Report contents
Industry landscape
Where the company competes: a chip-level oligopoly inside a crowded module market
Coherent earns its money at one narrow, technically demanding step in the electronics supply chain: growing compound semiconductor crystals and turning them into devices that emit, detect and steer light. That step is concentrated among a handful of firms that own fabs and process know-how. The assemblies built around it, above all optical transceivers, are not. Coherent sells into both layers, and the gap between them explains most of what happens to its gross margin from one quarter to the next.
The company makes indium phosphide and gallium arsenide laser chips, pump lasers, optical transceivers for datacom and telecom networks, industrial lasers, and precision optics for lithography and life-science instruments. Coherent's 10-K reports three segments: Networking, Materials and Lasers. Networking, which supplies transceivers and optical components to hyperscale cloud operators and telecom equipment vendors, is the largest by revenue and the one that sets the stock's direction. Materials sells the compound semiconductor substrates and engineered materials that feed both Coherent's own fabs and outside customers. Lasers sells fiber, carbon dioxide and ultrafast laser systems into machine tools, welding, marking and medical device manufacturing.
Bargaining power sits with the buyers at the top of that chain and with a few material suppliers at the bottom. On the customer side, a small group of hyperscalers and network equipment makers runs formal qualification programs, qualifies two or three sources per design, and renegotiates pricing at every speed transition. On the supply side, germanium, indium, gallium and certain specialty gases come from a limited set of producers, some of them sole-sourced. Coherent's partial answer is vertical integration: it grows its own crystals rather than buying finished laser chips, which protects supply and captures the epitaxy and fab margin that a pure module assembler leaves on the table.
The competitive geography is global but the negotiation is bilateral. Coherent manufactures in the United States, Europe and Asia, and its customers sit in North America and China. Because each transceiver design is qualified customer by customer, the effective market is not one pool of buyers shopping on price but a series of two- or three-supplier relationships inside large cloud accounts. Industrial lasers behave differently: they sell through machine-tool builders and distributors, carry service and applications engineering, and compete regionally, where a local service technician matters as much as a specification sheet.
Market structure: five or six credible suppliers per arena, and different barriers in each
The competitive set depends on which arena you draw the line around.
- Datacom and telecom optics: Lumentum, Broadcom, InnoLight Technology, Eoptolink, Hisense Broadband and Applied Optoelectronics, alongside Coherent. Japanese and Chinese component makers also supply chips and subsystems.
- Industrial lasers: IPG Photonics, Trumpf and nLIGHT are the names that recur in fiber and solid-state systems.
- Compound semiconductor materials: a longer list of substrate and epitaxy suppliers, several of them Asian, competing on wafer size and defect density.
Concentration is real but uneven. In indium phosphide laser chips, the count of suppliers with volume production and hyperscaler qualification is small enough to sit on one hand, which supports pricing and makes a new entrant's path expensive: a fab, years of yield learning, and a customer qualification cycle that runs across several quarters before the first meaningful shipment. At the module assembly step, the count is larger and the barriers are lower. Assembly and test are more labor-intensive, capital needs are modest relative to a wafer fab, and each new speed generation, from 400G to 800G to 1.6T, resets the design. That reset is the opening through which new suppliers win sockets, and it is why a company can hold a strong chip position and still face price pressure on the box it ships.
Coherent's portfolio is broader than any single rival's, a consequence of the July 2022 merger with Coherent Inc., which combined II-VI's compound semiconductor and optics business with a laser systems company. Breadth cuts both ways. It gives Coherent content in more designs and internal demand for its own materials, but it also puts the company into arenas where it is one of several capable suppliers with no structural edge.
The practical read: this is a market where the leading chipmakers can be counted on one hand, so entry barriers at the device level are high. At the module level they are medium at best, and that is where Coherent books a large share of its revenue. The question the rest of this analysis has to answer is whether the chip-level position is strong enough to keep the module business priced like an oligopoly rather than a contract assembly shop.
Competitive advantage
Does the company have a moat? Moderate, because pricing power has not reached the returns
The barrier is real and sits in one place: qualified fabs that can grow, process and test indium phosphide laser chips at volumes hyperscalers will accept. The financial record is what keeps the rating short of strong. Coherent's gross margin in fiscal 2026 was 37.5%, essentially the 38.2% it earned in fiscal 2022 on less than half the revenue, and its estimated return on invested capital in this, its best year in five, sits near the bottom of the range that usually signals a protected franchise.
