Chroma ATE Value Investment Deep-Dive

Hunting ten-year five-baggers among great growth stocks — pressing the upside question: "Can it get much bigger?"

• QHow high is its market ceiling? Is it expanding an existing market, or creating an entirely new one?5/10

  A

  The ceiling is not low, but Chroma is “expanding and taking a larger share of an existing cake,” not creating a new market from scratch. It sells test and measurement and automated test systems, a field that has existed for decades. What Chroma is doing is riding new demand curves in AI data centers, power electronics, and EV/batteries to raise its share and average selling price within the existing test market, rather than opening a category that has never existed before.

  Start with whether the cake is actually growing. The report cites the IEA Global EV Outlook 2026: global EV sales in 2025 exceeded 20 million vehicles, about one-quarter of new car sales, while EV battery deployment reached 1.2 TWh, up nearly 30% year over year. This directly pulls demand for battery, EV, charging/discharging, and power electronics testing. The semiconductor side is also in a real upcycle: Teradyne’s semiconductor test revenue in Q1 2026 exceeded USD 1 billion for the first time in a single quarter and more than doubled year over year, with AI-related demand already accounting for nearly 70% of that quarter’s revenue; Advantest plans in 2026 to expand tester capacity from 3,000 units to 5,000 units. Demand is genuinely strong, and the report says so plainly.

  But a “high ceiling” and “creating a new market” are two different things. Chroma itself is climbing on the back of these existing demand lines: the report discloses that 2025 revenue from power electronics test equipment grew 55% year over year, while semiconductor test revenue grew 40% year over year. Both engines are “existing customers expanding capex,” not a new demand category invented by Chroma. It is more like making the corners of the mature test cake that belong to AI power chains, BBU/PSU, and SLT system-level testing wider, deeper, and higher priced.

  The real boundary of the ceiling is also clear: Chroma is not the absolute leader in cutting-edge semiconductor ATE. The highest-end profit pool still belongs to Advantest and Teradyne. So its serviceable addressable market (SAM) is “the test and measurement segments where it has product-line capability”: power electronics, EV/batteries, server power, parts of semiconductor/system-level testing, photonics, and integrated measurement. This is an existing market with several strong growth curves, but its ceiling is shared with others and rises and falls with downstream capex cycles. It is not a winner-takes-all new empty field.

  In one sentence: the cake is getting bigger and Chroma is taking more of it, but it is a strong player with broad capabilities inside a mature testing cake, not a category creator; by Baillie Gifford’s highest standard of “creating a new market,” it earns only half credit.

  Jun 10, 2026
• QCan its revenue at least double over the next five years? Will growth be driven mainly by volume, price, or new businesses?5/10

  A

  A five-year revenue doubling, from 2025 consolidated revenue of NT$28.311 billion to about NT$56.0 billion, implying roughly 14.6% per year for five years, is achievable and not unlikely. But this is a “high-cycle tailwind” assumption, not steady-state capability. The driver is mainly “volume,” with “price/product mix” as a secondary contributor; new businesses are not yet the main force.

  First, why a doubling is not far-fetched. Current growth is far above the 14.6% required for a doubling: the report discloses Q1 2026 revenue growth of 73% year over year and net profit attributable to the parent growth of 82% year over year; public data also confirms that Chroma’s April 2026 consolidated revenue was NT$4.866 billion, up 128.89% year over year, with January-April cumulative revenue of NT$16.725 billion, up 86% cumulatively year over year; Q1 consolidated revenue was NT$11.86 billion, net profit was NT$3.86 billion, and EPS was NT$9.12. Sell-side numbers have also been revised upward: under the FactSet framework, 15 analysts raised the median 2026 EPS estimate to NT$38.3, about 38% higher than the 2025 reported EPS of NT$27.70. As long as strength in the AI power chain and semiconductor testing lasts another two or three years, the revenue base will be lifted quickly. The arithmetic threshold for a five-year doubling has in fact already been covered by current growth.

