ACM Research
Makes the machines that wash and copper-plate silicon wafers between the steps of making a chip, and sells almost all of them to factories in mainland China.
The scores
Growth3/4 ↑
Guidance rose to 25%-30% for FY2026 and orders grew 105%, but the flagship cleaning line fell 14% year over year in Q2 2026.
Business quality3/4 →
An 87% repeat-order rate signals a real qualification-based switching cost, but a named local China entrant is a live, growing threat.
Financial position2/4 ↓
The plan needs outside money already raised once (May 2026) and being raised again via the Hong Kong listing, while operations burn cash.
Execution record3/4 →
Beat consensus EPS in 6 of the last 8 quarters and raised guidance twice this year, but PECVD and Track slipped to year-end again.
Per share capture2/4 ↓
A shrinking 74.6% subsidiary stake and a below-market equity raise both dilute ACM Research's per-share economic claim.
Potential3/4 ↑
A credible $4 billion long-term target rests on PECVD, Track and panel-level plating, which are still mostly pre-production.
Risk profile2/4 →
Historic revenue is overwhelmingly China-sourced, and both operating subsidiaries now sit on the U.S. Entity List.
Valuation3/4 →
22x TTM earnings is level with ACM's own three-year median near 21x, so it asks less than at 32x, but cash burn bars cheap.
Evidence quality3/4 →
Filings, transcript and FINRA data cross-check cleanly with no vendor/filing gaps, but the $4 billion target and SAM rest on management-selected third-party data.
If it works 1.8× · if it fails you keep about 100%.
The Verdict
We rate ACM Research neutral. It builds machines that clean and plate silicon wafers, and customers keep coming back, but nearly all its sales go to mainland China, its arms sit on a US trade blacklist and its profit has not turned into cash, so the business is mixed. The price is cheap against its larger rivals and now level with its own past, so it is fair. As a bet it is defensive: the new tool lines add little if they win, and little is lost if they never sell. The view turns on those new tools winning production orders and on cash finally matching the profit.
The long caseChina wants its chip plants to buy home-made tools, and deposition and coating are big markets; if ACM's new tools pass customer trials, buyers who already trust it can order.
The evidence under this - 9 claims
- 2026-10-23editedACM's own notice of 19 October sets the third-quarter results for 5 November, not the 4 November a vendor estimated; the refresh date moves with it
- 2026-10-23editedre-struck at the 22 October close from 31 August: ACM's EV/EBIT 29.1x to 19.1x against a peer median of 42.1x to 30.65x; Lam's refreshed growth (33.4%) now exceeds ACM's 27.3%, so ACM is second fastest, not fastest
A related new platform, horizontal panel-level plating, already booked its first production order in Q2 2026 (C35), which management points to as evidence the evaluation-to-production path is real rather than a standing promise.
- 2026-10-23editedACM's own notice of 19 October sets the third-quarter results for 5 November, not the 4 November a vendor estimated; the refresh date moves with it
A skeptic could read the same discount as the market correctly pricing structurally lower returns on capital (ROIC 4.4% vs a 23.5%-40.8% peer range, C25) rather than as a temporary gap that closes as growth continues.
- 2026-10-23editedre-struck at the 22 October close: 29.1x EV/EBIT to 19.1x, with the cohort median 42.1x to 30.65x; the reading moves from 'a narrowing discount' to a wider one, with the business evidence unchanged
- 2026-10-23editedre-struck at the 22 October close (49.12) from the 31 August close (72.82): P/E 32.25x to 22.47x and EV/EBITDA 14.83x to 9.92x, so ACM no longer trades above both of its own medians
The by-product-line SAM figures are described as based on 'third-party reports and customer information' without naming which third parties beyond the general Frost & Sullivan citation used elsewhere, so the individual product-line splits cannot be independently checked against a named external source.
The company
A chip is built up in hundreds of steps, and between many of them the silicon wafer must be washed free of specks far too small to see. ACM Research makes machines that do that washing one wafer at a time, replacing older tools that dip whole batches into a bath.
Cleaning was about 70% of 2025 sales. The rest is copper-plating machines, furnaces, packaging equipment and two newer lines called PECVD and Track.
The evidence under this - 2 claims
How it makes money
A chip factory first tests a tool on its own line, then buys it outright, and spare parts and service follow. The next sale comes when that factory adds capacity, or a sister factory copies a line that already works.
