units, gas turbines, and helicopter transmissions.

The formal launch came on August 28, 2016. What had happened, in substance, was a divorce. China's engine assets — roughly forty units and, by contemporary accounts, well over RMB 100 billion of total assets — were extracted from AVIC and given their own corporate parent, their own balance sheet, and their own line to the State Council.⁴ Three AVIC-controlled listed companies moved across in the process, including stock code 600893.

Why a carve-out was the strategy

The logic was a diagnosis of why the engine problem had persisted. Inside AVIC, engines competed for capital and talent with aircraft, avionics, helicopters and missiles — and lost, repeatedly, because engines have the worst return profile of any of them. They take fifteen to twenty-five years to develop, consume enormous capital, and produce very little revenue until certification. Any conglomerate optimising for near-term output naturally starves them.

The remedy was to remove the choice. Give engines a dedicated parent whose only mandate is engines, fund it directly, and measure it on technology milestones rather than on profit. This is industrial policy operating on organisational design rather than on subsidy, and it is a genuinely interesting piece of state capitalism: the state's diagnosis was not that China lacked money or engineers, but that it lacked an institution whose survival depended on the problem being solved.

In March 2017, the Shanghai-listed shell was renamed AECC Aviation Power — a public signal that this would be the group's primary listed manufacturing vehicle rather than one asset among many.

The roll-up, and what was actually bought

What made 600893 investable as "the AECC stock," though, was not the renaming. It was a restructuring that ran from 2019 to 2021, and it has a wrinkle that most summaries get wrong.

The story begins with debt. Under State Council Document No. 54 of 2016 — the policy framework for 市场化债转股 market-oriented debt-to-equity swaps — three of AECC Aviation Power's key manufacturing subsidiaries carried balance sheets bloated by historical borrowing. Those three were 中国航发沈阳黎明航空发动机 AECC Shenyang Liming, the Shenyang plant that builds the WS-10 family of fighter engines; 中国航发贵州黎阳航空动力 AECC Guizhou Liyang, the mountain-country turboshaft and turboprop plant; and 中国航发南方工业 AECC South Industry, the small-turbofan, turboshaft and APU specialist.⁵

A consortium put RMB 6.5 billion of cash and converted debt into those three subsidiaries — AECC group itself, the National Development Fund, the National Military-Civilian Integration Industry Investment Fund, Bank of Communications Financial Asset Investment, Xinmaisui Investment, China Orient Asset Management, and Gongrong Jintou. AECC group separately contributed RMB 1.98 billion of state-appropriated capital reserve. Leverage came down. But the price was that the listed company's ownership of its own most important factories was diluted.⁵

The 2019 transaction, announced on July 19, 2019, undid that dilution. AECC Aviation Power issued shares to those seven counterparties to buy back the stakes they had received: 31.23% of Liming, 29.14% of Liyang, and 13.26% of South Industry, restoring the listed company to full ownership.⁵ The consideration was roughly RMB 8.49 billion of newly issued stock — an initial 412.92 million shares priced at RMB 20.56, later adjusted to 415.75 million shares at RMB 20.42. The deal completed in 2021.⁶

The M&A question that actually matters

Frame this as a deal, because that is how a sophisticated investor should read it — and then notice what is missing from it.

There was no competing bidder. There was no auction. There was no market-tested price. The transaction documents are explicit that the final consideration would be set by the appraised value in a valuation report filed with the state asset regulator.⁵ Every counterparty was a state entity or a state-backed fund. The buyer's controlling shareholder was also one of the sellers.

So the M&A question here is not the usual one — did the acquirer overpay in cash? No cash moved. The question is a governance question: were minority shareholders diluted on fair terms in an intra-group transaction with no independent price discovery? An appraisal-based price is a professional opinion, not a market clearing price, and appraisals of defence assets whose revenue, cost, capacity and order books are legally undisclosable involve extraordinary judgment. Minority holders had to take it on trust that the appraisers, the state asset regulator, and the parent all landed on a number that did not systematically favour the sellers.

There is no public evidence that they did not. There is also no mechanism by which an outside investor could verify it. That asymmetry — depend on the parent's fairness, because competitive discipline is unavailable — recurs throughout this story and is the single most durable governance feature of the name.

What the roll-up did accomplish is not in doubt. Overnight, 600893 stopped being a Xi'an engine plant and became the consolidated manufacturing base for essentially all of AECC's fighter, transport and helicopter engine output — four production complexes spanning Shaanxi, Liaoning, Guizhou and Hunan. No product launch in the company's history changed its investment identity as much as this piece of paperwork.

Which raises the obvious follow-up: if you now own a national monopoly, why don't you earn like one?

---

↑ Back to Top

IV. The Economics Nobody Prices In: Cost-Plus and the Pricing Reform

The answer sits in a document most investors have never read: the 1996 《军品价格管理办法》 Measures for the Administration of Military Product Prices, which governed how the People's Liberation Army paid its suppliers for more than two decades.

Its guiding principle was rendered in four characters — 保本、微利、免税 — break even, thin profit, tax exempt. The formula was arithmetic of almost childlike simplicity: price equals audited cost, plus five percent of audited cost.⁷

Sit with that for a moment, because it inverts every instinct a commercial investor has. In this system, a contractor cannot earn a superior return by being better. It can only earn more money by spending more money. Halve your manufacturing cost through brilliant process engineering and your revenue halves and your profit halves with it. Let costs drift upward and both rise. The formula was designed to protect the state from profiteering, and it did — while systematically destroying any incentive to be efficient. Chinese defence-sector analysts have written about this bluntly for years: the regime produced enterprises with, in the standard phrase, no cost consciousness at all.

The 2019 reform, and why it changed less than advertised

Beijing understood the problem. In 2019 a new set of rules, the 《军品定价议价规则》 Military Product Pricing and Negotiation Rules, replaced the flat markup with a three-part structure: pricing cost, plus five percent of a negotiated target price, plus an incentive profit calculated from the gap between actual cost and target cost. Where a contractor beats target, it keeps roughly seventy percent of the saving and the military takes thirty. Where it overruns, it absorbs seventy percent of the overrun. The incentive component is capped, both up and down, at five percent of the target price.⁷

This was a real improvement. It converted a pure cost-plus contract into something closer to a fixed-price-incentive arrangement, and it gave a well-run plant a genuine reason to attack its own cost base. Companies that reduce costs now keep something.

But look at the ceiling. In the best imaginable case — a contractor that systematically beats every target cost by a wide margin — the structure delivers roughly ten percent of target price as profit. That is the maximum. It is not a negotiating position, it is a policy parameter.