  But the driver mix matters, and it is the key to the answer:

  • Volume is the main engine. The report is explicit about the two major growth engines in 2025: rapid data center expansion drove strong demand for high-power power electronics testing, with equipment revenue up +55% year over year, while AI IC system-level testing supported semiconductor demand, up +40% year over year. This is essentially “customer capex expansion → more shipped test equipment units/systems.” It is volume.
  • Price and product mix are amplifiers. Gross margin rose from 48% in 2021 to 61% in 2025, and further to 63% in 2026Q1, showing that higher-value AI power chain, SLT, and semiconductor solutions are taking a larger share, with better average selling prices and mix. This gives revenue growth “leverage,” but it depends on the cycle. It is not price increases pulled out of thin air.
  • New businesses are not yet the main force. Services and others account for only 4%–6% of the parent company, MAS accounts for 3% of the group, and Turnkey only 2%. In other words, the doubling would rely almost entirely on volume growth in the core testing equipment business. High-stickiness revenue such as software, services, and consumables, which would make a doubling more durable, is still a low share.

  So the honest conclusion is: a five-year doubling is likely achievable, but it depends heavily on the condition that “AI/power electronics/semiconductor test demand maintains high growth for many years,” and it is volume and mix rising together under a cyclical tailwind, not a subscription-like structural inevitability. The report itself also gives the counterevidence: revenue fell in 2023 from NT$22.067 billion in 2022 to NT$18.676 billion, proving that this growth line can be interrupted by downstream capital spending rhythms. It can double, but the “quality” of that doubling is cyclical, not compounding.

  Jun 10, 2026
• QFive years from now, what will take over as the next growth engine? Does this “second curve” exist today?4/10

  A

  Chroma’s “second curve” looks more like new seasonal branches growing from the same tree than an engine independent of the trunk. Candidates are visible today, but most are still extensions of the main “test equipment” curve. A genuinely different second growth pole has not yet taken shape.

  First, acknowledge that it has the base for “handoff.” Chroma is not a single-product company: its website discloses solutions spanning semiconductor and IC testing, power electronics, server power and BBU, EV and batteries, energy storage and PCS, passive components, electrical safety, photonics, display, RF wireless, PXI, smart manufacturing, Turnkey automation, and other scenarios. A broad product line means that when one demand line cools, another may step in. This is one reason it has been able to move through cycles in the past.

  But we need to distinguish “the next growth curve” from “a new section of the same curve.” The engines pulling it today are still fundamentally inside the main logic of “selling test equipment to customers making capital expenditures”:

  • AI power chain / BBU, PSU, and high-power power electronics testing: growth is very strong, with 2025 power electronics test equipment revenue up +55% year over year, but this is the current main engine. It belongs to the “first curve already being realized,” not the handoff five years from now.
  • Semiconductor system-level testing (SLT) / AI IC testing: 2025 semiconductor test revenue grew +40% year over year. This is also a higher-end extension of the main curve.
  • The candidates with true “second curve” characteristics are high-stickiness services, software, calibration, and long-term operations and maintenance. The report explicitly notes that if the share of these revenues continues to rise, the business model would move from “capital-goods cyclical” toward “compounder.” But in reality they are too small today: services and others are only 4%–6% of the parent company, MAS accounts for 3% of the group, and Turnkey only 2%.

  So when Baillie Gifford asks, “what takes over five years from now, and does that line exist today,” the honest answer is: the candidates exist today, but they are still seeds. Chroma’s regenerative growth currently comes from its ability to “keep growing new branches on the testing trunk (EV → energy storage → AI power → SLT),” not from having already cultivated a second curve that can carry the flag independently and is different in nature from equipment sales.

  This is exactly the gap between it and a “perpetual compounding machine.” One of the report’s core concerns is that operating cash flow and Owner Earnings have not grown in step with the revenue surge. In 2025, net profit attributable to the parent was NT$11.692 billion, while operating cash flow was only NT$5.258 billion. That shows current growth is still the old path of “heavy capital input in exchange for equipment shipments.” For the second curve to be real, the key is whether the share of high-stickiness revenue can rise materially over the next several years. Until then, the handoff logic only earns “partially passed.”

