Quantum Computing Market

IonQ closed its purchase of SkyWater Technology on 31 July 2026, a move its second-quarter 2026 results filing calls the first vertically integrated, full-stack quantum platform, and it marks the year a qubit maker bought its own foundry. The quantum computing market covers paid access to quantum processors, whether as on-premises systems, cloud time, software or integration work, sold to labs, governments and companies. Douglas Insights sizes it at USD 1.38 billion in 2025, built as 1,960 paying accounts at an average annual spend of USD 702,400, and expects USD 12.8 billion by 2035 at a revenue CAGR of 25.01% for 2026 to 2035. The study belongs to our Information Technology and Semiconductors coverage, and every step of the build follows our research methodology.

Which companies win quantum computing contracts, and what is each position built on?

IonQ is the revenue leader among disclosed quantum computing vendors, with USD 130.0 million of 2025 revenue, a 9.4% share of the USD 1.38 billion market on Douglas Insights arithmetic. IBM, Quantinuum, Google Quantum AI, D-Wave and Rigetti follow, each built on a different qubit platform and sales route.

IonQ reported Q2 2026 revenue of USD 80.1 million, up 287% on the year, and raised its 2026 guidance to USD 280 million to USD 290 million, according to its August 2026 filing. That guidance excludes SkyWater. IonQ sells trapped ion systems, cloud time and networking, and it also bought Nexus Photonics for integrated photonics. Quantinuum, the other trapped ion supplier, reported Q2 2026 revenue of USD 8.0 million, up 279%, guided 2026 to USD 28 million to USD 32 million and raised USD 1.7 billion of gross proceeds in its initial public offering (IPO), according to its Q2 2026 earnings release. Its Nexus software platform now counts 180 organisations, and Oracle will offer its Helios machine as an Oracle Cloud Infrastructure service.

IBM builds superconducting processors and sells them through its cloud and on-premises deals. Its 120-qubit Nighthawk chip carries 218 tunable couplers, and IBM targets the first large-scale, fault-tolerant quantum computer by 2029, according to its 12 November 2025 announcement. Google Quantum AI holds a technical rather than a commercial position: its 105-qubit Willow chip finished a random circuit sampling benchmark in under five minutes, but Google books no separate quantum revenue. Science, not sales.

D-Wave sells quantum annealing systems and cloud time. Its 2025 revenue was USD 24.6 million, and first-half 2026 bookings reached USD 35.5 million, including a USD 20 million system sale, per its Q2 2026 results. Rigetti, a superconducting rival, booked USD 5.1 million of Q2 2026 revenue and signed a letter of intent with the U.S. Department of Commerce for up to USD 100 million over three years, per its Q2 2026 release.

Douglas Insights estimates the top-three concentration at 29.6% of 2025 quantum computing revenue: IonQ on its disclosed USD 130.0 million, IBM and Quantinuum on our modelled figures, since IBM does not report quantum revenue. The rest splits across dozens of national and start-up suppliers.

Which offering makes the money in quantum computing: cloud time, on-premises systems or software?

Cloud access and quantum computing as a service makes the most money, with 38.6% of 2025 revenue, or USD 531.4 million. Most buyers still test algorithms before they commit to a machine, so metered time on shared quantum processors outsells hardware across the USD 1.38 billion total.

Cloud access and quantum computing as a service holds 38.6% because a pilot costs a fraction of a machine and the leading processors sit behind vendor and hyperscale clouds; it grows fastest at 26.9% a year, reaching USD 5.75 billion by 2035, as the Oracle and HPE tie-ups pull quantum time into existing compute contracts. On-premises quantum systems take 33.2%, or USD 457.1 million, because national labs and supercomputing centres want sovereign machines; they grow 22.7% a year to USD 3.54 billion, slower because each deal is lumpy. Software and algorithms hold 15.7%, or USD 216.1 million, and grow 26.6% as error suppression and workflow tools are licensed per seat. Consulting and integration services take 12.5%, or USD 172.1 million, the slowest leg at 22.1%, since buyers move work in-house once teams are trained.

