Report contents
Industry landscape
Where the company competes: a three-player superconducting niche funded by governments
Rigetti designs and builds superconducting quantum processors: chips cooled to near absolute zero that execute gate-based circuits. It sells systems to research buyers, sells time on its machines through cloud platforms, and takes government development contracts, including a $100 million funding agreement with the US Department of Commerce that management described on the November 10, 2026 earnings call. That mix sets the arena. The company is not selling a general-purpose computer yet; it sells access to a physics program that a small set of buyers pays to advance.
The margin today sits upstream and in procurement
Upstream of Rigetti sit dilution refrigerators, microwave control electronics, superconducting thin-film fabrication, and the thin pool of physicists who can run error-mitigation work. Rigetti fabricates its own multi-chip modules, which most of its startup peers do not, so it owns the fabrication step instead of paying an outside foundry. Downstream, two channels carry the product: direct sale or lease to national laboratories and government programs, and cloud marketplaces such as Amazon Braket and Microsoft Azure Quantum, where the platform owner sets the access price and keeps the customer relationship. Suppliers and cloud platforms collect margin with far less technical risk than the system builder carries. Rigetti's gross margin reached about 61% in the third quarter of 2026 from 21% a year earlier, per the same call, but on a contract-driven revenue base that swings with delivery timing, mix explains that jump better than pricing power does.
The submarket: 100-qubit superconducting gate-model systems
The precise boundary is gate-model superconducting processors at the 100-qubit class, sold to government labs, universities and cloud users who rent time by the shot. That excludes annealing machines built for optimization, trapped-ion systems and photonic designs, each with its own physics and its own buyers. Competition is a national procurement game, not a global consumer market: customers cluster around US federal agencies and a few allied governments in the UK, Japan and Germany, while suppliers compete worldwide for those budgets. Buyers choose on two-qubit gate fidelity, qubit count, uptime and queue depth, not on price per minute, and a single order can move a quarter.
The competitive set:
- IBM and Google Quantum AI run superconducting gate-model programs at larger qubit counts, with their own cloud distribution.
- Rigetti is the third superconducting gate-model vendor and differentiates on multi-chip modules, which link smaller dies into one processor; its 108-qubit Cepheus-1-108Q reported 99.5% median two-qubit gate fidelity on the third-quarter call.
- IonQ and Quantinuum build trapped-ion machines and reach customers through their own clouds plus AWS and Azure.
- D-Wave sells annealing systems for optimization, a different buyer and a different use case.
- PsiQuantum pursues photonic quantum computing and has focused on government-backed, utility-scale builds.
Published market shares do not exist for a field this young, with revenue concentrated in a few contracts, so structure has to be read from capability, funding and who controls distribution.
Barriers are high for a startup and thin for an incumbent
Judged by the practical test (if the leaders cannot be counted on one hand, entry barriers are usually weak), this market splits. Across quantum hardware broadly, more than half a dozen funded programs are shipping or say they soon will, too many for durable pricing power. Inside Rigetti's own slice, three players matter, and that is a genuinely narrow field. Reaching 100 qubits at competitive fidelity takes a fabrication line, a cryogenics lab and a standing team of specialists; Rigetti alone ran $34 million of operating expense in the third quarter of 2026, most of it R&D.
The harder structural problem is who the buyers are. Federal agencies fund several vendors at once, deliberately, to keep options open, which caps any single supplier's leverage over price. And the deepest-pocketed potential customers, IBM, Google, Microsoft and Amazon, each run their own hardware or cloud programs. A market where the customer can also be the competitor is a difficult place to hold a premium, and it means Rigetti's pricing power depends less on its fidelity lead than on whether governments keep paying several suppliers to stay in the race.