  Jun 10, 2026
• QWhat is its core competitive advantage? Will this moat widen or narrow over the next three to five years?5/10

  A

  Chroma’s moat is an engineering barrier built from “application know-how + customer validation cycles + product-line breadth + localized service.” It is medium-strength, rated 3/5 by the report. Over the next three to five years, it will probably widen in niches such as the AI power chain, power electronics, BBU, and PSU. But at the level of the overall test equipment industry, relative to top ATE players such as Advantest and Teradyne, the moat will not become “unchallengeable.”

  First, clarify what its moat is and is not. The report breaks it down item by item, with a sober conclusion:

  • What really carries weight is switching cost, medium to above-medium. Once a test platform enters a customer’s validation, mass-production, and quality systems, switching requires retesting, recalibration, and requalification. This is especially visible in power electronics, semiconductor, and automotive-grade scenarios. It is the fundamental reason part of its revenue can repeat and cross cycles.
  • Brand, scale, and channels are only medium. It has engineering credibility in niche fields, but it is not a consumer-grade superbrand. Revenue reached NT$28.311 billion in 2025 and coverage is broad, but in the highest-end global semiconductor ATE profit pool it is still not first tier.
  • Network effects and a data flywheel are almost absent. This is not a platform business. More customers do not raise value exponentially.
  • Patent/license-style institutional barriers are limited. There are certification and measurement capability thresholds, with its calibration laboratory holding ISO/IEC 17025 certification, but there is no institutional monopoly like pharmaceuticals or utilities.

  The evidence supporting “the moat is widening” is sustained R&D investment: the report discloses that 2025 parent-company R&D expense was NT$2.362 billion, 11% of parent-company revenue, and that ratio stayed stable at 11%–13% from 2021 to 2025. Continued spending on the technology platform, combined with multi-product-line engineering capability, means competitors would need more than 3–5 years to replicate Chroma’s solutions and customer validation experience in AI power chains, power electronics, and BBU. That is the source of widening.

  But the risk of “narrowing” is just as real and must be stated symmetrically:

  • In top semiconductor ATE, the deeper moat belongs to others. The report says directly that in cutting-edge semiconductor ATE technical barriers, Advantest and Teradyne are actually deeper, and both are expanding on the AI tailwind: Teradyne’s 2026Q1 semiconductor test revenue more than doubled year over year, with AI accounting for nearly 70% of its revenue, and Advantest is expanding tester capacity from 3,000 units to 5,000 units. Stronger players are also adding capacity in the arenas Chroma wants to expand.
  • Its current boom will attract followers, and the supply chain is relatively fragmented. It procures from two or more suppliers and materials are complex, showing that it is not locking the industry through extreme low cost.

  Overall judgment: in the “multi-product-line + AI power chain/power electronics niches where Chroma is strong,” the moat is more likely to widen over the next 3–5 years. But this is the width of a “strong engineering company,” not the width of an “irreplaceable monopoly champion.” It is closer to “a good company with broad capabilities in a good industry” than to the deepest moat one can hold blindly for ten years. That is also why the report gives the moat 3/5, not 4–5 points.

  Jun 10, 2026
• QIf its core business is disrupted, does it have the DNA to reinvent itself? How does it handle mistakes and bad news?4/10

  A

  Chroma has the base for “reinvention” but has not been through a real life-or-death test. With multiple product lines and sustained R&D, it has historically been able to shift to another leg when one demand line cools. In facing bad news, management has shown rare candor by proactively reminding investors to exclude one-time gains. But it has never experienced a complete disruption of its core business that forced it to tear itself down and rebuild, so this “reinvention DNA” is more inference than proven fact.

  First, the evidence for reinvention, in two layers.

  The first layer is business regenerativity. Chroma does not walk on one leg: its product lines span semiconductor/IC testing, power electronics, server power/BBU, EV and batteries, energy storage, photonics, display, RF, PXI, smart manufacturing, and more. This breadth gives it the ability to “find light in another direction when one side goes dark.” Historically it extended from test instruments into EV/batteries and then into today’s AI power chain and semiconductor SLT. The hardest piece of evidence is this: in 2023, when the industry weakened and revenue fell from NT$22.067 billion in 2022 to NT$18.676 billion, the company still generated NT$3.979 billion in net profit attributable to the parent. Maintaining profitability in a headwind year shows resilience in its engineering organization and cost structure. Sustained R&D investment, with 2025 parent-company R&D expense of NT$2.362 billion, 11% of parent-company revenue, and 11%–13% maintained for years, is the fuel for this regenerative capability.