Which qubit technology holds the largest quantum computing revenue pool in 2025?

Superconducting machines hold an estimated 41.3% of 2025 quantum computing revenue, ahead of trapped ion at 29.7%. IBM, Google and Rigetti anchor the superconducting pool, while IonQ and Quantinuum carry trapped ion, on Douglas Insights splits of the USD 1.38 billion total.

Neutral atom systems take 12.9%, photonic machines 9.3% and quantum annealing 6.8% of the quantum computing total. Superconducting chips lead because they ride standard lithography; IBM moved wafer fabrication to a 300mm line in Albany. Trapped ion gains because its qubits show high fidelity, and the trapped ion vendors grew revenue by 287% and 279% in Q2 2026, per the IonQ filing and the Quantinuum release. Quantum annealing stays small but sells whole machines, as the USD 20 million D-Wave order shows. Photonic and neutral atom suppliers sell mostly research access today, which is why their combined 22.2% outruns their installed base.

Why are governments and drug makers signing larger quantum computing contracts?

Government procurement, chemistry research and finance pilots lift paying quantum computing accounts by 17.6% a year, the volume leg of the 25.01% revenue CAGR. Douglas Insights splits that leg across four drivers, with state buyers the largest at 6.8 points.

Public money and sovereign machines

National programmes add 6.8 points to quantum computing account growth. Rigetti signed a letter of intent for up to USD 100 million of CHIPS research funding over three years and is delivering a 108-qubit system to India’s C-DAC, per its Q2 2026 release. Quantinuum signed its own Department of Commerce letter of intent. Each sovereign machine pulls in universities and agencies as paying users around it. A 9-qubit Novera system bound for the Pittsburgh Supercomputing Center TangleLab, funded by a National Science Foundation grant, shows how one public award seeds a cluster of research accounts.

Chemistry and pharmaceutical simulation

Drug and materials research adds 4.9 points. Quantinuum, NVIDIA and a Fortune 100 drug maker showed AI-driven quantum simulation for molecular property work, and D-Wave signed or renewed work with Shionogi. Pharmaceuticals and chemicals buyers pay for repeat runs, not one-off demonstrations, so each new account carries an above-average spend of roughly USD 0.9 million a year in our model.

Finance and telecom optimisation

Banking and financial services and telecom add 3.4 points. D-Wave lists AT&T, Nasdaq Verafin and Unisys among new or renewing customers in its Q2 2026 results, and its first-half 2026 bookings rose 1,120% to USD 35.5 million. Optimisation problems such as fraud screening and network routing map well to annealers and hybrid solvers, so these quantum computing accounts convert from pilot to contract faster than chemistry users. D-Wave also names one of the world’s largest gambling and entertainment companies among its new clients, a sign that optimisation demand now reaches beyond banks.

Hybrid quantum and supercomputer stacks

Integration with classical high-performance computing adds the last 2.5 points. HPE works with Rigetti and the Pittsburgh Supercomputing Center on a hybrid quantum-classical supercomputer, and with Quantinuum on combining quantum, HPC and AI. Buyers already running GPU clusters for the Artificial Intelligence Applications Market add quantum time to the same job queue. Oracle will host Quantinuum Helios inside its cloud, and 180 organisations already use Quantinuum Nexus to submit jobs. So 6.8 plus 4.9 plus 3.4 plus 2.5 points gives the 17.6% volume leg.

What slows quantum computing adoption while logical qubits stay scarce?

Three headwinds remove 4.2 points a year from quantum computing account growth that would otherwise run near 21.8%. Error rates take the most, 2.1 points, because few production workloads beat classical computers before fault-tolerant machines arrive around 2029. Supplier losses, skills and export friction take the rest.

Error correction is the first brake. Timing matters most. IBM itself expects the first verified quantum advantage cases only by the end of 2026 and fault tolerance by 2029, so many quantum computing pilots stop at proof of concept. Douglas Insights removes 2.1 points of account growth for this delay. Revenue is also lumpy: D-Wave’s first-half 2026 revenue fell 67% to USD 5.9 million because the year-earlier half held a USD 13.7 million system sale, per its filing.