And the reported results are consistent with a company nowhere near that ceiling. AECC Aviation Power's aero-engine and derivative products segment — which is to say, essentially the whole company — produced a gross margin of 8.50% in 2025, down 1.04 percentage points year on year. Manufacturing overall came in at 9.09%.¹ Revenue nearly doubled from roughly RMB 28.6 billion in 2020 to RMB 47.9 billion in 2024 before slipping to RMB 46.3 billion in 2025, and across that entire arc the gross margin barely moved out of an eight-to-nine percent band.¹⁸

That is the tell. When a business grows revenue by two-thirds over five years and its gross margin does not budge, the margin is not being determined by scale, mix, learning curves or bargaining. It is being determined by a formula. Operating leverage — the reliable mechanism by which growing manufacturers get more profitable — has been switched off at the source.

What this means for how you value the company

Two consequences follow, and they run in opposite directions.

The first is protective. A cost-plus-derived contractor is insulated from the ordinary disasters of manufacturing. It does not face price competition, because there is no competitor. It is largely protected from input-cost inflation, because higher input costs pass into the pricing cost base. It cannot be undercut on a bid it does not have to win. In a genuine industrial downturn, this business does not see a demand collapse; it sees a procurement calendar.

The second is corrosive. Because profit is anchored to cost rather than to value, the company captures almost none of the strategic worth of what it produces. A WS-15 turbofan is, by any reasonable measure, one of the most valuable manufactured objects in China — it is the difference between an air force that can fly and one that cannot. Its manufacturer books roughly eight and a half cents of gross profit on every dollar of it, and after selling, administrative, R&D and financial expenses, keeps a little over one cent as net income.

Global engine makers do the reverse: they frequently lose money on the initial engine and make it back many times over across decades of spare parts and overhauls. That model — the one that generates the profits investors associate with the word "aerospace" — is examined in Section VIII. AECC Aviation Power does not have it yet.

For now, hold on to the structural fact: this is a company whose profitability ceiling is set in Beijing, not in the market. Which makes the technical achievement of the last decade all the more remarkable, because it was pursued under economics that would have killed the programme at any Western firm.

---

↑ Back to Top

V. The WS-15 Saga: Fifteen Years to a Clean-Sheet Engine (2005–2026)

In January 2011, a J-20 prototype lifted off from Chengdu with then-U.S. Defense Secretary Robert Gates in Beijing, and the world got its first look at China's stealth fighter. Analysts pored over the planform, the canards, the radar-absorbent coating. And then they looked at the exhaust nozzles, and a quiet consensus formed among people who follow propulsion: those are Russian engines.

They were right, and they would keep being right for more than a decade. The J-20 — an aircraft designed around supercruise, high-altitude persistence, and long-range missile employment — flew for years on engines that could not deliver what its airframe was designed to exploit. First Russian AL-31s. Then, as domestic industry matured, the WS-10C, an uprated member of the WS-10 Taihang family that Shenyang Liming had been grinding away at since the 1980s.

The WS-10C was a genuine achievement and an important one. It was, however, a derivative. The engine the J-20 was drawn around was something else entirely: the WS-15, a clean-sheet high-thrust turbofan whose development began in the mid-2000s and which was supposed to make the aircraft what it was meant to be.

The problem with hot metal

To understand why it took fifteen years, understand what a high-pressure turbine blade has to survive.

Air enters a fighter engine, gets squeezed by the compressor until it is several hundred degrees hot from compression alone, gets fuel injected and lit, and emerges from the combustor at a temperature well above the melting point of any metal humans can currently make. That gas then hits the first row of turbine blades — small objects, roughly the size of a hand — which must extract enough energy from it to drive the compressor, while spinning at speeds that generate centrifugal loads of tens of thousands of times gravity.

The blades survive by three tricks. They are honeycombed with internal cooling passages so fine they cannot be drilled, only cast around a ceramic core that is later dissolved out. They are coated in a thermal barrier of ceramic thinner than a sheet of paper. And crucially, they are grown as a single crystal — one continuous, unbroken metallic lattice with no grain boundaries at all, because grain boundaries are exactly where hot metal under sustained load starts to creep and eventually tears.

Growing a single crystal turbine blade is a manufacturing process of almost absurd delicacy: molten superalloy is withdrawn from a furnace at a rate measured in millimetres per hour, through a spiral "grain selector" that permits precisely one crystal orientation to propagate. Get the withdrawal rate, the thermal gradient, the alloy chemistry or the mould chemistry slightly wrong and you get a blade that looks perfect and fails at nine hundred hours instead of nine thousand.

This is where the WS-15 spent years. Chinese and Western reporting through the 2010s recorded repeated setbacks, and skepticism about whether the programme would reach serial production was widespread inside China as well as outside it. Rumours of core redesigns and material failures circulated persistently. In a system with normal capital discipline, a programme burning that long with that little to show would have been cancelled twice over.

The warm-up act

What kept it alive — and this is the underappreciated part of the story — was the WS-10C reaching maturity first.

Through the late 2010s and early 2020s, WS-10 derivatives progressively displaced Russian AL-31s across the J-11, J-16 and J-10 fleets. This did two things at once. Operationally, it gave China's air force a domestically supplied engine for the bulk of its combat aircraft, which is most of the strategic benefit of self-sufficiency. Industrially, it did something more valuable: it forced Shenyang Liming to learn, at volume, how to make hot-section parts that hold up in squadron service. Metallurgy, coating processes, inspection discipline, blade yield — the WS-10C ramp was the school in which the WS-15's manufacturing base was educated.

De-risking a technology base by shipping the merely-adequate version at scale is a pattern any technology investor will recognise. It is what made the WS-15 possible.

December 2025

The company itself is legally barred from telling you what happened next. Under the disclosure exemption granted by the industry regulator, AECC Aviation Power's annual report explicitly states that for products involving national weapons and equipment, the company is exempt from disclosing model designations, capacity, orders, plans, prices, revenue, cost, profit, and technical and R&D data.¹ There is no product line item. There is no order book. There is no delivery number.