The evidence under this - 2 claims
Where the money comes from, stream by stream
| Stream | Share of sales (FY2025) | How it earns it |
|---|---|---|
| Single-wafer cleaning, Tahoe and semi-critical cleaning | 69.5% | Wet-cleaning tools sold to chip fabs, mostly in mainland China, that strip particles and residue between process steps. |
| ECP (front-end and packaging), furnace and other technologies | 22.1% | Copper-plating tools for chip wiring and packaging, plus furnaces and newer deposition tools, sold per system. |
| Advanced packaging (excluding ECP), services and spares | 8.4% | Wet tools for chip packaging lines, plus spare parts and service on the installed base. |
Who runs it
David Wang founded ACM Research in 1998 and still runs it; the same finance chief has served since 2019, and the CEO's brother runs the Chinese subsidiary. Through shares with twenty votes each, Wang alone controls most of the vote, so outside holders cannot outvote him. Bonuses have no formula: the board decides, citing sales growth and margins. Long-term pay is stock options, which pay only if the share price rises.
The evidence under this - 3 claims
The CD&A credits 'improved operating margins' for 2025, while the filed operating margin fell from 19.3% in FY2024 to 12.1% in FY2025 (S01); the proxy does not say which margin it meant, which is what a pay plan with no stated measure allows.
Why it wins, and who could beat it
Why customers come back
A chip factory spends months approving a tool against its own recipe. Once a tool passes, the next order goes to the same machine, because a rival's would have to pass all over again.
The trace is in the orders: more than 1,255 of the 1,435-plus tools shipped since 2009 were repeats, and the plating line went from its 500th chamber in 2022 to its 2,000th this year. Lam Research, the largest rival, credits its own installed base with the same pull in its annual report.
The evidence under this - 2 claims
The threat
NAURA, a Chinese toolmaker, sells cleaning, plating and furnace machines to the same mainland factories, and ACM's own annual report warns of more local entrants. NAURA grew strongly in 2025 on China's push to buy home-grown tools, though a trade paper saw thinner margins ahead.
The evidence under this - 3 claims
The rivals: LRCX, AMAT, and 1 more that is private, in-house or listed abroad.
Who else does this, rival by rival
| Ticker | Rival | Where it competes | How far along | Priced against it |
|---|---|---|---|---|
| NAURA Technology GroupPriced against it: Listed abroad | Cleaning, ECP plating, furnace and advanced-packaging tools | On sale | Listed abroad | |
| LRCX · AMAT | Lam Research, Applied Materials and Suzhou JingtuoPriced against it: Yes | PECVD and Track platforms | On sale | Yes |
How big it could get
How big this could get
The world's chip factories will spend about $138.5B on equipment in 2026. ACM reckons its current machines can bid for $7.3B of cleaning, $5.3B of PECVD, $3.0B of Track, $2.6B of furnaces and $1.5B of plating. It sold $901.3M last year.
The catch is where. Mainland China's equipment spending is forecast to fall 9.9% in 2026, from $39.3B to $35.4B.
The evidence under this - 4 claims
The by-product-line SAM figures are described as based on 'third-party reports and customer information' without naming which third parties beyond the general Frost & Sullivan citation used elsewhere, so the individual product-line splits cannot be independently checked against a named external source.
Who has signed
Very little. Last quarter brought the first disclosed orders for a tool that plates large square panels instead of round wafers: one production order and one trial machine, both from unnamed customers and neither with a value.
The evidence under this - 1 claim
Where the growth is coming from, business by business
Where the growth is coming from
| Business | Share of revenue | Growth: last quarter | Last half-year | Last year |
|---|---|---|---|---|
| Total Single Wafer and Semi-Critical Cleaning Equipment | 45% | -14.2% | -10.2% | -0.6% |
| ECP Front End And Packaging Furnace And Other Technologies | 44% | +167.7% | +181.3% | +107.9% |
| Advanced Packaging (exclude ECP), Services & Spares | 11% | +153.3% | +103.1% | +98.5% |
| Whole company | 100% | +36.0% | +35.2% | +27.3% |
The filed split through 2026-06-30. These rates are not carried forward: the reading below says why.
How we read it: The blend. The fast line is plating, an established product, not PECVD or Track, still in evaluation; its 181% rides a weak 2025 base and China capacity spending SEMI expects to fade.
The results
The numbers
Sales grew 27% over the twelve months to Jun 2026, faster than the 15% in the year to Dec 2025.
Operating margin is 13%, down from 17% in the year to Dec 2023.
Running costs take 31¢ of each sales dollar, up from 29¢ in the year to Dec 2021.
Each dollar put into the business earns 7.1% a year after tax, about where it was (8.6%) in the year to Dec 2021.