Skills and losses take the other 2.1 points. Buyers notice. Vendors spend far ahead of sales: D-Wave lost USD 48.0 million in Q2 2026 on USD 3.1 million of revenue, per its results, so buyers worry about supplier survival and sign short quantum computing contracts. We take 1.3 points for scarce in-house quantum talent and 0.8 points for export licensing friction on cross-border system sales.

Where do quantum computing buyers sit, and which region grows fastest?

North America leads quantum computing with USD 651.2 million, or 47.3% of 2025 revenue, because IonQ, IBM, Google and Rigetti sell first to U.S. agencies and banks. Asia Pacific grows fastest at 26.4% a year to 2035, as India and Japan fund sovereign and corporate machines.

North America reaches USD 5.87 billion by 2035 at 24.6% a year. Europe holds USD 341.4 million, or 24.8%, and rises at 24.3% to USD 3.01 billion, helped by sovereign machines such as the Advantage2 capacity a buyer in Lombardy, Italy, took for EUR 10 million. Asia Pacific starts at USD 311.1 million, or 22.6%, and climbs to USD 3.24 billion: India’s C-DAC order and Japanese research buyers such as Shionogi and Oki show state and corporate money arriving together. Latin America is small at USD 28.9 million and grows 25.1% to USD 271.3 million. The Middle East and Africa is the wildcard at USD 44.1 million, rising 25.9% to USD 441.1 million, because Gulf sovereign funds buy whole machines when they buy at all.

How much does a quantum computer or a year of quantum computing access cost?

Whole quantum computing systems sell for about USD 2.9 million to USD 20 million in disclosed deals, while the average paying account spends USD 702,400 a year in 2025. Price per account rises 6.3% a year as buyers move from pilots to production contracts.

At the low end, Rigetti’s two Novera on-premises systems carried about USD 5.7 million of purchase orders together, near USD 2.9 million each, and its 108-qubit C-DAC system about USD 8.4 million. At the top, D-Wave reported a USD 20 million system sale in its first-half 2026 bookings, and a European client paid EUR 10 million for half the capacity of one annealer. Cloud quantum computing is billed by time or by shot; Douglas Insights puts the 2026 average at USD 746,651 per account, rising to about USD 1.29 million by 2035.

Which use cases turn quantum computing pilots into repeat contracts?

Optimisation carries an estimated 36.4% of 2025 quantum computing use-case revenue, ahead of chemistry and materials simulation at 31.8%. Both have customers who renew; quantum machine learning and cryptography research remain mostly funded exploration, paid from research budgets rather than operating lines.

Quantum machine learning takes 19.6% and cryptography research 12.2% of the quantum computing use-case pool on our split. Optimisation renews. It wins because a hybrid solver returns a usable schedule today. Chemistry and materials simulation renews because drug makers repeat runs across molecule libraries. Cryptography research stays a niche of buyers testing the threat to public-key systems, a topic covered in the Cybersecurity Mesh Market.

How did the SkyWater and Quantum Circuits deals reshape quantum computing supply?

Two 2026 acquisitions moved quantum computing toward vertical integration: IonQ bought foundry SkyWater, and D-Wave bought Quantum Circuits, spending USD 252.8 million of net cash on acquisitions in the first half. Owned fabs and gate-model teams now sit inside the two largest listed suppliers.

D-Wave completed the Quantum Circuits purchase in January 2026, adding gate-model superconducting work to its annealers, per its Q2 2026 filing, and booked USD 9.3 million of one-off deal costs. IonQ’s SkyWater deal gives a quantum computing vendor its own wafer fab, much as IBM moved to the Albany 300mm line. Margins come first. Douglas Insights expects owned fabrication to lift supplier gross margin before it lifts price, which is why our price leg stays at 6.3%. Readers tracking brain-inspired chips will find a parallel in the Neuromorphic Chip Market.