Competitive advantage
Does the company have a moat? Weak, because the difficulty is technical, not commercial
Rigetti has cleared a real barrier and kept almost none of the profit that barrier should protect. The company fabricates its own multi-chip superconducting processors, a step most quantum startups rent from a foundry, and it competes in a slice of the market where only three gate-model superconducting vendors matter. Yet revenue has been reported at roughly $10 million in every fiscal year from 2021 through 2025, and the operating loss has grown from $30 million in fiscal 2021 to $80 million in fiscal 2025 and about $110 million over the last four quarters. The barrier raises what it costs to compete. It has not raised what Rigetti can charge.
Concentration here runs the wrong way for the seller. Supply is narrow, with three superconducting gate-model vendors, but demand is narrower: a handful of government program offices and a few cloud marketplaces, and those buyers deliberately spread work across suppliers to keep alternatives alive. When the customer can also build the product and funds several vendors at once, a thin field of suppliers does not produce pricing power. The proof is in the returns, not the technology.
Market position: third of three superconducting gate-model vendors
Market position: Moderate. Rigetti ranks behind IBM and Google Quantum AI on every visible measure in superconducting gate-model systems: qubit count, published two-qubit fidelity, and the cloud relationships that carry machines to users. It is not a distant also-ran on engineering; its 108-qubit Cepheus-1-108Q reported 99.5% median two-qubit gate fidelity, which puts it in the same technical conversation as far larger programs. Rank has to be read from capability and revenue, though, because revenue of roughly $10 million a year is a rounding error against what IBM and Google spend on the same problem.
Market position stability: a held rank with no measured share
Market position stability: Moderate. No published share series covers gate-model superconducting systems, so a five-year share trend cannot be confirmed and a table would be false precision. What can be said is that Rigetti was one of a small group of superconducting vendors in 2021 and is still there in 2026, neither dropped out nor acquired. If spending on quantum hardware rose over that stretch, as most industry forecasts assume, then flat revenue implies Rigetti captured very little of the increase. The ranking has held; the share is unproven.
Profitability: five straight years of negative returns
Core returns: Weak. Return on assets was negative in every fiscal year from 2021 through 2025, in a range from -41% to -90%. An estimate of after-tax ROIC points the same way: a trailing four-quarter operating loss of about $110 million, no meaningful tax charge, and invested capital of roughly $490 million (equity of $540 million less $50 million of cash, with no debt) give about -22%. GuruFocus calculates trailing ROIC at -39.48% and the September 2026 quarter at an annualized -44.65%; the gap comes from how each version measures invested capital, not from disagreement about the sign. ROE is useless here, because equity is small relative to the losses and net income swings on non-operating items: in the March 2026 quarter Rigetti reported $30 million of net income against a $30 million operating loss.
Margins and scale: Weak. Gross margin ran 23.5% in fiscal 2021, 78.1% in 2022, 76.7% in 2023, 52.8% in 2024, then -86.1% in 2025 and 60.7% in the September 2026 quarter. A company with pricing power does not post a negative gross margin in any year, and a swing that wide on roughly unchanged revenue is contract mix, not price. Fiscal 2025's net margin of -3,050% is arithmetically real and analytically empty: a $220 million net loss over about $10 million of revenue, much of the loss coming from revaluing warrants and investments rather than from selling machines. A like-for-like margin comparison with IonQ or D-Wave is not possible, because each company mixes hardware, cloud access and research contracts in different proportions.
Overall rating: Weak
Overall rating: Weak. The barriers are genuine: a 100-qubit superconducting program needs a fabrication line, a dilution refrigerator and a standing team of specialists, and only a few organizations have assembled all three. That is why the rating is not None. But a moat has to show up as returns, and Rigetti's returns have been negative in every year on record without converging quickly. The rating sits below the Moderate market position because a third-place rank in a subsidized field protects nothing until someone pays above cost.