It used $119M of cash over the last twelve months, so the $1.07B it holds lasts about 9.0 years at that pace.
It holds $1.07B of cash against $341M of debt, $733M more cash than it owes.
The share count rose 9.9% in 4.5 years, and pay in shares costs 2.5% of sales.
All the numbers
Filed figures from the company's 10-K and 10-Q, read 2026-10-04. Fiscal years end in December.
| Dec 2021 | Dec 2022 | Dec 2023 | Dec 2024 | Dec 2025 | 12 mo to Jun 2026 | |
|---|---|---|---|---|---|---|
| Size and growth | ||||||
| Sales | $260M | $389M | $558M | $782M | $901M | $1.04B |
| Sales growth | +66% | +50% | +43% | +40% | +15% | +27% |
| Profit | ||||||
| Gross margin | 44% | 47% | 50% | 50% | 44% | 44% |
| Operating margin | 15% | 15% | 17% | 19% | 12% | 13% |
| Net margin | 15% | 10% | 14% | 13% | 10% | 14% |
| What growth costs, per sales dollar | ||||||
| Research and development | 13% | 16% | 17% | 13% | 16% | 16% |
| Sales and marketing | 10% | 10% | 8.4% | 8.4% | 8.5% | 8.0% |
| General and admin | 5.9% | 5.8% | 7.3% | 8.9% | 7.6% | 6.9% |
| Pay in shares | 2.0% | 2.0% | 4.9% | 6.3% | 3.7% | 2.5% |
| Capital spending | 3.5% | 23% | 11% | 11% | 6.2% | 11% |
| Cash | ||||||
| Cash from operations | −$40.1M | −$62.2M | −$75.3M | $152M | −$10.3M | −$6.60M |
| Free cash flow | −$49.2M | −$153M | −$137M | $70.0M | −$66.6M | −$119M |
| Balance sheet | ||||||
| Cash and short-term investments | $592M | $268M | $203M | $427M | $793M | $1.07B |
| Debt | $35.0M | $77.0M | $92.1M | $183M | $288M | $341M |
| Net cash | $557M | $191M | $111M | $244M | $505M | $733M |
| Returns and shares | ||||||
| Return on capital | 8.6% | 6.1% | 9.9% | 12% | 6.9% | 7.1% |
| Shares (diluted) | 65.4M | 65.3M | 64.9M | 66.2M | 67.3M | 71.8M |
What growth costs
Growing has not made ACM Research cheaper to run. Its spending on research, selling and administration has stayed about the same share of every sales dollar while sales more than tripled, and management plans to keep it there. Profit margins have risen and fallen with the price it gets for its tools, not with its size.
The evidence under this - 2 claims
Operating margin did widen from 14.9% in FY2021 to 19.3% in FY2024 before falling to 12.1% in FY2025, but the move tracked gross margin (44.2%, 50.1%, 44.4%), not operating cost; and FY2025 R&D includes $26.7 million of components for development tools (S01), lumpy rather than structural.
Management's 'no change' framing came the same quarter the BIS Entity List restriction was already in effect and component costs were described as rising; the reaffirmed range has not yet been tested through a full quarter of the local-supplier substitution management described.
- 2026-10-23editedACM's own notice of 19 October sets the third-quarter results for 5 November, not the 4 November a vendor estimated; the refresh date moves with it
The last report (7 Aug)
Quarter ended June 2026, filed figures.
The June quarter looked strong: sales up 36%, and the full-year forecast raised to $1.125B to $1.175B.
Underneath, the flagship cleaning line shrank 14%; the growth came from plating and packaging. Reported earnings per share of $1.23 were double the company's own adjusted $0.61, because of a $69.6M paper gain on investments. And the cash told a third story: operations used $6.4M, the fifth quarter out of six in which they consumed cash.
What it did to the thesisIt showed the growth is real but sits where the thesis said: in the newer plating and packaging tools, while the flagship cleaning business shrank. Earnings were flattered by a paper gain and cash still went out, so the view stays Neutral until a new line wins production orders.
The quarter in detail
Sales growth from the previous quarter, filed figures.
Earnings per share $0.61 against $0.42 expected; $1.23 as filed, the gap being what the company adds back.
The evidence under this - 4 claims
- 2026-10-23editedACM's own notice of 19 October sets the third-quarter results for 5 November, not the 4 November a vendor estimated; the refresh date moves with it
Management said the cleaning decline reflects a product-cycle transition, not lost share: SPM (about a third of the cleaning market by ACM's own estimate) had 'very little revenue today' but a handful of tools shipped in H1 2026 with more than 20 guided for H2, so the segment total is depressed by a new-product ramp rather than only by losses at existing accounts.
