Report contents
Industry landscape
Where the company competes: a merchant photonics supplier in a race run by larger firms
Quantum Computing Inc. sells into what is still mostly a research budget line, and the precise arena matters more than the label. Its submarket is merchant photonic quantum hardware: thin-film lithium niobate (TFLN) photonic chips, the fabrication work that turns them into devices, and small optical computing systems sold to labs, universities, government programs and a few systems integrators. That is narrower than "quantum computing," where the leaders are trying to own entire machines. QUBT is largely a component and tooling supplier to that race, which sets both the size of the profit pool it can reach and the bargaining position it brings to it.
Value chain: the scarce inputs sit upstream and the margin sits downstream. Upstream are lithium niobate wafers, electron-beam lithography and thin-film deposition tools, cryostats, lasers, and the trained engineers who can pattern a photonic circuit. QUBT occupies the middle, designing chips and running its own small fabrication line. Downstream sit research groups, defense and intelligence agencies, national laboratories, and cloud providers that rent time on finished machines. In most hardware chains the durable margin accrues to whoever owns a working system plus a distribution channel: cloud access, or a government prime contract. Fabrication and component work is usually reimbursed closer to contract manufacturing, through milestone payments and non-recurring engineering fees. If that pattern holds here, QUBT earns a slice of research and defense spending rather than a share of the end-market value, and its pricing power depends on being one of very few places a customer can get a custom photonic chip made.
Geography: global, talent-bound and shaped by export controls. Competition does not turn on freight radius or local distribution. It turns on where wafers, tools and PhDs are. Buyers cluster in the United States, Europe and Japan, while Chinese institutions face restrictions on importing advanced lithography and cryogenic equipment. That makes the relevant scale economic rather than geographic: a photonic fab competes on wafer starts and on reusing designs across customers, so a merchant supplier with a qualified process serves a worldwide niche from one site. A regional footprint would be worth little.
Market structure: many credible systems players, few merchant chip suppliers. The system level has at least seven or eight well-funded contestants, split across superconducting circuits, trapped ions and photonics. The merchant TFLN component level is thinner. Revenue shares for quantum hardware are not published on a consistent basis, so the table orders players by their position in the chain instead.
| Player | Technology | Position in the chain |
|---|---|---|
| IBM, Google Quantum AI | superconducting circuits | full systems plus cloud access |
| IonQ, Quantinuum | trapped ions | full systems plus cloud access |
| D-Wave | superconducting annealing | systems plus cloud access |
| PsiQuantum | silicon photonics | full systems, foundry partners |
| Quantum Computing Inc. | TFLN photonics | chips, foundry work, small systems |
Barriers to entry: medium at the chip level, weak at the system level. Building a photonic fab takes capital, process know-how and qualified engineers, and the customer list of government labs and defense programs is short and relationship-driven. Those are real obstacles for a new entrant. The systems layer is a different story: seven or eight credible, well-capitalized competitors is too many for any one of them to count on leader-style returns from entry barriers alone, and the technical benchmarks that would settle a winner are not yet agreed. What protects QUBT, if anything, is the reverse side of a small market. Few firms can make a custom TFLN device, but few buy one, so concentrated supply has not translated into concentrated demand or pricing power.
That asymmetry is the open question for the moat assessment. A defensible niche supplier would show repeat customers, rising design wins and stable or improving gross margin on foundry work as volume grows. Evidence of one-off government awards and lumpy system sales would point the other way, toward a contract business earning a return on specialized labor rather than on a durable position. Watch whether purchase commitments and customer concentration in the filings reflect recurring buyers or a handful of programs that renew at the government's discretion.
Competitive advantage
Does the company have a moat? Weak: the scarce input is customers, not chips
Moat rating: Weak. Quantum Computing Inc. designs thin-film lithium niobate (TFLN) photonic chips and the small optical systems built around them, and that process is not something a new entrant can stand up next quarter. The constraint sits on the other side of the trade. Its buyers are a short list of university groups, national laboratories and defense programs spending research budgets, and a supplier that few can replace but few need sets neither price nor volume. A moat shows up as concentration plus returns that stay above the cost of capital. QUBT has a narrow version of the first and has never had the second.
The income statement makes the case. Gross margin fell from 55.1% in FY2022 to 45.3% in FY2023, 30.0% in FY2024 and 9.8% in FY2025, then 4.5% in the quarter ended September 30, 2026, according to the 10-Q filed November 9, 2026. Mix explains part of that: system deliveries carry hardware content that chip work does not. Mix does not explain all of it. With revenue this small, one order moves the ratio, so read the series for direction rather than precision, and the direction is a supplier conceding price or absorbing cost to keep work moving.
Market position: a niche supplier whose share no one publishes
Market position: Weak. The company does not break out revenue by product line, and no third party publishes a comparable share series for merchant TFLN suppliers, so a rank cannot be established. What the statements do show is scale: revenue rounds to zero at the reporting precision used through FY2025, and the September 2026 quarter shows $0.01 billion. The competitive set at the systems level includes IBM, Google Quantum AI, IonQ, Quantinuum, D-Wave and PsiQuantum, all of which sell finished machines or cloud access. QUBT sells into that group's supply chain, where the buyer holds the pricing lever.
Share stability: Weak. With fewer than three reliable years of share data, stability cannot be confirmed. The observable proxy is lumpiness: quarterly gross margin ran from 32.8% in the September 2025 quarter to negative 49.5% in the December 2025 quarter, which is what a business looks like when a handful of orders set the quarter.
