ACM Research

Phase A — Understand the business

Lens 1 · Company Overview

ACM Research designs and sells front-end and packaging wafer-fabrication equipment to chipmakers: single-wafer wet cleaning (its anchor franchise), Tahoe (low-chemistry sulfuric cleaning), semi-critical cleaning, electro-chemical plating (ECP), furnace/Thermal-ALD, PECVD, Track (coat/develop), and stress-free Cu polishing. Tools run from $0.5M to >$5M each. Management sizes its addressable WFE slice at ~$21B of the 2025 market (cleaning ~$7.3B, PECVD ~$5.3B, Track ~$3.0B, furnace ~$2.6B, ECP ~$1.5B).

The structure is the story. ACM Research, Inc. is a Delaware holdco founded in 1998. It runs "a substantial majority of product development, manufacturing, support and services in mainland China through ACM Shanghai," in which it holds 74.6%. ACM Shanghai is separately listed on Shanghai's STAR Market (688082.SS), with a market cap that has at times exceeded the US parent's. So a US shareholder owns ~75% of a Chinese-listed operating company through a Nasdaq wrapper — and a 22.8% non-controlling interest leaks out of consolidated net income before it reaches ACMR holders.

Contract terms: equipment is sold with a "first-tool" acceptance model — initial units sit at customer sites and revenue is recognized on acceptance, which is why advances from customers ($168.8M at Q1'26) and deferred revenue feature heavily. Demand is order/backlog-driven, not recurring.

Lens 2 · Supply Chain

Upstream → ACM → end customer, named where the filings name them:

• Inputs / subsystems: ACM buys components, modules, and US-origin technology globally; it performs additional subsystem production in Korea via ACM Korea (a subsidiary of ACM Shanghai). Raw materials, wages, and supply-chain payments are largely in RMB and KRW; products are sold mostly in USD — a structural FX mismatch.
• The company: manufacturing concentrated in Shanghai; long-lived assets are $321.7M in Mainland China vs $8.9M Korea / $9.5M US at YE2025 — i.e. the physical company is ~95% in China.
• End customers: Chinese fabs scaling mature-node and advanced capacity. SemiAnalysis and Kerrisdale public work identify ACM tools winning at SK Hynix and Intel historically, but the revenue base is now ~entirely China (Lens 4).
• The chokepoint that defines the company: ACM Shanghai and ACM Korea are on the BIS Entity List (Dec 2, 2024). Any party worldwide is broadly prohibited from furnishing US-export-controlled hardware/software/technology to those subsidiaries without a license. The single-source dependency is not a vendor — it is continued access to US-origin components and tools under license. This lens does not stay generic: the named chokepoint is US export-control jurisdiction over ACM's own Chinese and Korean legs.

Lens 3 · Competitive Advantages (moats)

• Process IP / founder-invented core tech. Founder-CEO David Wang invented stress-free Cu polishing and holds 100+ patents. The SAPS/TEBO megasonic cleaning IP is a genuine technical differentiator in wet cleaning.
• The real moat is geopolitical, not technological: localization. ACM Shanghai is the leading domestic Chinese supplier of wafer-cleaning tools at a moment when Beijing is forcing supply-chain self-sufficiency. Against Lam (cleaning) and DNS/SCREEN, ACM's edge inside China is that it is the local champion the policy favors. That same fact is the bear case (Lens 13) — a moat granted by politics can be revised by politics, on both sides of the Pacific.
• Bargaining power: moderate-to-weak. Customers are concentrated (top-4 = 52% of revenue, Lens 4); ACM needs them more than they need any single tool line, except where it is the only licensed/local option. Upstream, ACM is a price-taker on US-origin componentry it may not freely import.
• Switching costs: real once a tool is qualified into a process flow (re-qual is expensive), which supports the installed-base/services tail (8.4% of revenue).

Lens 4 · Segments

By product category, FY revenue ($000s):

By geography: Mainland China $897.98M of $901.31M = 99.6% in FY2025; "Other Regions" collapsed to $3.3M (from $16.8M in 2023). This is not a global equipment maker — it is a China-domestic capex play.