So the record has to be assembled from outside. China's state engine maker announced the start of WS-15 serial production in 2023, the same year a J-20 first flew with two of them installed.⁹ Janes reported preparation for mass production.¹⁰ Then, in December 2025, footage confirmed that the first batch of J-20s with twin WS-15 engines had completed serial production, with a first flight of a serial-production aircraft reported on December 27.[^11] By January 2026, Chengdu Aircraft Corporation had released images of J-20A airframes in acceptance-testing primer, and industry reporting indicated the WS-15 was expected to fully supersede the WS-10C in J-20 production during 2026.⁹

The performance claim, which cannot be independently verified, is that the WS-15 delivers at least 36,000 pounds of thrust with afterburner, against roughly 35,000 pounds for the F119 that powers the F-22 Raptor.⁹ Whether or not that number is precise, the qualitative point stands: this is a fully clean-sheet design rather than an enhanced derivative, and it is the first time Chinese industry has taken a high-thrust military turbofan from blank sheet to squadron service.

The scale of the runway, and the size of the prize

Outside estimates put the J-20 fleet at roughly 300 aircraft by mid-2025, with production running near 120 airframes per year and projections of approximately 1,000 aircraft by 2030.⁹ If those numbers approximate reality, they imply a multi-year, largely visible production ramp — two engines per aircraft, several hundred engines per year, plus spares, plus eventual overhauls.

This is the section that carries the emotional weight of the story, and it deserves the comparison it invites: this was China's Boeing 747 moment, a bet-the-institution programme on a technology nobody was certain could be made to work. Except the payoff took three times as long as Boeing's, and — this is the punchline that the rest of the analysis has to reckon with — the company that builds the engine barely profits from selling it.

Fifteen years of national effort, a genuine technological breakthrough, a production ramp with real visibility. And in the year the breakthrough reached serial production, the company's earnings fell by a quarter.

---

↑ Back to Top

VI. What the Company Actually Looks Like Today

On August 28, 2025, AECC Aviation Power published its interim report, and the numbers landed with a thud. First-half revenue of RMB 14.10 billion was down 23.99%. Net profit attributable to shareholders of RMB 91.78 million was down 84.57%. Stripping out non-recurring items, profit of RMB 38.41 million was down 92.97%.¹¹¹²

For a company whose flagship product had just entered serial production, an eighty-five percent earnings decline is a jarring artefact. Management's explanation was threefold: an adjustment in the timing of military orders, a slower-than-expected ramp in civil products, and delayed deliveries in the international subcontracting business. Financial expenses also climbed as interest-bearing debt grew.¹²

The full year recovered some of it but not all. Revenue of RMB 46.33 billion was down 3.23%; net profit of RMB 634 million was down 26.27%.¹ The second half was clearly much stronger than the first — which is consistent with a timing story rather than a demand story — but the year still ended lower on both lines.

What the company sells

Strip away the complexity and the business is one thing. Aero engines and derivative products generated RMB 43.48 billion of 2025 revenue, some 94% of the total, at an 8.50% gross margin.¹ This is military fighter, transport and helicopter propulsion out of the four merged plants, plus a growing sideline in 航改燃机 aero-derivative gas turbines — engines re-purposed as industrial and marine turbines for offshore platforms, emergency and distributed power, data centres and ship propulsion. Management confirmed at its April 2026 results briefing that civil gas turbine revenue is booked inside this same segment rather than separately.¹³

Two smaller businesses round it out. International subcontracting — machining parts for Western aerospace supply chains — brought in RMB 1.95 billion, down 5.94% on weaker export orders, but at a 20.61% gross margin, more than double the core.¹ Non-aviation products and other business contributed RMB 232 million, up 17.83%, at 25.61%.¹

There is a quiet lesson in those margins. The only parts of this company that earn a normal manufacturing return are the parts that sell into markets rather than into a procurement system. They are also, together, about 4.7% of revenue. Nobody should build an investment case on them, but they are a useful control experiment: the low margin in the core business is not a statement about AECC's manufacturing competence.

The number that should worry you

The income statement is not where the interesting risk lives. The balance sheet is.

At the end of 2025, accounts receivable stood at RMB 44.45 billion — up 24.44% in a year when revenue fell 3.23%, and equal to 36.12% of the company's entire asset base.¹ The company's own explanation is that a portion of the amounts had not yet reached settlement date.¹ Receivables turnover fell to 1.16 times from 1.69 times.¹ Third-party data puts days sales outstanding at 398 days for 2025, against 179 days as recently as 2021 and just 141 days in 2020.⁸

Add inventory of RMB 32.20 billion at year end — with inventory turnover slipping to 1.30 times — and the operating cycle stretches past 675 days. Net of payables, the cash conversion cycle worked out to roughly 445 days in 2025, against 155 days in 2021.¹⁸

Set against that, the advances the company receives from customers are modest: contract liabilities of RMB 5.50 billion at year end, up 22.59%.¹

The consequence is exactly what you would predict. Operating cash flow was negative RMB 4.42 billion in 2025 — an improvement on negative RMB 14.31 billion in 2024, but still an outflow, and it followed negative RMB 6.74 billion in 2023.¹ Free cash flow has been negative in four of the last five years.⁸ With capital expenditure running above eight percent of revenue and financing inflows of RMB 5.47 billion in 2025, the shortfall is being plugged with debt — which is why financial expenses jumped 37.35% to RMB 627 million, a figure now equal to roughly the entire net profit of the company.¹

Read plainly: this business is financing its customer's payment cycle out of its own balance sheet, at interest, and the interest bill has grown to consume the earnings. Accounting profit exists. Cash does not. And an auditor flagged receivable impairment as a key audit matter, noting a gross receivables balance of RMB 45.26 billion at year end requiring significant estimation judgment about recoverability.¹

To management's credit, the annual report does not hide this. Financial risk is named explicitly, with the acknowledgment that high levels of what Chinese industry calls 两金 — receivables and inventory — create periodic funding pressure, and a remediation plan built around inventory control, production scheduling, tighter collections, and cheaper policy-supported financing.¹ Whether it works is an empirical question with a quarterly answer.

The early read is genuinely mixed. Asked at the April 2026 briefing whether the swing to positive operating cash flow of RMB 3.37 billion in Q1 2026 was seasonal or sustainable, management attributed it primarily to lower cash paid out for purchases — a working-capital timing effect rather than a collections breakthrough — while restating the same list of remedies.¹³ That is a candid answer. It is not yet a solved problem.

One more thing the income statement obscures. Of the RMB 634 million of reported net profit, a large share was non-operating: subsidiary South Industry realised roughly RMB 260 million of investment gains selling shares in sister listed company 航发控制 AECC Control, and there were land and property compensation proceeds from the Anshun municipal government's takeback of Liyang's Ping'ba site.¹¹³ Underlying profit of RMB 294 million is the number that reflects the engine business.¹ Management stated it had built its 2026 budget around the non-repeatability of those gains and expects underlying profit to improve — a claim with a checkable answer in about eight months.¹³

That reliance on the parent's ecosystem for both revenue and, in a pinch, for gains, leads naturally to the question of who is actually running this company.