  The second layer is the attitude toward mistakes and bad news, where the report gives a positive assessment:

  • The Q1 2026 earnings call materials explicitly state under the ROE/ROA table that “these ratios exclude one-time capital gains.” Management proactively reminded investors to distinguish “operating earnings” from “capital item gains,” and did not package one-time gains as operating strength. This is more candid than simply promoting headline EPS.
  • It also disclosed management’s own shareholding changes: the report records that Chairman Huang Chin-Ming transferred 10 million shares at NT$967/share on March 12, 2026 into newly established Hong Chang Investment, as payment for shares. This was not an open-market reduction, but it was a relatively large equity-structure adjustment, and the company disclosed it plainly so it can be tracked.

  But the answer must be complete and avoid exaggeration:

  • Its “reinvention” so far has been “growing new branches on the testing trunk,” not a true paradigm shift. There has never been an event where “the core test equipment business was completely disrupted by some technology and the company was forced to transform into something entirely different.” So the harshest hypothesis, whether it can reinvent itself when the core is disrupted, has no historical sample to falsify. This differs from companies that have survived close-range crises and rebuilt under pressure.
  • The capital-goods cyclicality of the business itself is the ceiling on regenerativity. The report repeatedly stresses that it is still mainly an equipment-sales, project-heavy capital-goods company, with low shares of services/subscription/consumables. That means its “leg switching” still circles inside the same heavy-capital business model, rather than jumping to a lighter and more disruption-resistant structure.

  Conclusion: management’s honesty and the business’s horizontal regenerativity let Chroma score reasonably well on “handling bad news” and “demand rotation.” But the gene for “reinventing from zero after the core is disrupted” has not been tested by a real event. By Baillie Gifford standards, it can only be marked as “has the base, untested.”

  Jun 10, 2026
• QDoes management, especially the founder, have a long-term view and deep alignment with the company? Is it willing to sacrifice current profit for five to ten years out?5/10

  A

  Management’s alignment with the company is decent, and its long-term view and capital allocation are generally sound, with the report giving management and capital allocation 4/5. But this is a “professionalized, steady, honest” team, not a founder-style all-in structure that is willing to sacrifice current profits substantially for five to ten years out. The depth of alignment and the aggressiveness of long-term bets are both upper-middle, not maximum.

  First, alignment. The evidence is concrete:

  • Shareholding alignment is upper-middle. The report discloses that in the 2026 major shareholder list, Chairman Huang Chin-Ming directly holds 2.55%; his spouse holds 1.01%, and holds 3.53% through related enterprise Hong Chang Investment, implying total related shareholding of about 7.09%. At the same time, as of the 2025 annual report printing date, there were no share pledges. Management is not a pure professional-manager structure with “zero skin in the game,” but a roughly 7% stake is also not founder-heavy control.
  • Share capital is extremely restrained, with almost no dilution. Capital stock was almost unchanged from NT$4.254 billion in 2021 to NT$4.252 billion in 2025. The 2022 employee restricted stock plan totaled 3 million shares, with 2.96 million issued, about 0.6961% of shares outstanding, an acceptable level. The company has not relied on large-scale issuance to “buy growth,” which is a friendly signal for long-term shareholders.
  • The willingness to pay dividends is strong. Cash returns mainly come through cash dividends rather than buybacks: from 2020 to 2025, cash dividend per share was about NT$4.50, NT$6.98, NT$8.00, NT$6.60, NT$9.00, and NT$19.50, stepping up over time.

  Next, long-term view and “willingness to sacrifice current profit for the future.” This is the question Baillie Gifford cares about most, and the answer is “willing to keep investing, but not an aggressive sacrifice type”:

  • It keeps paying for the technology platform: 2025 parent-company R&D expense was NT$2.362 billion, 11% of parent-company revenue, and has been stable at 11%–13% for years. This is an attitude of “not merely harvesting the cycle, but willing to pay for long-term capability.”
  • But its R&D/capital investment is “steady additions following the cycle,” not “sacrificing current-period profits to bet on a distant grand vision.” Chroma still has a high-dividend, low-leverage, stability-seeking financial style. You do not see founder-style aggressive behavior that intentionally suppresses current profits and pours cash into uncertain long-term projects. Its long-termism is more “steady continuation” than “all-or-nothing.”