The cash consumption is working-capital growth (inventory and receivables) behind a growing, mostly-GAAP-profitable business, not an operating loss, and the balance sheet holds $1.0 billion of net cash against it (C34) - a multi-year buffer at the current burn rate.
What management did with the money
It has been raising money, not returning it. In May it sold 2.9 million new shares at $52.00 for $149.85M, and in February it sold part of its stake in the Chinese subsidiary for about $111M. That subsidiary also raises its own outside money, which shrank ACM Research's share of it to 74.6%.
Insiders in 2026Q3: 10 purchases and 20 sales, from their Form 4 filings.
The evidence under this - 4 claims
74.6% is still a clear controlling, consolidating majority stake, and ACM Shanghai paid a $36.8 million dividend up to ACM Research and minority holders in the same September 2025 window (C33), so capital has flowed back as well as out.
ACM Shanghai also paid a $36.8 million dividend up to ACM Research and NCI holders in September 2025 (C33), so capital has flowed back as well as out, and 74.6% remains a clear majority stake with consolidation and control intact.
The price, and where we disagree
In short. Fair: cheap against its rivals, level with its own history. The discount pays for China, the US blacklist and cash that has not arrived; it narrows only if the new product lines sell.
The company has raised its full-year forecast twice this year and says first-half orders more than doubled. The September-quarter results on 5 November test whether that reaches the parts the price depends on: cleaning growing again, and a first production order for PECVD or Track, both still on trial.
The evidence under this - 4 claims
- 2026-10-23editedACM's own notice of 19 October sets the third-quarter results for 5 November, not the 4 November a vendor estimated; the refresh date moves with it
Order intake is not independently verifiable and is a figure management chooses to disclose selectively. ACM Shanghai's own backlog figure is now published (C43) and is up 88.2% on a year earlier, which points the same way, but it is a mix of shipped-not-yet-recognised and future-shipment orders on Chinese accounting, so it does not confirm the 105% order-intake figure itself.
- 2026-10-01editedThe backlog disclosure this claim said was pending was published on 2026-09-30; the counter-evidence now points to it (C43) instead of saying it was awaited.
- 2026-10-23editedACM's own notice of 19 October sets the third-quarter results for 5 November, not the 4 November a vendor estimated; the refresh date moves with it
Management said the cleaning decline reflects a product-cycle transition, not lost share: SPM (about a third of the cleaning market by ACM's own estimate) had 'very little revenue today' but a handful of tools shipped in H1 2026 with more than 20 guided for H2, so the segment total is depressed by a new-product ramp rather than only by losses at existing accounts.
A related new platform, horizontal panel-level plating, already booked its first production order in Q2 2026 (C35), which management points to as evidence the evaluation-to-production path is real rather than a standing promise.
- 2026-10-23editedACM's own notice of 19 October sets the third-quarter results for 5 November, not the 4 November a vendor estimated; the refresh date moves with it
How we read the multiple
Buyers pay about 19 times last year's operating profit for the whole company, debt included. Lam and Applied cost about 31 times. That makes sense only if sales keep growing and the gap to those two stays wide.
ACM grows second fastest of the three, is no longer dear against its own past, and earns the lowest return on the money invested in it.
Against its own history. About level with its own recent years after the shares fell by a third since August. The business is bigger and faster growing, but cash has not followed profit.
The evidence under this - 3 claims
A skeptic could read the same discount as the market correctly pricing structurally lower returns on capital (ROIC 4.4% vs a 23.5%-40.8% peer range, C25) rather than as a temporary gap that closes as growth continues.
- 2026-10-23editedre-struck at the 22 October close: 29.1x EV/EBIT to 19.1x, with the cohort median 42.1x to 30.65x; the reading moves from 'a narrowing discount' to a wider one, with the business evidence unchanged
- 2026-10-23editedre-struck at the 22 October close (49.12) from the 31 August close (72.82): P/E 32.25x to 22.47x and EV/EBITDA 14.83x to 9.92x, so ACM no longer trades above both of its own medians
- 2026-10-23editedre-struck at the 22 October close from 31 August: ACM's EV/EBIT 29.1x to 19.1x against a peer median of 42.1x to 30.65x; Lam's refreshed growth (33.4%) now exceeds ACM's 27.3%, so ACM is second fastest, not fastest
Priced against its rivals
The competitors its own annual report names, each confirmed against the rival's filing. Measured on company value for each $1 of operating profit, cheapest first:
| Ticker | Company | company value for each $1 of operating profit |
|---|---|---|
| ACMR | ACM Research | 19.1× |
| LRCX | Lam Research | 29.4× |
| AMAT | Applied Materials | 31.9× |
The middle of the other 2 is 30.6. ACM Research, Inc. is cheapest of 3.