---

↑ Back to Top

VII. Current Management & Capital Allocation

On the afternoon of April 28, 2026, in a conference room at the Marriott on Shanghai's Yangpu waterfront, with a parallel webcast on the Shanghai Stock Exchange roadshow platform, AECC Aviation Power held its annual results briefing. Around twenty institutions dialled in or attended — 易方达 E Fund, 招商基金 China Merchants Fund, 南方基金 China Southern, 大成基金 Dacheng, 华宝基金 Fullgoal's peers among them, plus brokers including 招商证券 China Merchants Securities and 国信证券 Guosen.¹³

Note who was in the room from the company's side: board secretary and chief accountant 任立新 Ren Lixin, the head of the board office, the finance department director, and a securities affairs representative.¹³ Not the chairman. Not the general manager. For a company whose flagship programme had just entered serial production and whose earnings had fallen 26%, the most senior person facing investors was the CFO.

Who is in charge

Chairman 牟欣 Mou Xin, 55, took the role on December 5, 2024. His career reads as a group career: deputy general manager at AECC Harbin Dong'an Engine, then executive director and party secretary at Liyang Power, and — critically — he concurrently serves as board secretary of AECC group and head of its asset management department.¹

Read that concurrency carefully. The chairman of the listed subsidiary is simultaneously the person at the parent responsible for managing the parent's assets, including this one. The annual report's remuneration table records his pre-tax pay from the listed company as blank, with a "yes" in the column asking whether he receives compensation from a related party.¹ He is paid by the parent. His accountability runs upward.

General manager 沈鹏 Shen Peng, 52, is the operator. He came up inside the company — chief quality officer, then deputy general manager — and has served as director and general manager since September 2024.¹ His total pre-tax remuneration from the listed company in 2025 was RMB 866,000, about US$120,000.¹ Ren Lixin, who carries the extraordinary combined title of deputy GM, chief accountant, board secretary, general counsel and chief compliance officer, was paid RMB 816,600.¹

For the entire board and senior management team combined, 2025 pre-tax remuneration totalled RMB 9.42 million.¹ For context, that is less than the company's daily revenue run rate.

The ownership problem

Now the number that matters most for alignment. In the shareholding column of the directors and officers table — start of year, end of year, change — the entries read zero. For the chairman: zero. For the general manager: zero. For every executive vice president: zero. The only non-zero entry in the entire table belongs to a deputy general manager who departed on August 31, 2025, holding 5,900 shares.¹

Not a single serving director or executive officer of AECC Aviation Power owns a meaningful stake in it. There is no stock-based compensation of any kind; the reported figure is nil.⁸

This is not a scandal — it is entirely normal for a Chinese central state-owned enterprise, where senior appointments are made through party and group personnel processes and pay is set within state guidelines. But investors should be clear about what it means analytically. There is no owner-operator here. There is no one on the executive floor whose personal wealth moves with the share price. The people running this company are stewards executing a group plan, evaluated on delivery of national objectives, and they will be rotated to other posts within the system when the system decides.

Governance did tighten in one respect. During 2025 the company abolished its supervisory board in line with regulatory reform, transferring its functions to the board's audit committee, and revised five governance documents including the articles of association.¹ That is a modernisation of form; whether it changes substance is unproven.

Capital allocation is not a management decision

The FY2025 dividend proposal was RMB 0.72 per 10 shares, or roughly RMB 192 million against a share count of 2.67 billion — approximately 30% of reported net profit.¹ It was a cut from the RMB 0.97 per 10 shares distributed on the prior year.¹ At the current price that is a yield of about 0.19%.²

Almost everything else goes back into the ground. Management disclosed that capital expenditure across the 14th Five-Year Plan period approached RMB 16 billion, funded from a mix of state appropriations and self-raised money, directed at production line layout and capacity building for the full development-test-serial production chain.¹³ The FY2026 fixed-asset investment plans by subsidiary — RMB 1.21 billion at Liming, RMB 794 million at Liyang, RMB 221 million at South Industry, each with a substantial government-funded component — tell you the ramp is being built now.¹

The R&D line, meanwhile, is arresting. Total R&D investment in 2025 was RMB 885 million — RMB 773 million expensed, down 22.28%, plus RMB 112 million capitalised — equal to 1.91% of revenue.¹ For a company whose entire investment case is technological breakthrough, spending under two percent of sales on research looks impossible. The explanation is structural: in the Chinese defence system, engine design sits at dedicated research institutes and development is largely funded by the state, so the listed manufacturer's own R&D line captures process and production engineering rather than clean-sheet design. That is a legitimate explanation. It is also a warning about what this listed entity actually is. It is a manufacturer of designs it does not own, in a system where the intellectual property and the state funding both sit outside the listco.

The credibility test

Assessing management credibility here requires grading on the right curve. These executives do not set strategy, choose customers, negotiate prices, or decide capital allocation independently — the group's five-year industrial plan does. Judging them on capital allocation judgment is close to a category error.

What can be judged is disclosure quality and explanatory rigour, and here the record is mixed but improving. The FY2025 annual report attributes the revenue decline to "changes in customer demand, deliveries below expectation" and the margin decline to "the maturity of new products needing improvement."¹ Those are honest phrasings and, for a new engine entering serial production with early-life yield problems, plausible ones. But they are not specific, and they were not accompanied by any quantification — nor could they be, given the disclosure exemption.

More encouraging is the tone of the April 2026 Q&A itself. Analysts pushed on uncomfortable subjects — the Bank of Guizhou fair-value losses running through a subsidiary's books, the sustainability of one-off gains, the receivables balance, whether Q1 cash flow was real — and management answered directly rather than deflecting.¹³ That is worth something.

But the fundamental test has not been run. This team has not yet been through a full cycle of missing, explaining specifically, committing to a fix, and being measured against it. The phrase "order timing" is a hypothesis, not a finding. The next two reporting periods are where it gets tested.

The deeper reason management has so little room to manoeuvre is the structure of the industry it sits in — which is unlike almost any other in the world.

---

↑ Back to Top

VIII. Competitive & Industry Structure

Play a war game. You are a strategy team asked to build a competitor to AECC Aviation Power inside China. Where do you start?

You cannot. There is no path. The customer is the People's Liberation Army, which does not run open competitions for fighter propulsion. The designs originate at state research institutes assigned to AECC. The specialised metallurgy, the qualified supply chain, the security clearances, the test facilities — all sit inside the same state perimeter. A new entrant would need permission from the entity it intends to compete with.