  Two cautions need to be retained, and the report states them directly:

  • Honesty is acceptable but needs continuous monitoring. A rare positive is that management proactively noted in the 2026Q1 earnings call that “ROE/ROA excluded one-time capital gains,” and did not treat one-time gains as operating strength. This is more candid than packaging headline EPS.
  • On March 12, 2026, the chairman transferred 10 million shares at NT$967/share into newly established Hong Chang Investment, as payment for shares. Although this was not an open-market reduction, it was a relatively large equity-structure adjustment and deserves ongoing tracking.

  Overall: this is a trustworthy management team with rational capital allocation, upper-middle alignment, and good honesty. In the report checklist, both “is management trustworthy” and “is capital allocation rational” are marked as passed. But by Baillie Gifford’s highest yardstick of “whether the founder is willing to sacrifice today for ten years out,” it is “steady long-termism,” not “aggressive long-term betting.” So it earns 4/5, not 5.

  Jun 10, 2026
• QIf it disappeared tomorrow, how much would customers miss it? Is its growth sustainable, without relying on social harm or regulatory arbitrage?5/10

  A

  If Chroma disappeared tomorrow, customers would find it “very inconvenient and would pay a meaningful cost in retesting/requalification,” but it would not be “the whole production line immediately stops with no substitute.” In power electronics, semiconductor, and automotive-grade test steps where it has already been adopted, it has real indispensability through medium-to-above-medium switching costs, but it is not a sole-source node whose failure halts production. Its growth method is sustainable and does not rely on harming society or regulatory arbitrage. That is its clean side.

  First, “how much would customers miss it” needs a dual lens: indispensability plus substitutability.

  • The indispensable side: the report notes that once a test platform enters a customer’s validation, mass-production, and quality systems, switching requires retesting, recalibration, and requalification, especially in power electronics, semiconductor, and automotive-grade scenarios. In other words, for customers that have embedded Chroma equipment into production lines and quality systems, losing it suddenly would create real time cost and validation cost. That is the source of “missing it,” and also the root of part of its recurring revenue.
  • The replaceable side: the report also says plainly that it “is not irreplaceable.” In the highest-end semiconductor ATE, Advantest and Teradyne have deeper moats and can take over; in broader electronic measurement, Keysight is also an important reference. So Chroma is not the throat node where cutting it off stops the whole industry. It is more like a key supplier that is costly to replace, but where substitutes exist in the ecosystem. Customers would miss it, but not to the point that they cannot live without it.
  • One quantitative supporting point: the report discloses that in 2025 the largest customer contributed about NT$5.100 billion in revenue, about 18.0% of that year’s revenue, versus about 16.6% in 2024. This shows it has significant value to major customers and is heavily adopted. But “indispensability” runs both ways: customers matter to Chroma, and Chroma matters to customers, yet neither side is bound to the other at the level of survival.

  Next, “is growth sustainable and not reliant on social harm or regulation.” This is a positive point for Chroma:

  • The source of growth is legitimate industrial demand, not regulatory arbitrage or harmful externalities. Demand comes from testing needs in the AI data center power chain, semiconductors, EV/batteries/energy storage. The IEA Global EV Outlook 2026 records that global EV sales exceeded 20 million vehicles in 2025 and battery deployment reached 1.2 TWh, up nearly +30% year over year. Test equipment is exactly the step that helps ensure these products are “safe, compliant, and mass-producible,” with a positive social role: helping customers make batteries, power supplies, and chips more reliable. This is not growth through exhausting users, evading regulation, or creating negative externalities.
  • Compliance and certification are assets, not burdens: its calibration laboratory holds ISO/IEC 17025 certification, and electrical safety testing itself is a tool to help customers meet regulation.