Profitability: no positive operating year in the record
Profitability: Weak. Return on equity was negative in every year from FY2021 through FY2025: -181.3%, -94.9%, -40.4%, -77.9% and -2.2%, with return on assets tracking the same path from -171.4% to -2.1%. Operating income was negative in each of those years and in each of the three quarters reported for FY2026, so after-tax ROIC (return on invested capital, profit after tax divided by the debt and equity funding the business) is negative and not worth computing as a positive figure.
The FY2025 ROE of -2.2% is not an improvement in the business. Equity grew from $0.11 billion at the end of FY2024 to $1.60 billion at the end of FY2025, mostly through a $1.48 billion financing inflow, so the same dollar of loss is divided by a much larger base. There is no debt on the balance sheet, which means these returns reflect operating losses rather than leverage.
Against a rule of thumb that average after-tax ROIC of 15% to 25% over a decade or more signals a protected franchise, and 6% to 8% signals none, QUBT sits below both because it has never cleared zero. Operating cash flow was negative in every year from FY2021 through FY2025 and in the first nine months of FY2026, totaling about $130 million of cumulative cash consumption, funded by share sales rather than by customers.
Overall rating: Weak
Two facts carry the call: gross margin fell roughly 50 percentage points between FY2022 and the September 2026 quarter while revenue was still forming, and operating income has been negative in every period the statements cover, from FY2021 through the nine months ended September 30, 2026. The business has technology worth owning and a customer base that is real but tiny and discretionary. Every sub-rating lands at Weak, so the overall rating follows without adjustment. What keeps this from None is that the TFLN process and the fabrication relationship are genuinely uncommon; what keeps it from Moderate is that the scarcity has never shown up in a price.
Where the advantage comes from and how long it can last: know-how buyers have not yet paid for
The single metric that tests this moat is gross margin on product and foundry revenue. A supplier with a protected process and repeat design wins should earn more per revenue dollar as it spreads engineering across more customers. QUBT's margin went the other way. Three possible sources are worth weighing: process know-how, the buying relationships, and scale. Each is rated below.
Supply-side advantage: Moderate, and tied to one process
Supply-side advantage: Moderate. The mechanism is requalification cost. A research group or defense program that has a chip designed and fabricated on QUBT's process cannot simply move the design to another shop; the masks, the process design kit, the calibration and the yield learning all have to be rebuilt, and that costs months of a program's schedule. That gives QUBT real protection on the next order from an existing customer, which is why the rating is not lower.
The protection stops at the edge of the current design. Each customer's spend is small enough that re-competing a new program is a routine decision, and the know-how sits with a limited number of process engineers who can be hired away. Patents are not evidently the barrier: the 10-Q's risk factors list failure to protect intellectual property among the things that could hurt results, which is standard language but also an accurate description of a process business protected by tacit knowledge. The clearest sign that the edge is thin is what it is worth. At December 31, 2025 the company held $1.60 billion of equity, no debt and $0.74 billion of cash, so roughly $0.86 billion of invested capital supported FY2025 revenue of about $0.7 million (estimated by dividing the reported net loss by the reported net margin). That works out to about $1,200 of invested capital for every $1 of annual revenue, a ratio that describes a laboratory, not a franchise.
What breaks it: a competing foundry offering shared multi-project wafer runs at a lower price, or a design that ports cleanly onto a standard silicon photonics process, either of which turns a bespoke relationship into a commodity order.
Demand-side advantage: Weak, because programs renew and customers do not
Demand-side advantage: Weak. Inside a funded program, switching is genuinely painful. A lab that has qualified a QUBT module into an experiment or a defense subcontractor that has written it into a specification faces schedule risk if it changes suppliers mid-cycle, and that inertia explains why the company keeps winning repeat work at all.
The captivity does not travel. Each award is competed on its own, the customer list is short, and a photonic chip does not become more valuable because another lab bought one, so there is no network effect to lean on. The pricing test is the one that decides it: a supplier with captive buyers does not surrender margin as volume arrives, and QUBT's margin compressed over exactly the years its revenue base was building. Export controls on advanced lithography and cryogenic equipment shape who can buy from whom, but QUBT holds no license or quota that shuts a rival out.
What breaks it: a systems leader or cloud provider standardizing on an in-house photonic module and buying the component layer on price, which would leave merchant suppliers bidding against each other for the same handful of programs.
Economies of scale: Weak, because the market caps the minimum efficient scale
Economies of scale: Weak. A scale barrier needs fixed costs that are large relative to the market, so that a new entrant must win a big share before its unit costs match. QUBT's fixed costs are large relative to its own revenue and small relative to the market: merchant TFLN demand is thin enough that a rival reaches comparable unit costs with a few customers. The company's $1.63 billion balance sheet does not change that, because capital is abundant across this sector and does not exclude anyone. Buying scale is the harder version of the same problem, and the balance sheet shows it happening: goodwill rose from $0.06 billion to $0.16 billion in the nine months to September 30, 2026, while cash fell from $0.74 billion to $0.17 billion. Acquisitions can add technology and people; they do not add customers who had nowhere else to go.
Durability: a bet on a modality that has not been chosen
The largest risk is that the niche stops mattering. Superconducting and trapped-ion machines may never need a TFLN component, and near-term demand may be absorbed by classical GPU clusters that are cheaper per useful calculation. If photonics wins, the value could migrate to whoever integrates it into a full machine, which is the position IBM, Google and PsiQuantum are building. A merchant component supplier then competes for a smaller slice of a market defined by someone else's architecture.
So the moat is real but shallow, and it will announce itself in one place if it deepens: gross margin recovering and holding above roughly 30% on rising revenue, which would mean the process know-how is finally being priced. Until that shows up in the statements, the evidence supports a supplier with useful technology and no demonstrated power to charge for it.
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 Quantum Computing Inc 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 Quantum Computing Inc
Quantum Computing Inc
Business, moat and key questions
Open company page