This is not a moat. It is a decree.

The five forces, rewritten by the state

Run Porter's framework and watch four of the five forces get overwritten by policy.

Rivalry is essentially absent. Across the four merged plants, this company is effectively the sole manufacturer of PLA fighter, transport and helicopter engines. The nearest related listed company, 航发控制 AECC Control (000738.SZ), makes engine control systems — a supplier within the same group, not a competitor. AECC Aviation Power's own subsidiary held equity in it and sold some down for a gain in 2025.¹ That is not a competitive relationship in any meaningful sense.

Barriers to entry are absolute, and not for the usual reasons. Even with unlimited capital, a new entrant would need decades of empirical hot-section knowledge — the institutional memory discussed earlier — which cannot be bought and cannot be hired quickly.

Supplier power is low and structurally so. The top five suppliers accounted for 55.05% of 2025 purchases, of which related parties within the AECC system were 30.46%.¹ Roughly a third of what this company buys, it buys from its own family. Those are administered relationships, not negotiations.

Buyer power is where the whole edifice inverts. The top five customers accounted for 90.78% of 2025 sales — and of that, related-party sales within the AECC group were only 3.10% of the total.¹ So the concentration is not intra-group; it is the military procurement system. One buyer, or one buyer's several arms, takes nearly everything this company makes, sets the price via the pricing rules, and controls the settlement calendar. That is as much buyer power as exists anywhere in global manufacturing, and it is the direct cause of both the eight-percent gross margin and the 398-day receivables cycle.

Substitutes do not exist domestically. Which leaves the fifth force operating at a different level entirely: the real competitive contest is geopolitical, not commercial. AECC competes against GE Aerospace, Pratt & Whitney and legacy Russian propulsion — not for orders, but for capability parity.

The global benchmark, and the heart of the "why not" case

Here is where the comparison gets uncomfortable.

Global commercial propulsion is a genuine oligopoly: GE Aerospace, RTX's Pratt & Whitney, Safran through the CFM joint venture, and Rolls-Royce. What makes it one of the best businesses in industrial history is not the engines. It is what happens after.

An engine maker often sells the initial unit at thin margin or a loss, because the airline is buying a thirty-year relationship. Over those thirty years the engine returns to the shop four to six times for overhaul, and every one of those visits is a high-margin sale of proprietary spare parts and labour that only the manufacturer can legally and practically supply. The installed base becomes an annuity. Aftermarket work accounted for roughly 60% of Pratt & Whitney's 2023 sales.¹⁴ At GE Aerospace, full-year 2025 commercial services revenue rose 26% while LEAP deliveries rose 28%, and the Commercial Engines & Services segment margin reached 26.6%.¹⁵

Twenty-six point six percent, against AECC's eight and a half.

That single comparison contains both the risk and the opportunity in this name. AECC Aviation Power is, today, almost purely an initial-delivery manufacturer. It builds engines, ships them, and books cost plus a formula. It does not have — or does not disclose — a meaningful spare-parts and overhaul annuity of the Western kind.

Could it get one? In principle, yes, and this is the most interesting long-dated question about the company. Engines being delivered today will need overhaul in the 2030s. A PLA fleet of several hundred J-20s plus the J-10, J-11, J-16 and transport fleets is exactly the kind of aging installed base that generates sustainment demand. And a domestic civil fleet, if the CJ-1000A ever certifies and scales, would add decades more.

But three things must be true for that to become an earnings event, and none is yet established. The maintenance work must actually be booked at the listed manufacturer rather than at military depots or separate group entities. It must be priced on a basis that permits aftermarket-style margins rather than the same cost-plus formula. And it must be disclosed clearly enough for investors to see it. Not one of those three has been demonstrated. Anyone modelling an aftermarket inflection into AECC's numbers should be explicit that they are modelling a hope.

It is also worth noting that even the incumbents are unhappy with the model. RTX leadership has publicly questioned whether investing heavily up front and relying on shop visits over a 25-year period remains the right structure for next-generation single-aisle engines, and is looking at capturing more value at delivery.¹⁶ The annuity model is durable but not immutable.

Which power does AECC actually have?

Run Hamilton Helmer's 7 Powers over the business honestly and most of them fail.

Scale economies: real, but they translate into cost, and cost translates into a lower price under a cost-plus formula rather than into profit. Network economies: none. Switching costs: nominally infinite, but with one state buyer and a state supplier, the concept is meaningless. Brand: irrelevant. Counter-positioning: none — there is no incumbent to counter-position against. Cornered resource: partially, in the accumulated metallurgy and single-crystal know-how, and in the exclusive designation as the group's manufacturing vehicle.

Process power comes closest. The decades of accumulated engineering knowledge — how to cast a blade that survives, how to inspect what you cannot see — are genuinely non-replicable and genuinely valuable. Combined with state-granted exclusivity, the position is durable in a way that few businesses anywhere can match.

But durability and profitability are different properties. AECC Aviation Power has an extraordinarily durable position that converts into a return on invested capital of about 1.05%.⁸ A moat that protects a puddle is still a moat; it just does not make you rich.

That gap between strategic centrality and financial return is exactly the space in which the market prices optionality — and the largest single option sits outside this company's own income statement.

---

↑ Back to Top

IX. The Civil Aviation Optionality: CJ-1000A

There is a Y-20 transport aircraft in China with an unusual asymmetry: three of its engines are the standard type, and hanging from the fourth pylon is something visibly different. That fourth engine is the 长江-1000A CJ-1000A, China's clean-sheet large-bypass commercial turbofan, and it has been flying that way since 2023 as a testbed for the engine intended to replace the Franco-American LEAP-1C on the C919 narrowbody.

Two structural points before any excitement. First, the CJ-1000A is not developed by AECC Aviation Power. It sits with 中国航发商用航空发动机 AECC Commercial Aircraft Engine, a sister entity inside the same group. Second, whatever revenue AECC Aviation Power derives from it today is buried inside the aviation engine segment as component supply, and it is small. Management confirmed at the April 2026 briefing that it does not separately disclose purchases from the commercial engine unit; total sales to all entities within the AECC system in 2025 were RMB 1.363 billion — under 3% of revenue, and that figure covers everything, not just CJ-1000A work.¹³

What management does say is directional and appropriately hedged: the company is an important supplier in commercial propulsion and the principal manufacturer of complete general aviation engines; the strategic importance of the civil business will rise progressively; and business scale is expected to enter a rapid growth phase after domestic commercial engine models are type-certified and reach serial production.¹³

That is a conditional statement with the condition doing all the work.