Rivals' figures come from a data provider, so read the spread of the group rather than any single pair.
Cheaper than both listed rivals its own filing names, on every measure. The position is the company's own, and the risks below are what it pays for.
Private or foreign rivals that cannot be priced: Tokyo Electron Ltd., SCREEN Holdings Co., Ltd., NAURA Technology Group Co., Ltd., SEMES Co. Ltd., Kokusai Semiconductor Equipment Corporation, Suzhou Jingtuo Semiconductor Technology Co., Ltd.
The bet: what it pays if it works
This bet is on one business only: PECVD and Track, two new kinds of chip-making tool that customers are still testing. Everything else ACM sells is counted separately, at the company's own price for its sales, and if it meets this year's guidance it covers today's price on its own.
So today's price asks nothing of the new tools. But even a first foothold in their market adds little to the company's value, which makes it an expensive bet; if neither tool ever wins an order, little is lost.
| PECVD and Track platforms revenue in two to three years | Points of that market needed | At the stated multiple |
|---|---|---|
| $83.00M (the big win) | about 1 at 100% kept | 1.8× |
| $360.00M (market forecast) | about 4 at 100% kept | 2.2× |
| $415.00M | about 5 at 100% kept | 2.3× |
SEMI's path carries PECVD and Track near $12 billion by 2029, about 30% in China; a 10% share of China, below Piotech's in PECVD, is about $360 million of revenue. On that forecast, $360.00M pays 2.2× from today's value: a Fair bet on the market case. PECVD and Track platforms would be worth about $1.40B, and the price pays nothing for it.
To pay 2× (a Fair bet) it would need about $197.26M of PECVD and Track platforms revenue (about 2 points of that market); to pay 5× (a Cheap bet), about $2.22B of PECVD and Track platforms revenue (about 27 points of that market), at the same multiple.
Each row is PECVD and Track platforms alone, plus the rest of the company (Cleaning, ECP plating, furnace and advanced-packaging tools) counted at about $4.49B.
The steps it has to take
- Done PECVD and Track evaluation tools placed with customers
- Slipped Production qualification of PECVD or Track due 2026-12
- Not started A first disclosed production order for PECVD or Track due 2026-12
- Not started Repeat production orders from the qualifying customer, the path most of ACM's tools follow due 2027-12
- Not started PECVD and Track revenue near one point of their combined market, about $83 million a year due 2028-12
The detail behind the label
The evidence under this - 4 claims
A related new platform, horizontal panel-level plating, already booked its first production order in Q2 2026 (C35), which management points to as evidence the evaluation-to-production path is real rather than a standing promise.
- 2026-10-23editedACM's own notice of 19 October sets the third-quarter results for 5 November, not the 4 November a vendor estimated; the refresh date moves with it
The by-product-line SAM figures are described as based on 'third-party reports and customer information' without naming which third parties beyond the general Frost & Sullivan citation used elsewhere, so the individual product-line splits cannot be independently checked against a named external source.
- 2026-10-23editedACM's own notice of 19 October sets the third-quarter results for 5 November, not the 4 November a vendor estimated; the refresh date moves with it
- 2026-10-23editedre-struck at the 22 October close from 31 August: ACM's EV/EBIT 29.1x to 19.1x against a peer median of 42.1x to 30.65x; Lam's refreshed growth (33.4%) now exceeds ACM's 27.3%, so ACM is second fastest, not fastest
Where we disagree with the market
A Chinese equipment winner riding AI and the push for home-grown tools, worth buying on the dip: 8 of 10 analysts say buy.
The growth is real, but it sits in unproven tool lines and in the one market forecast to shrink.
The market sees two raised forecasts and earnings beats in 6 of the last 8 quarters. But June's reported beat was a paper gain, and the cash statement and the shrinking stake in the Chinese arm sit far from the headline.
If the new lines win production orders, cleaning grows again and cash turns positive, management's confidence was earned and we were too cautious.
What analysts say, and the evidence
The evidence under this - 8 claims
- 2026-10-23editedACM's own notice of 19 October sets the third-quarter results for 5 November, not the 4 November a vendor estimated; the refresh date moves with it
The cash consumption is working-capital growth (inventory and receivables) behind a growing, mostly-GAAP-profitable business, not an operating loss, and the balance sheet holds $1.0 billion of net cash against it (C34) - a multi-year buffer at the current burn rate.