  But the real variable in sustainability is not “society/regulation,” but “cycle and geopolitics.” The report lists geopolitical tensions, export controls, a pullback in the AI investment cycle, and customer concentration as risks. Its customers span the Americas (about 31.9%), Taiwan (about 31.6%), and China (about 25.5%), exposing it to both U.S.-China technology investment cycles and supply-chain restructuring. This is the fragile part of growth sustainability, but it is cyclical and geopolitical risk, not a moral/regulatory flaw of “growing by harming society.”

  Conclusion: customers would genuinely miss it, but there are substitutes; growth is clean and aligned with social and regulatory direction. Chroma scores positively on this question, with deductions only because “indispensability” has not reached sole-source level.

  Jun 10, 2026
• QWhat are the unit economics of this business, in gross margin and incremental returns? Do they improve or deteriorate as scale increases? Where does the money it earns go?7/10

  A

  Chroma’s unit economics look excellent at the “reported gross margin” level and improve with scale, but they need a discount at the “cash return” level. A large amount of incremental money is absorbed by working capital and capex, so this is a capital-goods model with “high accounting profits and weak cash conversion,” not a software-like high-incremental-return business that gets lighter as it grows. The money it earns mainly goes to working capital, equipment/engineering investment, and cash dividends.

  Start with the bright side: gross margin and scale effects are real:

  • Gross margin rose from 48% in 2021 to 61% in 2025, and further to 63% in 2026Q1. This shows that product mix, with higher shares of AI power chain, SLT, and semiconductor solutions, and pricing power are indeed improving. As scale grows, unit gross profit is getting better.
  • Net margin reached 41.3% in 2025, but this figure must be unpacked. It was heavily flattered by one-time gains, discussed below, and should not be treated as steady-state unit economics.
  • Interest coverage is almost not a constraint: under the report’s framework, 2025 operating profit was NT$9.198 billion and finance cost was only NT$0.029 billion, with interest coverage above 300 times. Capital structure is not a drag on unit economics.

  Now the dark side, which is the core of this question: incremental returns are not good on a cash basis:

  • Reported profit is far above operating cash. In 2025, net profit attributable to the parent was NT$11.692 billion, but operating cash flow was only NT$5.258 billion. A large part of the gap is cash absorbed by working capital: the report discloses that accounts receivable increased by NT$3.021 billion and inventory increased by NT$2.654 billion that year.
  • In high-growth periods, the business actually “gets heavier as it grows.” In Q1 2026, while revenue surged 73% year over year, operating cash flow was only NT$1.016 billion and free cash flow was instead -NT$0.825 billion. This is the most telling set of numbers: incremental revenue did not bring incremental free cash in step; working capital and construction investment consume a significant share of cash when a high-cycle volume surge is underway. This is the opposite of excellent software/platform unit economics where marginal cash increases sharply after scale is reached.
  • The 41.3% net margin was lifted by one-time gains. In 2025, net non-operating income was as high as NT$4.720 billion, including NT$3.197 billion of gains recognized on disposal of non-current assets held for sale and NT$3.884 billion of disposal proceeds received. After stripping this out, the real operating unit economics are not as stunning as the accounts suggest. On this basis, the report estimates conservative Owner Earnings at about NT$4.358 billion, calculated as operating cash flow of NT$5.258 billion less about NT$0.9 billion of maintenance capex, far below accounting profit of NT$11.692 billion.

  Finally, “where does the money it earns go,” in three directions:

  1. Working capital (accounts receivable +NT$3.021 billion, inventory +NT$2.654 billion, 2025): it supports volume growth but ties up large amounts of cash.
  2. Capital and engineering investment: in 2025, cash paid for PP&E was NT$0.622 billion and intangible assets were NT$0.114 billion, with another NT$1.661 billion in equipment and engineering prepayments, plus R&D expense of NT$2.362 billion, 11% of parent-company revenue, continuing to fund the technology platform.
  3. Cash dividends: the 2025 cash dividend per share was NT$19.50, more than twice the NT$9.00 in 2024. Dividends are the main return method, with almost no buybacks and almost no dilution from issuance.

  Conclusion: unit economics are excellent and scale-friendly at the gross-margin level, but incremental returns are weak on a cash basis during high-growth periods, even temporarily negative in FCF. This is a capital-goods business with “very high reported margins, but continuous feeding of working capital and capex.” That is the fundamental reason the report insists on valuing it by Owner Earnings rather than accounting profit, and remains cautious about the current price.