Where the programme stands

Chinese reporting through the first half of 2026 described substantial progress: 317 airworthiness compliance subjects completed and around 6,142 hours of testing accumulated across ground test stands, altitude test facilities and flight testing, covering plateau, icing and bird-strike conditions. A technical review was held in April 2026 and a final review in May, with the programme targeting a CAAC type certificate. The stated plan called for a first C919 fitted with CJ-1000A engines to go to China Eastern for real-world route testing, batch installation from 2027, and large-scale commercial operation around 2030.¹⁷

Treat those milestones with appropriate caution. They are drawn from domestic reporting rather than from a regulator's certificate, and Chinese aerospace timelines have a long history of arriving late. As of this writing, no CAAC type certificate for the CJ-1000A has been publicly confirmed.

How to size it

The right way to hold this is as an option, not a forecast.

The strategic logic is genuine. The C919 today flies on the LEAP-1C from CFM — the GE-Safran joint venture — which means China's flagship commercial aircraft programme depends on a Western supply chain for its propulsion, precisely the dependency that the whole AECC project exists to eliminate. Export controls have made that dependency feel considerably less theoretical than it did a decade ago. Substituting a domestic engine is not a commercial preference; it is the same self-sufficiency imperative that drove the WS-15, applied to the civil side.

And if it works, the prize is the thing AECC does not currently have: a genuine aftermarket. Commercial engines fly thousands of hours a year in the hands of airlines who pay market prices for overhauls. That is the business model examined in the previous section, and it is the only plausible route by which a Chinese engine manufacturer eventually earns Western-style margins.

But the honest accounting today is simple. It is pre-revenue at this listed company. It is developed by a sibling. Its certification is not confirmed. Its commercial ramp, on the most optimistic published schedule, is late this decade. Nothing in AECC Aviation Power's 2025 or 2026 numbers depends on it.

It deserves exactly one clear mention in any investment case, and no more, because the thing actually moving this company's revenue, its capacity plans and its share price is military — and military is where the bull and bear cases collide.

---

↑ Back to Top

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

The share price tells the argument better than any summary. Within the past year AECC Aviation Power traded as high as RMB 63.78 and as low as RMB 31.86, closing at RMB 37.01 on August 18, 2026 — a peak-to-trough round trip of roughly half.² That is not noise in a RMB 99 billion state-controlled monopoly. That is the market repricing a thesis in real time, and the volatility is evidence about how much of the valuation rests on belief rather than on earnings.

Why it wins from here

The bull case has four legs, and three of them are solid.

The first is structural position, and it is the strongest thing about the company. Sole-supplier status by government design is not a competitive advantage that can erode. There is no disruptor, no low-cost entrant, no substitute technology. Whatever the PLA spends on fighter, transport and helicopter propulsion over the next twenty years flows through these four plants.

The second is a production ramp with unusual visibility. The WS-15 reached serial production and, on outside reporting, is displacing the WS-10C in J-20 output during 2026 — against a fleet already near 300 aircraft and building at roughly 120 a year toward perhaps 1,000 by 2030.⁹ Most industrial companies would kill for a demand picture that legible.

The third is fiscal support. China set its 2026 defence budget at roughly RMB 1.91 trillion, up 7% year on year — the eleventh consecutive year of single-digit growth, but still growth well above nominal GDP trend in a fiscally constrained year.¹⁸ Notably, this was the slowest increase since 2021.¹⁹ The direction is right; the acceleration is not there.

The fourth leg is the weakest, and it is worth being precise about why. Contract liabilities — customer advances — rose 22.59% to RMB 5.50 billion in 2025, which bulls read as forward order visibility.¹ It is a genuine positive signal. But scale it: RMB 5.5 billion of advances against RMB 46.3 billion of revenue is roughly six weeks of sales. Against RMB 44.4 billion of receivables, it is small. The customer is not prefunding this ramp; the company is.

Layer on the strategic-imperative argument — that ending foreign engine dependency insulates this programme from ordinary budget cyclicality — and the bull case is coherent. This is a mission, and missions get funded.

Why it may not

The bear case is not that the mission fails. It is that the mission succeeding and the shareholder being rewarded are different events.

Start with the ceiling. Margins are capped by defence pricing policy, and the 2019 reform raised the ceiling modestly without removing it. No amount of WS-15 volume changes the formula. A bull who models margin expansion has to explain what mechanism produces it, and the honest answer today is that only two exist — an aftermarket that is not yet visible, and a civil business that is not yet certified.

Then the cash. Four straight years of deteriorating cash conversion, negative free cash flow in four of five years, and a financing cost that has grown to swallow the profit line.¹⁸ The company is not converting its strategic win into cash, and the longer the ramp runs, the more working capital it absorbs.

Then the disclosure asymmetry. FY2025 demonstrated that delivery timing can move net income by double digits — 84.57% in a half year — with essentially no advance warning, because the company is legally forbidden from disclosing orders, capacity or delivery plans for military products.¹¹² An investor in this name is structurally unable to see the most important operating variable until it appears in a printed report.

Then alignment. Zero management shareholding, no equity compensation, executives appointed and rotated by a parent, and a chairman paid by that parent.¹ Nothing here is improper. But nobody inside the company is compensated for the share price, so nobody inside the company is optimising for it.

And finally, valuation. Enterprise value to EBITDA has ranged from roughly 28 times in 2023 to 60 times in 2021, sitting near 37.5 times for 2025 — against a return on invested capital of about 1.05% and a return on equity of 1.58%.⁸¹ The trailing price-earnings multiple is above 150.² Those multiples do not price a manufacturer earning one percent on capital. They price years of flawless future execution.

The activist stress test

There is no activist here and there never will be — majority state control forecloses the possibility entirely, and no proxy fight, board seat or breakup proposal is available. But the activist questions still discipline the analysis, and they are worth asking out loud.

An adversarial investor would open with the receivables. Nearly RMB 45 billion of gross receivables, growing at 24% while revenue shrinks, owed by a state procurement system, with the auditor treating impairment as a key audit matter and management describing recovery as an area for "assessment, incentive and accountability."¹ Where exactly is the ageing schedule, and what is the collection experience by vintage?

Second, on earnings quality: a majority of 2025's reported profit came from selling shares in a sister company and from a municipal land takeback.¹¹³ Management deserves credit for stating plainly that its 2026 budget assumes those do not repeat.¹³ But an investor should note the pattern — that in a difficult operating year, the reported number was rescued by asset disposals within the state family.