The trend that matters (decelerating core, mix-shifting growth): the anchor cleaning franchise grew only +8.1% in FY25 (after +43% in FY24) and then went negative −5.5% in Q1 2026 ($122.5M vs $129.6M). Total Q1'26 revenue still rose +34.2% to $231.3M — but the growth came entirely from ECP/furnace/other +204.9% ($84.2M vs $27.6M) and packaging/services +62%. The headline number is healthy; the engine underneath it is rotating away from the highest-quality, highest-margin product line.

---

Phase B — Measure performance

Lens 5 · Earnings Result

FY2025 (10-K):

• Revenue $901.3M, +15.2% YoY.
• GAAP gross profit $400.1M, gross margin 44.4% — down 570 bps from 50.1% in FY24. Cause per management: "revenue mix between product categories, and a higher provision for inventory". An inventory provision is a soft-demand tell.
• GAAP operating income $109.4M (12.1% margin), down 27.5% from $151.0M (19.3%) in FY24 — operating income fell while revenue rose, the single most important fact on the P&L.
• Net income attributable to ACMR $94.1M (10.4% margin) vs $103.6M (13.2%) — −9.2%.
• Diluted EPS $1.37 vs $1.53 — −10.5%.
• Cash flow is the flag: operating cash flow was −$10.3M (vs +$152.5M in FY24) and free cash flow −$67.1M. A company growing reported revenue 15% while burning operating cash is funding its growth through the balance sheet (receivables + inventory).

Q1 2026 (10-Q):

• Revenue $231.3M, +34.2% YoY; shipments $240.7M, +53.6%.
• GAAP gross margin 46.4% (vs 47.9%); non-GAAP 46.5% vs 48.2% — recovering off a "low-40s" Q3/Q4 2025 trough but still down YoY.
• Net income attributable margin fell to 7.6% (from 11.8%); GAAP diluted EPS $0.24; non-GAAP EPS $0.34 vs $0.28 expected — a beat.
• Balance-sheet flags worsening: accounts receivable $561.6M (~219 days of sales), inventory $738.0M (>3 quarters of COGS), allowance for credit losses $35.1M, up from $18.3M a year ago. Working capital is absorbing the growth.
• Market reaction: stock +14.5% on the print despite the margin compression — the tape rewarded the revenue beat and shrugged off the cash-flow/margin story.

Lens 6 · Earnings Calls (sentiment trend)

No transcripts/ on the shelf; sentiment is. Across the last several calls the arc is consistently upbeat on demand, increasingly defensive on margin:

• Management frames Q1'26 GM as "above the midpoint of the 42–48% long-term target and a good recovery from the low-40s in Q3/Q4 2025," explicitly attributing the dip to inventory provisions and mix.
• Recurring phrases: "new product ramps" (SPM single-wafer, Tahoe, vertical furnace/Thermal-ALD), "backlog visibility," "global expansion ambitions".
• The thing they now say that they downplayed before: "international trade policy uncertainties" named as an explicit outlook factor. The Entity-List reality has moved from footnote to stated risk in management's own framing — tone has shifted from pure-growth to growth-with-an-asterisk.

Lens 7 · Comps

Peer table — global WFE peers + the two domestic-China comparables. Multiples are, dated; where a field isn't cleanly sourced it is marked n/a. Do not treat the China A-share multiples as directly comparable (different float, investor base, accounting).

Read: ACMR's EBITDA-based multiple (44.7x) sits below LRCX (65x) and its fwd P/E (61x) is between LRCX and the Chinese names. But ACMR's ROE is a fraction of LRCX/KLA's (single-digit-ish vs 60–80%) because of margin compression and the NCI leak, and its EV/Sales of ~6.5x is being paid for a 99.6%-China, Entity-Listed, negative-FCF business. On quality-adjusted terms it is the most expensive name in the table, not the cheapest.