What if fault-tolerant quantum computing arrives on schedule in 2029, or slips?

The base case puts quantum computing at USD 12.8 billion in 2035, with 17.6% account growth and 6.3% price growth. A slower path ends near USD 7.81 billion, and a faster one near USD 19.2 billion, depending mainly on when error-corrected machines reach paying buyers.

The slower case assumes fault tolerance slips past 2032: accounts grow 13.4% and price 4.9%, for 18.96% a year. The faster case assumes IBM hits its 2029 target and verified advantage lands by the end of 2026: accounts grow 21.2% and price 7.4%, for 30.17% a year. Adding one point to account growth adds about USD 1.13 billion to the 2035 quantum computing value. Rival published forecasts run between 22.3% and 41.8% a year; our 25.01% sits in the lower half because we count paid access only, not national research budgets. IonQ’s closing of SkyWater on 31 July 2026 is one marker for the faster case, as its filing sets out a full-stack model.

Which export rules and post-quantum cryptography standards shape quantum computing sales?

Two rule sets shape quantum computing sales: U.S. export controls on quantum items from September 2024, and three NIST post-quantum encryption standards. We remove 0.8 points of account growth for the licensing friction, while the encryption standards add research demand from banks and agencies.

The Bureau of Industry and Security published an interim final rule on 5 September 2024 that controls quantum computers and related equipment, components, materials, software and technology, with a License Exception Implemented Export Controls (IEC) for countries running equivalent controls, according to the BIS announcement. NIST released FIPS 203, FIPS 204 and FIPS 205 in August 2024, which pushes banks to fund cryptography research on quantum computing time. Public funding also carries rules: Rigetti’s CHIPS letter of intent contemplates a federal equity stake in line with the up to USD 100 million award.

Douglas Exclusive: the quantum computing milestone calendar

The quantum computing milestone calendar is a Douglas Insights model built from 11 sourced inputs taken from five primary releases by IBM, IonQ, Quantinuum, D-Wave and Rigetti. It is our own compilation, not an official register, and it dates each promised milestone that would move revenue.

The finding: five of the 11 quantum computing milestones fall due in 2026, and the two revenue guides alone add up to USD 308 million to USD 322 million. Against our 2026 estimate of USD 1.72 billion, that is 17.9% to 18.7% of the market from two suppliers. If both guides land, our base case holds; a miss by IonQ alone would pull the 2026 figure toward the slower path.

What methodology turns 1,960 paying quantum computing accounts into USD 1.38 billion?

The quantum computing model multiplies 1,960 paying accounts by an average annual spend of USD 702,400 to reach USD 1.38 billion in 2025. It draws on 46 data points across 21 countries and checks the result against disclosed vendor revenue and published estimates.

Accounts grow 17.6% a year to about 9,916 in 2035, and spend per account grows 6.3%, which compounds to 25.01% revenue growth and USD 12.8 billion. The three listed pure plays with full-year 2025 revenue, IonQ at USD 130.0 million, D-Wave at USD 24.6 million and Rigetti at USD 7.1 million, total USD 161.7 million, or 11.7% of our figure. Our 2025 base sits 1.0% below the lowest published base-year estimate we read. Regions sum to the global figure in 2025 and 2035, and offering segments reconcile within 1%. Douglas Insights analysts drafted this study with AI assistance and checked every sourced figure against the linked primary pages.

Sources

1. IonQ (SEC Form 8-K) IonQ Q2 2026 results (2026)
2. Quantinuum (SEC Form 8-K) Quantinuum Q2 2026 results (2026)
3. D-Wave (SEC Form 8-K) D-Wave Q2 2026 results (2026)
4. Rigetti (SEC Form 8-K) Rigetti Q2 2026 results (2026)
5. IBM IBM Nighthawk and Loon announcement (2025)
6. U.S. Bureau of Industry and Security Quantum export controls (2024)
7. NIST Post-quantum encryption standards (2024)