74.6% is still a clear controlling, consolidating majority stake, and ACM Shanghai paid a $36.8 million dividend up to ACM Research and minority holders in the same September 2025 window (C33), so capital has flowed back as well as out.
A related new platform, horizontal panel-level plating, already booked its first production order in Q2 2026 (C35), which management points to as evidence the evaluation-to-production path is real rather than a standing promise.
- 2026-10-23editedACM's own notice of 19 October sets the third-quarter results for 5 November, not the 4 November a vendor estimated; the refresh date moves with it
Management said the cleaning decline reflects a product-cycle transition, not lost share: SPM (about a third of the cleaning market by ACM's own estimate) had 'very little revenue today' but a handful of tools shipped in H1 2026 with more than 20 guided for H2, so the segment total is depressed by a new-product ramp rather than only by losses at existing accounts.
What could go wrong, and how we would know
The risks
- Both operating arms blacklisted. Since December 2024 nobody anywhere may supply ACM Shanghai or ACM Korea with controlled American technology without a US licence.
- One country, forecast to shrink. Substantially all sales have gone to mainland China, where equipment spending is forecast to fall 9.9% in 2026.
- A few big customers. Four customers were 52% of 2025 sales and 62% of what customers owed at the year end.
- Profit that is not cash. Operations consumed cash in five of the last six quarters, as unsold inventory and unpaid customer bills piled up.
- Insiders selling. Insiders made 69 open-market sales and no purchases over the last year.
The evidence under this - 6 claims
A filed ACM Shanghai investor-relations record states all non-standard components are now fully localized and that progress has been made on localizing certain critical standard components (C42) - more concrete, filed detail than the general assurance management gave on the Q2 2026 call (C29).
The cash consumption is working-capital growth (inventory and receivables) behind a growing, mostly-GAAP-profitable business, not an operating loss, and the balance sheet holds $1.0 billion of net cash against it (C34) - a multi-year buffer at the current burn rate.
The filings classify these sales as tied to standing compensation plans (officer 'See Remarks' filings), not as a single discretionary decision to exit; the pattern is still a real absence of open-market buying at any point in the trailing year.
What we still do not know
- How far a planned Hong Kong listing of the Chinese subsidiary would shrink ACM Research's 74.6% stake.
- Whether swapping blacklisted American parts for local ones works, beyond the company saying so.
- How much more outside money the Chinese arm will raise, and at whose expense.
The evidence under this - 4 claims
74.6% is still a clear controlling, consolidating majority stake, and ACM Shanghai paid a $36.8 million dividend up to ACM Research and minority holders in the same September 2025 window (C33), so capital has flowed back as well as out.
Management said on the Q2 2026 call it 'cannot comment too much on the Hong Kong listing at this time' beyond a previously announced timeline, and the plan still requires shareholder approval, so timing and even completion are not certain.
- 2026-10-23editedACM's own notice of 19 October sets the third-quarter results for 5 November, not the 4 November a vendor estimated; the refresh date moves with it
ACM Shanghai also paid a $36.8 million dividend up to ACM Research and NCI holders in September 2025 (C33), so capital has flowed back as well as out, and 74.6% remains a clear majority stake with consolidation and control intact.
What would change our view
- Neutral→AttractiveThe ramps convert and the profit becomes cash
It needs all of:
- ○PECVD and Track win disclosed production orders by the end of 2026, and cleaning sales grow again after the second-half ramp.watched as it happens
- ○Operations bring in cash that matches the reported profit, instead of running negative beside it.checked at the next results, due by 5 Nov
- Neutral→UnattractiveThe China risks land
It needs any of:
- ○A further American export restriction limits what the Chinese subsidiary can buy or ship.watched as it happens
- ○The Chinese subsidiary lists in Hong Kong, further shrinking ACM Research's share of its own operating business.watched as it happens
We would also be wrong if
- Operations keep consuming cash for two more quarters despite the reported profit.
What is coming
- 5 November 2026 Q3 2026 results: Whether cleaning revenue rebounds sequentially, PECVD or Track book a production order, and the raised 2026 guidance and 42-48% gross margin hold.
Every dated event ahead
- 5 November 2026 Q3 2026 results: Whether cleaning revenue rebounds sequentially, PECVD or Track book a production order, and the raised 2026 guidance and 42-48% gross margin hold.