  Jun 10, 2026
• QWhat conditions must all hold for it to rise fivefold in ten years? Are those conditions realistic? What expectations are embedded in today’s share price?2/10

  A

  For Chroma to rise fivefold in ten years from about NT$2,390/share today (the June 9 report snapshot; it had fallen intraday on June 10 to about NT$2,250) to about NT$12,000/share, implying a market value of about NT$5 trillion, a long list of conditions would need to hold “simultaneously.” Stacked together, those conditions are unrealistic, because today’s share price has already priced “high growth continuing for many years” as the base case, leaving almost no room for future upside. This is the core reason the report rates it “Avoid.”

  First, quantify “what expectations are embedded in today’s share price.” This is the key point, and the numbers have been web-verified:

  • The current price implies a P/E of about 86 times based on 2025 reported EPS of NT$27.70; stockanalysis gives a trailing P/E of about 75.6 times. P/B is about 31–35 times, P/S about 35.8 times, and EV/EBITDA about 68 times. Market cap is about NT$1.01 trillion, share capital about 423.67 million shares, and the share price rose about 633% over the past year.
  • On an “Owner Earnings” basis, the valuation is even harsher: the report estimates conservative 2025 Owner Earnings at NT$4.358 billion, calculated as operating cash flow of NT$5.258 billion less about NT$0.9 billion of maintenance capex, or about NT$10.3 per share. The current price implies P/Owner Earnings of about 235 times.
  • Even under the most generous optimistic DCF, with starting Owner Earnings of NT$5.0 billion, ten-year growth of 15%, a 9% discount rate, and terminal growth of 3.5%, the discounted value per share is only about NT$540, still several times below today’s price. The report’s core intrinsic value range is NT$160–450/share, and NT$2,390 is a premium of about 199%–1,394% over that core range.

  In other words, the expectations embedded in today’s share price are: Chroma must keep delivering “super growth” for many years, cash quality must catch up with profits, and the market must be willing to maintain valuation multiples far above traditional capital goods for a long time. Only if all three hold do you barely avoid losing money.

  Specifically, which conditions must hold “simultaneously” (the three most fragile chains identified by the report plus my added arithmetic):

  1. Demand must keep growing at a high rate for many years without flaming out: AI infrastructure, power electronics, and semiconductor test demand must keep growing rapidly for many years, not just for a two- or three-year upcycle. There is real current-cycle evidence: 2026Q1 revenue +73% year over year, net profit NT$3.86 billion, EPS NT$9.12; April monthly revenue NT$4.866 billion, up +128.89% year over year; and 15 sell-side analysts lifted the 2026 median EPS estimate to NT$38.3. But sustaining high growth for ten years on an already very high base is extremely difficult.
  2. Cash quality must catch up with profit: working-capital absorption must ease materially, and operating cash flow/Owner Earnings must truly step up to more than NT$10.0 billion, realized through pure operations rather than asset disposals. Yet in 2026Q1, while revenue was +73%, operating cash flow was only NT$1.016 billion and FCF was -NT$0.825 billion. The direction is the opposite.
  3. Valuation multiples must not compress: the market must tolerate 80-plus times PE and 30-plus times PB for a long time, even though this is itself the most uncontrollable variable over ten years and the one most likely to mean-revert.
  4. Margins must stay high: gross margin must remain around 60% for the long term and not fall back into the 50% range because of intensifying competition or product mix normalization.

  Looked at individually, none of these conditions is absurd. Requiring them to hold simultaneously for ten years is highly unrealistic. If any one loosens, the fivefold target fails. The report’s reverse calculation is more direct: buying at the current price and holding for ten years, assuming valuation eventually returns to a more reasonable range, gives expected annualized returns of about -17%/-14%/-7% under conservative/base/optimistic scenarios, and those scenarios already assume substantial growth and terminal multiples.

  Conclusion: a fivefold return in ten years mathematically requires four things to stack together: demand does not fade, cash improves, high valuation does not compress, and margins do not fall back. The realistic probability is low. Today’s price is not paying for “upside imagination”; it has already pulled forward many years of good news. That is the definition of no margin of safety.