Concentration supports the possibility of a moat without proving one. A handful of firms own volume compound-semiconductor fabs, entry requires years of yield learning, and customer qualification runs across several quarters; that structure is why the industry can earn above its cost of capital in good years. The proof is whether the returns persist. Coherent has posted one good year after three poor ones, so the evidence supports a genuine but narrow barrier that has not yet been converted into durable excess returns.
Market position: a qualified supplier, with no confirmable share
Market position: Moderate. Fewer than a handful of vendors ship volume indium phosphide laser chips into hyperscale designs, and Coherent is one of them. A dated share series is not available, so ranking the company first or second in any of its arenas would be an assertion rather than a finding, and share stability across five years cannot be confirmed.
What is measurable is breadth and what it costs. Revenue rose from $4.71 billion in fiscal 2024 to $5.81 billion in fiscal 2025 and $7.12 billion in fiscal 2026, growth of 23.4% and 22.5%, after a fall from $5.16 billion in fiscal 2023. The 2022 merger with Coherent Inc. put the company into lasers, optics and materials alongside datacom transceivers. That spread gives it internal demand for its own substrates and content in more designs. It also leaves Coherent competing in arenas where several capable suppliers operate and where it holds no structural edge.
Profitability: returns sit near the cost of capital
Profitability: Weak. In fiscal 2026, operating income was $0.90 billion and the tax provision was $0.06 billion, about 7% of pre-tax income, a rate low enough to flatter the 11.4% net margin. Estimated after-tax ROIC: NOPAT of roughly $0.71 billion at a 21% tax rate on invested capital of about $13.0 billion (equity of $10.90 billion plus debt of $3.28 billion less cash of $1.16 billion), or 5.5%. The same calculation gives about 4% for fiscal 2025 and about 1% for fiscal 2024. Fiscal 2022, before the merger, came in near 8% on a $4.1 billion capital base.
Stripping the $4.38 billion of goodwill out of invested capital lifts the fiscal 2026 figure to roughly 8.2%, an estimate that measures the operating businesses instead of the price paid for them. Either way the company falls short of the 15% to 25% that would mark a protected franchise, and the goodwill is a claim on future excess returns that has not been earned yet. ROE tells the same story with a leverage assist: 8.5% in fiscal 2026, against 0.6% in fiscal 2025 and negative 2.1% and negative 4.5% in the two years before that.
Margins: a component supplier's spread, improving
Margins: Moderate. Gross margin went from 31.0% in fiscal 2024 to 35.5% in fiscal 2025 and 37.5% in fiscal 2026. The quarter ended September 30, 2026 reached 39.1% on revenue of $2.37 billion, up 50% from $1.58 billion a year earlier. Net margin swung from negative 3.3% to 11.4% over those two years, and net income reached $0.81 billion.
The level matters more than the direction. A 37.5% gross margin and a 12.6% operating margin are respectable for a hardware supplier and unremarkable for one with pricing power, and the margin is close to where it stood four years ago on a much smaller revenue base. Cash conversion reinforces the caution. Operating cash flow was $0.08 billion in fiscal 2026 against $0.81 billion of net income, and free cash flow was negative $1.02 billion after roughly $1.10 billion of capex, up from about $0.44 billion the year before. Growth is being bought with capacity, and the return on that capacity has not arrived in cash.
Overall rating: Moderate
Overall: Moderate. The decisive facts are a fiscal 2026 estimated ROIC of 5.5% to 6.5% and a gross margin that has not beaten its fiscal 2022 level despite revenue that is 2.1 times larger. Both point to real technical barriers alongside limited pricing power, and the Weak profitability rating caps the overall call at Moderate. The sub-ratings agree, so nothing here rests on a share judgment I could not document.
Where the advantage comes from and how long it can last: know-how and design-ins, both conditional
Two sources carry whatever advantage Coherent has: process know-how in compound semiconductors, and the qualification that puts a Coherent chip inside a hyperscaler's transceiver design. Scale in capacity is the third, and it reinforces the first two only while demand for indium phosphide optics holds.
One metric tests all three, and it is gross margin through the current speed transition. In this business the chip is where pricing power lives and the module is where it leaks. If the barrier is at the device level, the move from 800G to 1.6T should hold gross margin near 39% as volume grows. If margin slides back toward the low 30s while revenue climbs, Coherent is buying sockets with price.