Third, on related-party exposure: 30.46% of purchases from group entities at administered prices, with independent verification unavailable.¹

Fourth, and most bluntly: what is the credible path from a sub-2% return on equity to anything that justifies a defence-technology multiple, on a timeline an investor can underwrite? Management's own answer at the results briefing was that underlying profit should improve in 2026 once the one-offs are stripped out, driven by organic growth in the core engine business.¹³ That is a testable claim, and it is the right one to hold them to. But "improvement from RMB 294 million" is a long way from an earnings profile that supports the current multiple.

The fair summary is that the strategic case and the financial case are, at this moment, disconnected. The strategic case is close to unassailable. The financial case rests entirely on multi-year execution — WS-15 volumes scaling, cash conversion normalising, and eventually either an aftermarket or a civil franchise emerging. Every one of those is unproven.

---

↑ Back to Top

XI. Risk Radar

Skip the generic risk taxonomy. Five things could genuinely break this story, and each has a specific mechanism.

Order-timing and single-customer risk. With 90.78% of revenue from five customers, essentially all of it military, this company's quarterly results are a function of one buyer's procurement calendar.¹ FY2025 was the demonstration: a first half where profit fell 84.57%, then a recovery that still left the year down 26%.¹¹² The mechanism is not demand loss — the aircraft still get built — but the recognition timing sits with someone else, and the company cannot warn investors about it because model-level disclosure is legally exempt.¹ Expect this to recur.

Working-capital and financing risk. The 445-day cash conversion cycle means the company is effectively extending long-dated, unsecured, interest-free credit to its own controlling shareholder's procurement system while paying interest on the money it borrows to do so.⁸ Financial expenses rose 37.35% in 2025 on a larger interest-bearing debt load and foreign exchange losses.¹ The mechanism to watch: if receivables keep growing faster than revenue, the borrowing grows, the interest bill grows, and it consumes an ever-larger share of an already thin profit line. At current levels, finance costs alone are approximately equal to net income.

Execution risk. The WS-15 has entered serial production, not full-rate mature production, and the company's own explanation for its 2025 margin compression was that new product maturity needs improvement.¹ That is the honest language of early-life manufacturing: yields on hot-section parts are below steady state, scrap is high, and unit costs are above target. Under an incentive-pricing regime where the supplier absorbs seventy percent of a cost overrun, immature production is directly a margin problem.⁷ Any CJ-1000A transition would compound it with a second learning curve.

Policy risk, running in both directions. Further reform of defence pricing could compress margins if the state decides contractors are over-earning — a live possibility given the reform's stated aim of squeezing cost. Simultaneously, geopolitical tension around Taiwan and Western export controls on advanced materials and semiconductors raise both the urgency and the cost of self-sufficiency: controls that deny access to Western machine tools, inspection equipment or specialty alloys make every engine harder and more expensive to build, while making the programme more politically untouchable. Investors should not assume these effects net out; they operate on different lines of the P&L.

Worth noting too: China's 2026 defence budget growth of 7% was the slowest since 2021, set against a broader fiscal environment analysts have described as increasingly constrained.¹⁸¹⁹ Defence spending is a political priority, not a legal entitlement.

Governance risk. The 2019–2021 asset injection was priced by appraisal rather than by market. Roughly a third of purchases are from group entities. Board and management shareholding is nil. The chairman is paid by the parent and concurrently manages the parent's asset portfolio.¹ Individually, each is unremarkable in a Chinese SOE. Collectively, they mean minority shareholders depend on the state parent's continued fairness rather than on any enforceable market discipline. The specific forward-looking exposure: future asset injections from AECC group into the listco — which several analysts expect — would again be priced without independent discovery, and could dilute existing holders on terms they cannot contest.

One risk that deserves to be dismissed rather than listed: technological disruption. Electric propulsion is irrelevant at fighter scale for the foreseeable future, and no software or AI development threatens the business of casting turbine blades. This is one of the few industrial franchises in the world where the disruption question is genuinely not the point.

---

↑ Back to Top

XII. Durable Lessons & What to Watch

Three things generalise beyond this company.

Lesson one: understand the parent's restructuring intent before you value the subsidiary. Between 2016 and 2021, AECC Aviation Power's scope changed more from paperwork than from products. A State Council decision to carve engines out of AVIC, followed by a share-issuance roll-up of three sister plants, transformed a single Xi'an factory into a national monopoly inside a listed shell. An investor who studied the company's products in 2015 and again in 2022 would have been studying two entirely different businesses. In state-controlled corporate structures, the parent's five-year plan is a more important document than the subsidiary's annual report — and future injections can change the story again without any shareholder vote that matters.

Lesson two: national-champion defence stocks trade on strategic narrative, not near-term earnings — and the gap is measurable. Here it is quantified precisely: an enterprise value near 37 times EBITDA against a return on invested capital of roughly 1%.⁸ That spread is not an analytical error. It is the market explicitly saying that today's earnings are not the point, and that it is paying for a claim on a future that has not yet appeared in any financial statement. That can be a perfectly rational position. But it should be held consciously, because it means the investment is a bet on a narrative remaining intact, and narratives reprice fast — as a fall from RMB 63.78 to the low thirties demonstrated.²

Lesson three: in aerospace propulsion, the difference between selling equipment and servicing it is the whole business. GE's commercial services margin of 26.6% against AECC's 8.50% engine margin is not a story about Chinese manufacturing being worse.¹⁵¹ It is a story about where in the value chain the profit sits. Any long-term case for AECC Aviation Power is, when you strip it down, a bet on whether it can eventually cross from one side of that divide to the other — and that requires the installed base to age, the maintenance revenue to be booked at this entity, and the pricing regime to permit a normal return on it. Three conditions, none currently met.

The KPIs that actually matter

Most metrics for this company are noise. Three are not, and a reader tracking them quarter by quarter will know more than someone reading the headlines.

One: delivery cadence and the timing of military order recognition. This is the variable that broke in 2025 and the only real driver of near-term revenue. It cannot be observed directly — the disclosure exemption sees to that — so it must be inferred from the shape of half-yearly and quarterly revenue and from whether management's "order timing" explanation stops recurring. If a strong second half repeatedly rescues a weak first half, that is a structural feature of the procurement calendar, not a one-off, and it should be modelled as such.

Two: the receivables and cash conversion trend. Watch the direction of days sales outstanding and the cash conversion cycle, and specifically whether receivables growth reconverges with revenue growth. This single metric determines whether the WS-15 ramp turns into cash or into an ever-larger loan to the state. Management has committed to collections discipline; the number will show whether it worked.

