Kerrisdale Capital's Sahm Adrangi on $ACMR's price dislocation between Shanghai and NASDAQ listings
Summary
Sahm Adrangi’s core call is that $ACMR offers a Chinese wafer-fabrication-equipment growth company at a 60%-75% discount created by a mismatch between its U.S. listing and China-centered operations. At roughly $23, ACMR sat below a DCF near $70, comparable-company values of $60-$110, and the approximately $73 per-share value implied by its 82% stake in Shanghai-listed ACM Shanghai. “It’s just way below any valuation methodology you come up with.”
China’s localization mandate gives ACMR a protected route into an industry whose qualification costs normally entrench a handful of suppliers. Revenue grew from $30 million in 2016 to 2025 guidance of $850-$950 million because Chinese fabs are “mandated to go with a Chinese toolmaker if it’s remotely competitive.” Adrangi’s second leg is less certain: incubation at home might eventually let ACMR become a second or third source for international fabs.
Adrangi argues that tougher U.S. semiconductor restrictions are operationally bullish for ACMR even when the headlines hurt its stock. Its Entity List addition helped drive shares from roughly $19 to $15, yet January guidance was little changed from pre-list expectations; meanwhile, pressure on China increases Beijing’s incentive to support domestic suppliers. “The tougher the U.S. is on China…the more China is going to double down.”
Fraud and governance risks are central, but Kerrisdale’s conviction comes from channel checks rather than trusting the reported numbers alone. Those checks described ACMR as China’s leading cleaning-equipment supplier, credited its ECP tools, patents, and engineering team, and suggested its products are genuinely competitive. Even a more copycat-heavy reality could still support the thesis if fabs keep installing the tools and localization continues.
Walker identified management’s lack of urgency as a key challenge to the thesis. The valuation gap has several possible unlocks: Adrangi prefers a Hong Kong dual listing, a premium take-private and Asian relisting, or a strategic sale over shrinking a relatively illiquid U.S. float; he is reluctant to demand that a 30-year founder sell 20%-25% of his enterprise for a quick shareholder gain. Walker’s pushback was blunt: ACMR could monetize part of ACM Shanghai and repurchase the parent at an enormous discount, and “that does sound pretty nice.”
A slowing semiconductor-equipment cycle may be absorbed by localization-driven share gains. Adrangi cited forecasts for Chinese fabs’ domestic-tool penetration to rise from the mid-teens to the low 30s within roughly two years, potentially offsetting a 5%-10% decline—and perhaps even a 10%-20% decline—in overall spending. ACMR also has a land-and-expand opportunity from cleaning and electrochemical plating into advanced packaging, furnaces, PECVD, and track tools.
The ugliest financial signal is working capital, including inventory days above 500, but Adrangi offered a plausible business-model explanation rather than a clean exoneration. First tools require customization and evaluation periods of up to 24 months, with payment and revenue recognition delayed, while repeat tools are recognized on shipment and delivery. Adrangi expects conversion to improve as repeat tools rise; his best answer was that if the tools are competitive and entering fabs, “one way or another, they’re ultimately going to get paid.”
The risk Adrangi identifies most is an extreme U.S. ownership restriction, although he regards it as a remote tail risk that might paradoxically force the valuation unlock. A 365-day divestment window could accelerate a Hong Kong listing and share conversion; he thinks the present 68% jurisdictional discount might then narrow to 10%-25% relatively quickly. He acknowledges that restrictions could also strand U.S. holders from accessing value in the Chinese subsidiary.
Deep dive
1. China gives ACMR a route through an otherwise closed oligopoly
Adrangi framed wafer-fabrication equipment as a collection of “little oligopolies.” Applied Materials, Lam Research, KLA, ASML, and major Japanese suppliers sell the tools used inside semiconductor fabs; he considers the larger platform businesses companies one could plausibly own for 20 or 30 years because their competitive advantages are so durable.