- December 2026 PECVD and Track production qualification: Whether either platform leaves customer evaluation by year-end as guided; another slip would test the execution record and the potential axis.
- February 2027 Q4 and FY2026 results and the FY2026 10-K: Full-year revenue against the $1.125-1.175B guide, cash flow after the inventory build, the ACM Shanghai stake, and whether high-temperature SPM passed 20 systems.
- March 2027 First volume-production panel-level plating system delivered: Whether the panel-level plating line turns from one order into a shipped production tool, the first test of the advanced-packaging prize outside cleaning.
Since we wrote this
- 28 OctNewsACMR pre-announces Q3 revenue up 34-36%, with shipments growing slower than revenue. View heldACM Research's preliminary third-quarter revenue is $360-365 million, up 34-36%; shipments $335-340 million.
- 23 OctPrice moveShares fell a third; ACM now level with its own historyValuation claims were re-struck at the 22 October close. The discount to Lam and Applied widened as the whole group fell, the business evidence did not move, and the view held.
- 5 OctThesis reviewThe bet now names what it ridesThe owner asked that every bet name the macro trend behind it, after BWXT's bet was missing its chief executive's case for nuclear over gas.
- 4 OctThesis reviewOwner's checklist: revenue streams, cost of growth, management, catalystsThe owner's checklist found the method silent on who runs the company, what growth costs, where the money comes from and what is dated ahead; this pass fills those four blocks from the proxy and filings.
- 1 OctNewsShanghai backlog up 88.2%; view heldThe filed backlog of RMB 17.1 billion agrees with the order growth management cited. It does not say whether cleaning revenue rebounds or new platforms win orders, so the November print stays the test.
- 31 AugThesis writtenNeutral, medium convictionWritten after June-quarter results raised the full-year forecast while the core cleaning line shrank 14% and operations used cash again.
Also checked 2 times with nothing new to act on.
Sources 22
Sources - 22
Every claim above cites one of these, or states the reasoning that derived it. The quote is frozen at the date it was read: a link can rot, and an agent fetch returns a reading of a page rather than the page. Backs counts the whole claim ledger, so it can name a claim this page does not surface.
- S01primarycompany filing2026-03-02business description, competition, risk factors (export controls, customer concentration), ownership structure, dual-class structure, cash-flow drivers, customer/AR concentrationbacksC01C02C04C07C08C09C10C11C14C16C18C20C21C24C25C30C31C33C54C55
- S02primarycompany filing2026-08-07Q2 2026 GAAP/non-GAAP financials and reconciliation, guidance, cash and debt positionbacksC03C05C06C24C34C36
- S03primarytranscript2026-08-07Q2 2026 earnings call transcript (prepared remarks and Q&A)management commentary and figures with no filed home (SAM, long-term target, order growth, product-line color); analyst Q&A on China exposure, component supply, Hong Kong listing
For the first half of 2026, orders increased 105% year-over-year. This is a mix across all product categories with a heavier emphasis on some of our new products.
backsC05C12C13C18C22C23C29C34C35C45C50 - S04primarycompany filing2026-05-12terms of the May 2026 $150 million equity raisebacksC15
- S05primaryvendor data2026-08-31Research pack - market and vendor-fundamentals tiers (FMP, IBKR, Form 4 filings)price/valuation history, peer_valuation cohort multiples, earnings history and EPS-beat record, insider transactions, analyst consensusbacksC02C07C24C25C27C28C36
- S06primarygovernment data2026-08-14FINRA consolidated short interestshort interest percent of float and days-to-cover, settled 2026-08-14backsC26
- S07primarycompany filing2026-02-26original FY2026 guidance ($1.08-1.175bn, 21%-30% growth) later raised in Q2; the February 6, 2026 sale of about 1.3% of ACM Research's ACM Shanghai stake for ~$111 million; year-end 2025 net cashbacksC38
- S08primarycompany filing2026-08-07outside confirmation that 'clean' is a named Lam market, and that installed-base/repeat-order dynamics are a named competitive mechanism industry-widebacksC17
- S09primarycompany filing2025-12-12Semiconductor Systems segment description (deposition as one of several named process steps) used for the competitor-cohort perimeter judgementread, but no claim rests on it
- S10primaryindustry data2025-12-16SEMI, 'Global Semiconductor Equipment Sales Projected to Reach a Record of $156 Billion in 2027' (press release) SEMIan independent, non-company-selected global WFE and mainland-China spending forecast to check against the Gartner figures ACM's own 10-K cites
WFE segment sales are projected to expand 9.0% in 2026 and 7.3% in 2027 ... China is projected to maintain the leading position over the forecast period ... sales gradually decline from 2026
backsC19 - S11primarycompany filing2026-05-29terms of ACM Shanghai's proposed Hong Kong H-share listing: up to 7% of post-offering share capital plus a 15% over-allotment option, subject to shareholder approval, and the intended use of proceedsbacksC37
- S12secondarynews2026-04-20DigiTimes, 'Naura revenue climbs on China chip tool demand, margins tighten amid rising costs and competition'an outside, non-ACM account of NAURA's growth and the domestic-substitution dynamic named as a threat in machine.threat
China semiconductor equipment maker Naura Technology delivered strong 2025 revenue growth on rising chip tool demand and domestic substitution, but weaker margins and intensifying competition signal a tougher next phase.