Three: WS-15 full-rate production displacing the WS-10C. This is the physical milestone underneath everything. Because the company cannot disclose it, it must be tracked through defence-industry reporting and airframe production evidence. The distinction that matters is between limited-rate and full-rate production — the former is a technical achievement, the latter is an industrial one, and only the latter produces the volumes the bull case requires.

---

↑ Back to Top

XIII. Epilogue

Return to those Xi'an workshop halls, and to the strange position the company occupies in the summer of 2026.

By any historical measure, AECC Aviation Power has just delivered the thing it was built to deliver. Fifteen years after the WS-15 programme began, and roughly seventy years after Soviet engineers first laid out Factory 430, China has a clean-sheet high-thrust turbofan in serial production under its most advanced fighter. The engine dependency that shaped Chinese air power doctrine for three decades has, on the available evidence, ended.

And the stock is down more than forty percent from its high, on earnings that fell by a quarter, with a company financing its own customer at a cost that eats its profit.

What the next two to four quarters need to show is narrow and specific: whether 2025's dip was timing or a new baseline. If half-year and full-year 2026 results show revenue recovering with the aero-engine segment gross margin stabilising or rising as WS-15 yields improve, the "new product maturity" explanation will have been validated and the thesis holds. If revenue growth returns but margins keep slipping, the more troubling interpretation gains ground — that the new engine is structurally harder and more expensive to build than the one it replaces, and that under an incentive-pricing regime the manufacturer absorbs most of that difficulty.

Beyond the numbers, four things are worth watching. Whether AECC group injects further assets into the listed company, and on what terms. Whether the payout policy shifts — the FY2025 cut to RMB 0.72 per 10 shares from RMB 0.97 signalled reinvestment priority, and a reversal would signal something about the cash outlook.¹ Whether the CJ-1000A actually receives its type certificate and on what schedule. And whether WS-15 volumes scale toward the production rates the outside estimates imply.

There is also a smaller, more interesting signal. In February 2026 the company established 航发通航动力科技（上海）有限公司 AECC General Aviation Power Technology (Shanghai), consolidating general-aviation propulsion products currently spread across South Industry into a single contracting and delivery platform, explicitly framed as an entry into China's 低空经济 low-altitude economy policy push.¹³ Alongside the AES100 turboshaft receiving its production certificate, the AEP100 powering the first flight of the W5000 heavy cargo drone, and a QD280-based gas turbine genset achieving first grid connection on a national demonstration project, this is management trying to build something the group plan does not dictate: a second growth curve with commercial customers and commercial margins.¹³ It is early, it is tiny, and it may amount to nothing. It is also the only part of the story where this company gets to act like a company.

Which leaves the essential tension unresolved, and honestly so. AECC Aviation Power is a business whose entire investment case rests on an engine that took fifteen years to build, in a country that decided the engine mattered more than the return. The strategic achievement is real. The financial statements still look like a factory — high fixed assets, thin margins, ballooning receivables, cash going out — rather than like a franchise. Whether the factory ever becomes a franchise depends on things that have not happened yet: an aftermarket that does not exist, a civil engine that is not certified, and a pricing regime that would have to change.

The market is paying franchise prices for factory economics. That may eventually be vindicated. Right now it is a wager, and it should be held as one.

---

↑ Back to Top

XIV. Outro & Links

For readers who want to go deeper, the primary sources reward the effort more than the commentary does. The FY2025 annual report is unusually candid about working capital risk and states the disclosure exemption for military products plainly, which tells you exactly what you are not being told.¹ The 2019 restructuring plan lays out the debt-to-equity mechanics and the appraisal-based pricing framework in detail.⁵ The April 2026 investor activity record is the best available window on how management handles adversarial questions, and covers cash flow, non-recurring gains, capital expenditure and the general-aviation subsidiary in the analysts' own words.¹³ The interim report for 2025 remains the clearest illustration of how violently this company's half-year earnings can swing.¹¹

References

1. 中国航发动力股份有限公司2025年年度报告 (FY2025 Annual Report) — Shanghai Stock Exchange filing, 2026-04-02 ↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩

2. AECC Aviation Power Co., Ltd (SHA:600893) Stock Overview — StockAnalysis, 2026-08-18 ↩↩↩↩↩↩↩↩

3. 西安航空发动机公司 — 维基百科 (company history: Factory 430, 1958 founding, Spey MK202, WS-9/WS-10/WS-15) ↩↩↩

4. 军工央企中国航发悄然挂牌：注册资本500亿，北京国资入股 — 澎湃新闻 The Paper, 2016-06 ↩↩

5. 中国航发动力股份有限公司发行股份购买资产暨关联交易预案 — 2019 major asset restructuring plan, 2019-07-20 ↩↩↩↩↩

6. AECC Aviation Power completed the acquisition of stakes in Shenyang Liming, Guizhou Liyang and AECC South Industry — MarketScreener ↩

7. 深度解析军品定价与采购机制改革 — 格隆汇 Gelonghui ↩↩↩

8. AECC Aviation Power key metrics and ratios, FY2020–FY2025 — Financial Modeling Prep data via StockAnalysis ↩↩↩↩↩↩↩↩↩↩

9. China Is Now Building J-20A Stealth Fighters Around the Engine It Waited 15 Years to Perfect — 19FortyFive, 2026-07 ↩↩↩↩↩

10. Chinese WS-15 engine prepared for mass production — Janes ↩

11. 航发动力：2025年半年度报告 (H1 2025 Interim Report) — 新浪财经 filing index, 2025-08-28 ↩↩

12. 航发动力：2025年半年度净利润约9178万元，同比下降84.57% — 每日经济新闻 NBD, 2025-08-28 ↩↩↩↩

13. 中国航发动力股份有限公司投资者活动记录表 — 2025年度业绩说明会, 2026-04-28 ↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩

14. Pratt & Whitney looks to tweak business model, with more money upon delivery — FlightGlobal, 2026-07 ↩

15. GE Aerospace FY and Q4 2025 Earnings Thrust Higher Propelled by Services Growth, LEAP Volume and Expanding Margins — Leeham News, 2026-01-22 ↩↩

16. RTX 2025 Earnings: Commercial Aerospace Leads Growth as Pratt Advances GTF Recovery — Leeham News, 2026-01-27 ↩

17. CJ-1000A取证在即，C919何时换上"国产心" — 网易 NetEase, 2026 ↩

18. China's 2026 defense budget growth slows to 7 percent — Xinhua, 2026-03-05 ↩↩

19. China to boost defense spending by 7%, slowest pace since 2021 — CNBC, 2026-03-05  ↩↩