  Jun 10, 2026
• QWhy has the market not realized all this yet? Is it because investors do not understand it, look down on it, or cannot look far enough? What will become the “narrative inflection point”?3/10

  A

  For Chroma, this signature Baillie Gifford question, “why has the market not realized it yet,” has to be asked in reverse. This is not an overlooked growth stock that the market fails to understand. Quite the opposite: the market understands it too well, and has excitedly priced in all the good parts. So the perception gap here is not “undiscovered undervalued value,” but “the price is overextended and risks are covered by an optimistic narrative.” The narrative inflection point will not be “the market suddenly discovers it is good,” but “one day the market discovers it is not that perfect.”

  First, compare the three classic perception gaps of “does not understand / looks down on it / cannot look far enough.” All point to “not applicable in the undervaluation direction”:

  • Does not understand? Not true. Chroma is a star AI concept stock in Taiwan, with ample coverage: under the FactSet framework, 15 analysts already cover it, have raised 2026 median EPS to NT$38.3, and have a target price of NT$2,450. Institutional coverage is dense and sell-side estimates continue to rise. It is not an obscure stock with no coverage that investors cannot understand.
  • Looks down on it? The opposite. The share price rose about 633% over the past year, market cap is about NT$1.01 trillion, and trailing P/E is about 75.6 times, about 86 times under the report EPS basis. The market not only does not look down on it, it gives it a premium far above Advantest, Teradyne, and Keysight.
  • Cannot look far enough? The opposite. The report’s core argument is precisely that the market is looking too far out and has already discounted many years of high growth into today’s price, with P/Owner Earnings of about 235 times. This is not “a short-sighted market missing long-term value,” but “a market valuing it on the most optimistic long-term assumptions.”

  So where is the real “perception gap”? On the risk side, not the value side. That is the key to an honest answer:

  1. Earnings quality is commonly treated as operating strength. In 2025 net profit of NT$11.692 billion included a one-time gain of NT$3.197 billion from disposal of non-current assets held for sale, while net non-operating income reached NT$4.720 billion; yet operating cash flow was only NT$5.258 billion. If the market looks at headline EPS/FCF and concludes the “valuation is not expensive,” it is misreading the company’s real cash-generation ability. This is the blind spot covered by the optimistic narrative.
  2. Negative cash conversion in a high-growth period is being ignored. In 2026Q1, when revenue was +73%, operating cash flow was only NT$1.016 billion and FCF was -NT$0.825 billion. The faster volume ramps, the more cash working capital and capex consume. The market is excited by triple-digit year-over-year monthly revenue growth (April monthly +128.89%), but has not fully priced in that “growth burns cash.”

  So the “narrative inflection point” will be these things, not the realization of good news:

  • If growth loses speed and it is not seasonal: a clear slowdown in AI power chain/semiconductor test-related growth would immediately force the market to reassess the pricing premise of “many years of high growth.” The report lists this as the first trigger signal.
  • The next time “beautiful profits without cash” is recognized: if large non-operating asset disposal gains again inflate profit, or operating cash flow continues to lag net profit and Owner Earnings still cannot step up, the narrative will shift from “super growth” back to “cyclical capital goods.”
  • Margins fall from high levels: if gross margin persistently falls below the mid-to-high 50% range, even if it only moves “from excellent back to pretty good,” that can trigger violent multiple compression under a framework of 80-plus times PE and 30-plus times PB.
  • Customer concentration worsens or geopolitics hits: the largest customer accounts for about 18% of revenue, layered with U.S.-China technology investment cycles and export controls. Any negative development could become a sentiment inflection point.

  In one sentence: this is not an overlooked stock where “the market has not yet realized its greatness.” It is a high-expectation stock where “the market has fully realized it and already bought all the benefits in advance.” Its perception gap is that risk is underestimated, not that value is underestimated; the probability of a downward narrative inflection is far higher than an upward one. That is also the fundamental reason the report rates it “Avoid” and recommends putting it on a high-quality watchlist to wait for a good price, rather than chasing it here.

  Jun 10, 2026