The moat comes from joint development and qualification. Fabs pursuing more advanced chips work closely with incumbent toolmakers, while replacing a qualified tool imposes testing costs, operational disruption, staff retraining, and the inefficiency of supporting multiple systems for the same process step.
Walker’s pushback—worth keeping—was that ACMR’s own rise appears to prove these oligopolies can be breached. Adrangi’s answer: it could break in because Chinese fabs give domestic tools preferential treatment, whereas Intel or SK hynix must voluntarily accept the cost and risk of replacing an incumbent.
That distinction explains ACMR’s climb from $30 million of revenue in 2016 to 2025 guidance of $850-$950 million. China is effectively incubating domestic platforms; once their tools become competitive, they might pursue the harder second leg of becoming international second or third sources.
2. Every valuation route lands far above the U.S. share price
Asked what the market had missed for five years, Adrangi offered an “honest non-answer”: he does not usually try to inhabit the market’s psychology. He estimates future cash flows, discounts them, and buys when the quoted value is sufficiently below the result.
Here the explanation nevertheless looks straightforward: ACMR is a U.S.-listed company operating in a Chinese semiconductor sector that Washington is actively trying to constrain. Adrangi sees “the mismatch between the shareholder base and the listing and the underlying company” as the only discount explanation that really resonates.
The numerical gap was unusually wide. Kerrisdale’s DCF produced roughly $70 per share; peer analyses across U.S., Japanese, and Chinese equipment companies produced $60-$110; discounting a 2028 EPS estimate at a 30x P/E yielded about $63.
Most strikingly, ACMR’s 82% holding in Shanghai-listed ACM Shanghai implied approximately $73 per ACMR share when the U.S. parent traded near $23. Walker summarized the special situation as a roughly $1 billion parent controlling an operating company valued near $6 billion: “You can drive a truck through that valuation.”
3. U.S. restrictions may strengthen the operating thesis
Kerrisdale wants to reverse the market’s geopolitical reflex. ACMR fell from about $19 to $15 after an Entity List addition, but Adrangi argued that “all of these headlines…are actually good for ACMR” because restrictions accelerate China’s substitution of foreign semiconductor tools.
He did not claim the rules are harmless. Rather, the observed workarounds and China’s continued progress suggest they have not achieved their stated purpose, while ACMR’s January guidance remained broadly consistent with expectations formed before the listing.
Walker initially viewed ACMR’s exposure to the Chinese semiconductor industry as a negative. His change of mind captured the thesis: exposure to that industry becomes attractive when the government is actively encouraging every feasible piece of the supply chain to be sourced locally.
The non-China opportunity remains more speculative. Management historically envisioned a global supplier, and Adrangi heard encouraging indications from prospective overseas customers—he was most optimistic about SK hynix—but acknowledged that qualification outside China has moved slowly.
4. Channel checks carry more weight than reported profits
Walker raised the history of fraudulent U.S.-listed Chinese companies. Adrangi said Kerrisdale’s conviction in this name comes from channel checks, recalling that when the firm exposed Chinese frauds in 2011 and 2012, it checked facilities and spoke with customers.
The checks identified ACMR as China’s number-one cleaning player, described its electrochemical-plating product as innovative, and pointed to novel patents and a strong engineering core.
Some checks also suggested ACMR was “less of a copycat” than other domestic suppliers. Adrangi stressed that the thesis does not require technological purity: even reverse-engineered tools can take substantial Chinese share and eventually qualify internationally outside the most sophisticated process categories.
The load-bearing evidence is that competitive tools are entering real fabs. Profit leakage or agents “skimming some profit off the top” would matter, but less than whether ACMR is building an installed base capable of supporting revenue and attractive steady-state margins five, ten, or 15 years from now.
5. Governance protects holders imperfectly—and management is in no hurry
ACMR directly owns its Chinese operating company rather than relying on a VIE, and the parent is incorporated in Delaware. Adrangi described the CEO as holding voting control through Class B shares but only about 10% of the economics, so a low-priced take-private would expose him to the value gap on the 90% he does not own.