Supply-side advantage: crystal growth and fab yield, Moderate
Coherent grows its own crystals, runs the epitaxy, and processes and tests in-house, which a module assembler does not. A rival can buy a fab; it cannot buy the yield curve. Laser chips live or die on defect density, and that learning accumulates through years of production, which is why the supplier count stays small and why a new entrant's early wafers cost far more than Coherent's marginal ones. The catch is that this is know-how, not a patent estate. It walks out the door with engineers, and Chinese substrate and chip suppliers have been closing the gap on mature parts. A process edge of this kind buys years, not a decade.
What it is worth shows up in the capital being committed. Capex rose to about $1.10 billion in fiscal 2026 from $0.44 billion in fiscal 2025. Over the same year revenue rose $1.31 billion and gross profit rose $0.61 billion, so each dollar of new capex came with about $0.55 of added gross profit, an estimate taken before depreciation on the new tools. That is a fine return if the equipment stays loaded for five years and a poor one if demand pauses, because the depreciation arrives either way.
What would break it: a rival reaching comparable yields on 1.6T parts and winning a volume socket at a top-three cloud customer.
Demand-side advantage: qualification locks a socket, not a customer, Moderate
Within a design, swapping the laser chip means requalification and outage risk, and that is real captivity for the life of the program. It is why gross margin held in the mid-30s even through the loss years of fiscal 2023 and fiscal 2024. But the lock expires at every speed generation, hyperscalers qualify multiple sources on purpose, and the buyer list is short and each name is large enough to negotiate hard. The financial signature of partial captivity is that scale does not lift price. Revenue has more than doubled since fiscal 2022, and gross margin has moved from 38.2% to 39.1% at best.
The recovery is worth real money even so. Going from a 31.0% gross margin in fiscal 2024 to 39.1% in the September 2026 quarter is about $0.58 billion a year of gross profit at fiscal 2026 revenue, roughly 8 points on $7.12 billion, an estimate. That figure is the value of the current capacity tightness. It is a return to a prior level, not a new plateau, and treating it as permanent is the main way to overpay for this business.
What would break it: the first 1.6T sockets going to a competitor, or a hyperscaler adding a third qualified chip source.
Economies of scale: capex is the entry ticket, and the ticket is being punched, Moderate
Fixed costs here are fabs, epitaxy reactors, test capacity, and for industrial lasers a field service and applications engineering organization. Capex equaled 15.5% of revenue in fiscal 2026, against 7.6% in fiscal 2025, which is the entry ticket being paid. Scale helps only if the fabs stay loaded, and scale without captivity is not a barrier, because an entrant with a committed cloud customer behind it can finance a fab. Buyers have every reason to want that second source. The barrier therefore deters the unprepared entrant, not a strategic one.
The balance sheet gives Coherent room to spend: net debt of about $2.5 billion at September 30, 2026 against $11.3 billion of equity. The risk is timing. Free cash flow was negative $0.87 billion in that quarter alone, so a pause in AI datacom spending would leave depreciation and interest without the revenue that justified them.
Durability: silicon photonics is the substitute that matters
The substitute to watch is not another laser maker. It is co-packaged optics and silicon photonics, which move the light engine onto CMOS wafers where Broadcom and the foundries hold the process advantage. Coherent would still sell lasers into such systems, but the value would migrate toward whoever owns the photonic integrated circuit, and the margin pool with it. Whether Coherent's position in that architecture is as strong as its indium phosphide position is not settled by the evidence here.
Moat strength: partial captivity plus partial scale
The strongest combination is customer captivity reinforced by economies of scale; Coherent has partial versions of both, plus a cost edge that rests on know-how, the easiest source to copy. That is enough to keep a crowd of new entrants out of qualified chip sockets and not enough to earn 15% returns through a cycle. The rating is Moderate.
Two questions decide whether that changes. Can gross margin hold near 39% through the 1.6T ramp while capex falls back below 10% of revenue? That combination would show the chip position is being paid for, and it would support a Fairly Strong rating. Or does gross margin slip under 33% on rising revenue, which would say the module business sets the price and the chip merely wins the design? The next four quarters of margin against capex will answer more than any share estimate.
Generated by DeepVal AI. Data and models may contain errors or be outdated. For research only; not personalized investment advice.
Continue reading the full report
- Comparison and summaryHow Coherent Corp compares with its closest rivals, and our moat rating
Free to sign up · 20 research points · unlocking this report uses 10
20 points cover 2 full reports. No card needed.
Research reflects its generation date. Preview the first two modules.
More on Coherent Corp
Coherent Corp
Business, moat and key questions
Open company page