Lens 8 · Stock-Price Catalysts (last ~5 years)

What actually moves ACMR:

• Earnings beats/raises — repeatedly the dominant driver. Q1'26 +14.5% on a revenue beat; the stock has historically gapped on quarterly prints and guidance raises.
• Analyst target hikes — Roth Capital raised PT $100→$125 in June 2026; the stock hit an ATH ~$110.18 on 2026-06-18. Sell-side re-rating has amplified the move.
• China-policy / export-control headlines — the Dec 2, 2024 Entity-List addition is the standing overhang; trade-policy news is a recurring shock source.
• The ACM Shanghai (688082.SS) STAR listing — the A-share's valuation and any further capital raises there move the parent. The pattern: the market trades ACMR as a high-beta China-semicap momentum name — it reacts to revenue/backlog and sell-side targets far more than to margin or cash-flow quality. That is precisely the setup that breaks on a quarter where revenue also disappoints.

---

Phase C — Judge people & books

Lens 9 · Management

• Founder-CEO Dr. David Wang — CEO/President/director since founding in 1998. PhD/MEng (Precision Engineering, Osaka), BS (Tsinghua); inventor of stress-free Cu polishing; 100+ patents. This is a genuine technical founder, not a hired operator — the right archetype for a still-scaling equipment company.
• Skin in the game: Wang is a >10% owner (director/officer/10%-holder), ~802,708 Class A shares direct plus indirect family-trust holdings, and ACMR carries a dual-class structure giving founder/insiders voting control. Alignment is high but control is concentrated.
• Capital allocation: reinvestment-heavy (R&D 16.1% of revenue FY25, up from 13.5%) — appropriate for the land-grab, but ROE/ROIC is deteriorating on his watch as margins compress and the balance sheet bloats. No dividend; minimal buyback.
• Red flag to weigh, not to over-read: Wang has been a persistent seller via a Rule 10b5-1 plan adopted Nov 29, 2024 — three weeks before the Entity-List designation went effective. The plan's timing (just ahead of a known adverse catalyst window) deserves attention even though pre-planned 10b5-1 sales by a long-tenured founder trimming a large position are routine. Not alleging anything; flagging the optics.
• Founder vs professional: clearly founder-led; implies vision/continuity strength and key-person + governance-concentration risk.

Lens 10 · Forensic Red Flags

Act as a forensic analyst. Where cash and earnings diverge, look hard.

• Earnings-to-cash divergence is the headline. FY25 GAAP net income $121.9M but operating cash flow −$10.3M. The gap is working capital: receivables and inventory are growing faster than revenue.
• Receivables: AR (net) $526.5M at Q1'26 against ~$231M quarterly revenue → ~219 DSO. Top-4 customers = 60–62% of AR. Allowance for credit losses nearly doubled YoY to $35.1M — management is itself signaling collection risk in a concentrated, China-domestic book.
• Inventory: $738.0M (raw $377.9M / WIP $81.6M / finished $278.4M), >3 quarters of COGS, and the FY25 GM hit was partly a higher inventory provision. Rising finished-goods + provisions = demand/timing risk crystallizing.
• Revenue recognition: "first-tool acceptance" model + large customer advances ($168.8M) is legitimate for capital equipment but is judgment-heavy on timing — a lever that can flatter or defer a quarter. Watch acceptance-timing language.
• SBC and non-GAAP gap: FY25 SBC $33.6M; the company's preferred "adjusted operating income" ($143.0M) is 31% above GAAP operating income ($109.4M). The gap is widening relative to GAAP profit as GAAP margins fall — standard, but the non-GAAP framing increasingly carries the bull narrative.
• Government subsidies: mainland-China R&D funding is a recurring input to reported expense/capex — a Beijing-policy dependency embedded in the cost structure.
• NCI leak: 22.8% of consolidated profit belongs to ACM Shanghai's other (Chinese public) holders, not ACMR shareholders — easy to miss when reading consolidated headlines.