Adrangi thinks an M&A adviser would likely caution against a lowball take-private without a majority-of-the-minority vote, especially given ACM Shanghai’s visible valuation benchmark. A Delaware court could potentially view such a transaction as taking the company private at a significant discount; approval by the minority would make the fair-value argument easier.
Walker highlighted the unresolved challenge: management is aware of the large listing discrepancy yet has shown little urgency to exploit it. Adrangi’s “hope is just it’s a slow process,” a notably softer answer than his confidence in the operating business.
Selling 20%-25% of ACM Shanghai could fund a parent buyback or distribution, but Adrangi resisted telling a founder with a decades-long horizon to sell “a quarter of your business” for his own 12-month return. Walker conceded the horizon mismatch while reminding him: “You are a shareholder, and that does sound pretty nice.”
6. A Hong Kong listing looks cleaner than shrinking the float
Adrangi prefers aligning the shareholder base with the underlying business through a Hong Kong dual listing, premium take-private and relisting, or outright sale. A large buyback could make an already small, roughly $1 billion company less liquid and might take years to close the discount.
A strategic sale of ACM Shanghai to a larger Chinese equipment company could, by his estimate at the time, generate something like a “600% return overnight.” A 30% premium may not excite a founder, but the jurisdictional gap could make surrendering control potentially transformative.
Large capital transfers from China may also require government approval, which could be difficult if the purpose is merely returning cash to U.S. holders while Beijing wants the company reinvesting to become more competitive. That constraint reinforces the case for changing the listing venue rather than extracting the subsidiary’s capital.
Walker floated a more creative route: ACM Shanghai could offer its own shares for ACMR, letting the child acquire the parent at a premium while creating value for remaining Shanghai holders. Adrangi did not validate the regulatory feasibility, but reiterated that any bidder should need to pay an attractive premium.
7. Localization can outrun both a cycle downturn and new competition
Adrangi defended ACM Shanghai’s six-to-eight-times-revenue valuation with revenue growth above 40% for about six years and roughly 20% EBITDA margins, in an industry structure unlike Chinese EVs or solar panels. A fab will not efficiently support six different cleaning tools, limiting the number of viable competitors.
Cleaning is less sophisticated than lithography, but that does not make it a commodity. ACMR’s pole position and installed base still matter because fabs dislike ripping out qualified tools, while China’s enormous localization “white space” lets domestic suppliers pursue different categories without immediately stepping on one another.
Beyond leadership in cleaning and electrochemical plating, ACMR is expanding into advanced packaging, furnace tools, PECVD deposition, and track tools. The model is “land and expand”: secure a fab with one system, then qualify adjacent products.
Chinese WFE spending may slow after foreign-tool pre-buying ahead of restrictions, with cited forecasts around 5%-10%. But domestic tools were not similarly pulled forward, and localization was projected to rise from the mid-teens to the low 30s within two years; Adrangi thinks those share gains could offset even a 10%-20% market contraction.
8. Working capital is explainable, while U.S. policy remains the true tail risk
ACMR’s inventory days above 500 understandably trigger fraud alarms. Adrangi explained that first tools are customized for each logic or memory fab, can undergo evaluation for up to 24 months, and may not produce payment or recognized revenue until that process ends.
Repeat tools receive revenue recognition on shipment and delivery. Because ACMR grew roughly 40% annually from almost nothing, first-tool deployments consumed substantial cash; Adrangi expects—“hopefully”—cash conversion to improve as repeat systems become a larger share.
Channel checks make him comfortable that fabs want the tools, but he preserved the uncertainty: “One way or another, ultimately they’re going to get paid.” He called that his best answer on the working-capital issue.
The risk Adrangi cannot diligence away is extreme U.S. action against ownership. Yet a rule providing 365 days to divest could accelerate a Hong Kong conversion, where he believes the roughly 68% discount could narrow to 10%-25% quickly; thus the feared restriction “actually could be a catalyst for the upside,” though he still calls it an unlikely tail risk.