Where the advantage comes from and how long it can last: a fabrication capability no customer has paid for yet
Three things could plausibly protect Rigetti: its own superconducting fabrication, a hold on the customers who run circuits on its machines, and the fixed-cost scale of the hardware business. In a sub-industry whose buyers are funding agencies, only the first is a genuine barrier, and it is the weakest kind. The metric that tests all three is how much paying demand exists per dollar of fixed cost. Rigetti's trailing revenue of about $10 million covers roughly 8 cents of each dollar of operating expense, an estimate built from about $0.01 billion of revenue and a $0.11 billion operating loss, which implies roughly $0.12 billion of operating cost. At the September 2026 quarter's gross margin, revenue would need to reach about $200 million a year to break even, more than 20 times the current level. Until that ratio moves, arithmetic settles the moat question.
Supply side: an in-house fab is a capability, not a cost edge
Supply-side advantage: Moderate. Rigetti builds its own multi-chip modules, linking smaller superconducting dies into one processor. That matters technically. It shortens the loop between designing a chip and holding a packaged part in a refrigerator, and it lets the company chase qubit count by adding dies instead of redesigning a monolithic chip, while a fabless rival queues for a foundry and takes the process it is offered. The protection is real but narrow, and it has not turned into money: gross profit ran a few million dollars in the September 2026 quarter against an operating loss near $30 million, so the fab adds to the fixed cost base before it adds to margin. IBM runs its own fabrication at far greater scale and spreads the same cost across a cloud business, and a competitor buying dies from a foundry can match fidelity without owning the line. What would break it is another modality reaching error-corrected operation first, or a foundry-based superconducting program matching Rigetti's two-qubit fidelity at a lower cost per qubit. Engineering leads in hardware typically last a few product cycles, not a decade.
Demand side: government buyers fund several vendors by design
Demand-side advantage: Weak. The captivity test is whether a rival's discount takes the customer. Here price is not how buyers choose, which sounds encouraging until you look at who chooses: program offices that pick on two-qubit fidelity, uptime and queue depth, and that carry a mandate to keep more than one supplier alive. The $100 million agreement with the Department of Commerce shows both halves of that. It validates the technology, and it is a development award rather than a purchase a customer chose to make twice. It is worth roughly one year of current cash consumption, against $90 million of free cash outflow in the first nine months of fiscal 2026. Captivity would appear as a commercial buyer signing a second, larger contract at a higher price, with access revenue growing without a new government program behind it. So far the evidence only shows the reverse: revenue flat near $10 million for five years while the installed base of government-funded machines grew. A funding agency narrowing to fewer vendors, or a cloud partner routing users to its own superconducting hardware, would break the demand side outright.
Scale: too small to spread the fixed cost
Scale advantage: Weak. Research and development, cryogenics and fabrication set a minimum efficient scale far above Rigetti's revenue, and the companies that already carry those fixed costs are also the distribution channel. Scale protects a hardware maker only when it is paired with captivity, because without it a well-funded entrant can buy its way to the same qubit count. Here the entrants are incumbents with cloud revenue to fund the effort. The balance sheet, not scale, buys Rigetti time: equity of $540 million, no debt, and current assets of $410 million against $80 million of current liabilities at September 30, 2026 leave net current assets of about $330 million, or roughly three years at the present burn before any new financing.
Durability: the substitute risk sits outside superconducting circuits
The threat that matters is architectural. Trapped-ion programs at IonQ and Quantinuum, neutral-atom designs, and PsiQuantum's photonic approach all pursue the same customer need by different physics, and the first to reach fault-tolerant operation makes superconducting machines into laboratory instruments. A slower risk runs the other way: if error correction stays a decade out, the government pool that pays Rigetti may shrink before the technology pays it back. Neither is within management's control, and Rigetti is not the best-funded bet on its own architecture.
Two observable questions settle the durability case. Does a buyer that is not a government program sign a second contract at a higher price for access to a Rigetti machine? And does Rigetti reach the error rates that make its multi-chip approach useful before a rival architecture gets there? A clear yes on the first would be the first real evidence of customer captivity; a yes on the second keeps the fabrication capability valuable, and a no turns it into a historical footnote.
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 Rigetti 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 Rigetti Computing Inc
Rigetti Computing Inc
Business, moat and key questions
Open company page