backsC39 - S13primarycompany filing2026-06-22a filed (STAR Market Reg-FD) Q&A record with a different $4bn target split than the Q2 call, component-localization detail, Q1 2026 order growth (65% YoY, a filed figure distinct from the 105% H1 figure), and Lingang Plant A/B capacity detailbacksC41C42
- S14primarycompany filing2026-09-30ACM Shanghai total backlog RMB 17.1bn (USD 2.5bn) at 2026-09-29, +88.2% year on year, with the company's own caveats on the measure; and the dates of the nine-month results (28 October) and Q3 call (early November)backsC43
- S15primaryindustry data2026-04-01the industry body's fab-equipment spending path to 2029, the growth rate the PECVD and Track market is carried forward on; read through ir_fetch
Worldwide 300mm fab equipment spending is expected to increase 18% to $133 billion in 2026 and 14% to $151 billion in 2027, SEMI reported today in its latest 300mm Fab Outlook. ... Looking further out, the report projects investment will continue to increase 3% to $155 billion in 2028 and another 11% to $172 billion in 2029, respectively.
backsC44 - S16primarytranscript2026-04-30Tokyo Electron, FY2026 fourth-quarter earnings presentation and transcript Tokyo Electron (8035.T)the Track incumbent's own account: 91% world coater/developer share in CY2025 (Gartner), coater/developer 28% of FY2026 new-equipment sales of 1,775.4 billion yen, China 34.1% of FY2026 sales and 26.8% in Q4, a CY2026-27 WFE outlook of $150-170 billion a year; read through the ir_fetch client and pdftotext
In the coater/developer business, in particular, our share in the global market exceeds 90%. ... On the full-year basis in the fiscal year ended March 2026, proportion of sales in China was 34.1%.
backsC46 - S17primarycompany filing2026-04-28Piotech Inc. (688072.SS) 2025 Annual Report, Shanghai Stock Exchange filing via CNINFO Piotech (688072.SS)China's domestic PECVD leader's 2025 PECVD revenue and its SEMI-based sizing of thin-film and PECVD markets, globally and in mainland China; read through the ir_fetch client and pdftotext
公司 PECVD 系列设备实现销售收入约 51.42 亿元，同比增长 75.27% ... 根据 SEMI 历史统计，PECVD 是薄膜设备中占比最高的设备类型，约占整体薄膜沉积设备市场的 33%
backsC47 - S18primarycompany filing2026-04-18Shenyang Kingsemi (688037.SS) 2025 Annual Report, Shanghai Stock Exchange filing Kingsemi (688037.SS)China's domestic Track leader's 2025 revenue (RMB 1.948 billion, +11.11%) and its statement that front-end Track localisation in China remains low; read through the ir_fetch client and pdftotext
全球前道涂胶显影设备市场长期被海外厂商高度垄断，目前国产化率仍然较低。
backsC48 - S19primarycompany filing2026-04-27beneficial ownership and voting power as of April 13, 2026, the executive officers and their start dates, the CD&A's discretionary bonus and stock-option design, the 2025 bonusesbacksC51C52C53
- S20primarycompany filing2023-03-02the FY2021 and FY2022 income statement and capital spending, read through the frozen numbers card (content/ACMR/financials.json)backsC54C55
- S21primarycompany filing2026-08-07the first-half 2026 cost lines behind the trailing year, the financing cash flows (ACM Shanghai stake sale, the May 2026 share issue) and the June 2026 Shanghai office purchase and its Bank of China loanbacksC54C55
- S22primarycompany filing2026-08-27the delivery dates for the first volume-production panel-level plating system and the panel evaluation system, the high-temperature SPM shipment target for 2026, and the September 30 backlog disclosurebacksC56