Regulatory findings (required):

• SEC Litigation Releases / AAERs: none found naming ACM Research, 2021–2026.
• Non-SEC enforcement (web): no ACM-specific FTC/DOJ/FDA/CFPB action found. (The 2024–2026 DOJ/SEC short-seller cases surfaced in search relate to Andrew Left / Citron Research, unrelated to ACM.) The material government action against the company is the BIS Entity-List designation of ACM Shanghai and ACM Korea (effective 2024-12-02) and exposure to the Outbound Investment Security Program / COINS Act — regulatory, not enforcement/fraud.
• Item 3 Legal Proceedings (10-K): "From time to time we may become involved in legal proceedings… we currently believe that the final outcome of these ordinary course matters will not have a material adverse effect" — i.e. no material pending litigation disclosed.
• Verdict: No accounting-fraud or enforcement findings of record. The forensic concerns are earnings quality (cash-vs-earnings gap, receivables/inventory build, provisions) and structural/regulatory (Entity List, China subsidies, NCI), not malfeasance.

---

Phase D — Project & stress-test

Lens 11 · Forward Projection

Built bottom-up off FY25 actuals + FY26 guidance ($1,080–1,175M, ~25% midpoint growth, reaffirmed Apr 2026 ). Assumptions labeled; outputs. NCI ~22.8%, tax ~10%, diluted shares ~67.8M.

The valuation read: at ~$99 the stock trades at ~61x base-case FY26 GAAP EPS and ~48x FY27. Even on the bull path ($2.05 FY26) it is ~48x. EV/Sales has compressed to ~5.5x on the forward number but EV/EBITDA ~45x is rich for a business with decelerating core, negative FCF, 99.6% single-country revenue, and both opcos on the Entity List. The multiple embeds flawless execution and benign US-China policy.

No forecast.ts create in unattended watchlist mode (per skill rules). Forecast to log on a future interactive pass: ACMR FY26 GAAP diluted EPS >= $1.55, p≈0.55, resolves 2027-02-28.

Lens 12 · Bull vs Bear

Bull case. China's forced semiconductor self-sufficiency is a multi-year, policy-guaranteed capex super-cycle, and ACM Shanghai is the domestic cleaning champion positioned to take share from Lam/SCREEN inside China while expanding into ECP, furnace/Thermal-ALD, Track and PECVD — multiplying its served WFE from cleaning into a $21B opportunity. New product ramps (SPM, Tahoe, vertical furnace) are early; backlog and shipments (+54% Q1'26) lead revenue; a founder-inventor with 100+ patents runs it. If domestic Chinese demand stays vertical and margins normalize to the 46–48% model, FY27 EPS >$2 and the growth re-rates the stock further. Sell-side agrees (Roth $125, "Strong Buy").

Bear case (permanent-impairment risks). (1) Concentration on one geopolitically contested market — 99.6% China revenue with both operating subsidiaries on the BIS Entity List; a tightening of US export controls on the inputs ACM needs (or Chinese retaliation/localization that favors NAURA/AMEC over a US-parented firm) could structurally impair the business. (2) Margin + cash-flow deterioration — GM −570 bps in FY25, operating income down 27% on higher revenue, negative operating cash flow and FCF, receivables at 219 DSO and a doubling credit-loss allowance: growth that doesn't convert to cash. (3) Valuation — ~61x forward GAAP EPS prices perfection. Pre-mortem (18 months out, thesis broke): a single soft quarter — core cleaning down again, a large Chinese customer stretches payment or an AR write-off lands, and a fresh export-control headline — collapses the momentum multiple from 60x toward 25–30x, a >50% drawdown even with revenue merely flat-to-up. Contrarian view the market is refusing to see: the bulls are paying a global-semicap multiple for what is functionally a single-country, policy-dependent, sanctioned-subsidiary equipment maker whose own cash flows have already turned negative — the re-rating risk is to the downside, not the upside.

Lens 13 · Devil's Advocate (short-seller)

Dismantling the bull case:

• Where revenue is concentrated: 99.6% Mainland China; top-4 customers 52% of revenue, 60%+ of AR. If even one large Chinese fab cuts capex or stretches terms, both the P&L and the receivable book wobble at once.
• Why the moat is weaker than bulls think: ACM's edge inside China is localization favored by policy — but the most localized champions are NAURA and AMEC, Chinese-domiciled firms with no US-parent baggage. Beijing's self-sufficiency drive can just as easily route share to them as to a Nasdaq-listed, Delaware-parented company whose subsidiaries are on the US Entity List. The moat cuts both ways.
• Most dangerous competitor bulls underestimate: not Lam — NAURA (largest Chinese equipment maker, expanding into cleaning and hybrid bonding) and AMEC (etch leader pushing into deposition). Both are scaling product breadth on home turf.
• Worst capital-allocation/governance optics: founder selling via a 10b5-1 plan adopted three weeks before the Entity-List designation; dual-class control; a 22.8% profit leak to NCI; reliance on Chinese government subsidies in the cost base.
• Assumptions that must hold for $99: ~25%+ growth and GM recovery to ~47% and no incremental US export tightening and no Chinese-customer credit event and the momentum multiple persists. Five things, all favorable, simultaneously.
• If growth disappoints 20–30%: at ~$880M FY26 revenue with GM back to low-40s, GAAP EPS could fall toward ~$0.90; a momentum name that misses growth and shows margin/cash stress routinely loses half its multiple — a 50–60% drawdown is the plausible, not the tail, outcome.
• The single scenario that permanently impairs: a US rule extending the Entity-List restrictions to deny ACM Shanghai the US-origin components/tools it needs to build its own products (not just sell into the US), forcing a costly redesign or capacity loss — directly contemplated in ACM's own risk factors. Plausibility: non-trivial and bidirectional given the 2024–2025 policy trajectory (Entity List + OISP + COINS Act).

Lens 14 · Management Questions (ordered by information value)

1. If US export controls tightened to restrict the US-origin components ACM Shanghai imports to build its tools (not just US sales), what is your qualified second-source/redesign plan and the gross-margin cost of executing it?
2. Operating cash flow was −$10.3M in FY25 and −$45M in Q1'26; at what revenue level and working-capital discipline does the business turn FCF-positive, and when?
3. Accounts receivable are at ~219 DSO with the credit-loss allowance up to $35.1M — which customers are stretching, and what is the realistic write-off exposure?
4. Core single-wafer cleaning revenue declined 5.5% YoY in Q1'26 — is this timing, share loss to NAURA/local rivals, or a demand plateau in mature-node cleaning?
5. You reaffirmed 20–30% 2026 growth — what specifically in the backlog and acceptance schedule underwrites the back half, and what would cause a cut?
6. Gross margin fell 570 bps in 2025 on mix and inventory provisions — what is the steady-state product mix that holds the 46–48% model, and how much is structurally lower-margin furnace/ECP?
7. With $738M of inventory (incl. $278M finished goods at customer sites), how much is at risk of further provisioning if acceptances slip?
8. How do you defend domestic share against NAURA and AMEC specifically, given they carry no US-Entity-List constraint?
9. What is the long-term capital-allocation plan for the 22.8% NCI in ACM Shanghai — buy it in, let the STAR float grow, or status quo — and how do you weigh ACMR-holder vs ACM-Shanghai-holder interests?
10. What is your exposure to the OISP/COINS Act outbound-investment regime, and could it constrain capital flows between the US parent and ACM Shanghai?
11. How dependent is the FY cost structure on mainland-China government R&D subsidies, and what happens to margins if they are reduced?
12. The CEO's 10b5-1 plan was adopted just before the Entity-List event — what governance steps ensure insider-sale optics don't undercut investor trust?
13. What is the realistic non-China revenue ambition (it fell to $3.3M / 0.4% in FY25) and the timeline, given the Entity-List handicap abroad?
14. Where are you on PECVD and Track qualification at tier-1 customers, and what revenue do they contribute by FY27?
15. Under what scenario would you consider separating or fully consolidating the US and Shanghai listings to remove the structural complexity discount?

---