Quantum Computing Cleared $1 Billion in 2025: What the Milestone Proves and What It Doesn't

Global quantum computing revenue passed $1 billion in 2025, while three companies from the sector's first SPAC wave booked about $138 million against the roughly $1.2 billion they had forecast for the same period. The milestone and the miss are both real: one shows that paying customers exist, the other shows that early commercial promises ran well ahead of delivery.

For a general reader that combination looks like a contradiction. It is not. It is the first year in which this industry could be audited, and the audit is more interesting than either headline alone. Below: why the gap opened, what buyers are actually purchasing, which roadmap numbers can be trusted, and a short framework for judging quantum claims.

Why did two opposite quantum headlines appear in the same year?

Because the field crossed from prediction into verification. Technical progress, commercial adoption and market expectations can now be examined on separate grounds, each against a documented track record.

Five years ago that separation was almost impossible. Laboratories could publish results and vendors could publish roadmaps, but commercial history was too thin to test either claim. Some of those projections have now reached their due dates, which is why one year produced both an industry-wide first and a very public shortfall.

Independent trackers describe the same inflection. McKinsey's Quantum Technology Monitor put quantum computing firms' combined revenue between $650 million and $750 million in 2024, above the billion mark only the following year, with roughly $4.4 billion projected for 2028 and a $28-72 billion annual range sketched for 2035. Those figures describe the industry curve; the shortfall described above describes the distance between individual vendor promises and that curve. Different questions, different evidence.

How far did early forecasts miss, and what caused the gap?

The arithmetic is blunt: Roughly 12 cents of every forecast dollar arrived as revenue. An uncomfortable outcome, but in our reading it is not proof that quantum computing has failed. It is proof that commercial expectations were set long before the market had any way to discipline them.

A forecast can miss for several distinct reasons, and readers should ask which one applies:

• customers adopted more slowly than the sales plan assumed;
• a hardware milestone arrived later than promised;
• the cost of scaling rose faster than budgeted;
• an application that appeared commercially close actually needs a further system generation.

Engineers already judge their own work this way: a design target is a hypothesis until the built system confirms it. Public quantum firms have now reached the same test, where investors and customers can set promise beside delivery.

One structural cause is worth naming. Going public through a SPAC rewarded narrative before revenue, and it inflated headline value without guaranteeing that operating cash reached the company. That is a governance lesson with a much wider audience than quantum.

What is actually inside the first billion-dollar year?

Read closely, the revenue side is the most encouraging part of the story. Customers are paying for hardware, cloud access, software and services at a time when no broadly useful fault-tolerant machine exists.

Yet much of that spending remains exploratory. Organisations buy access in order to learn what the technology can do, to build internal skills and to pilot workloads. That activity is genuinely commercial. It does not demonstrate that a quantum computer outperforms the strongest classical option on a workload with economic value.

The mix of demand is shifting, and this is where reporting beyond the milestone helps. The Wall Street Journal reported that enterprise users spent about $300 million on quantum computing in 2025, outranking research laboratories and governments as the leading buyer group for the first time. McKinsey's later monitor counts more than 300 companies worldwide engaging with quantum computing and notes that around 72 percent of quantum computing use now sits in companies majority-owned by private entities, a swing away from the public-research era. IBM's Institute for Business Value, surveying senior executives, placed quantum at roughly 11 percent of research and development budgets on average, up from 7 percent in 2023.

So the next question is not how big the revenue number is. It is narrower: what exactly customers buy, whether they buy again, and whether spending migrates from exploration toward applications with a measured advantage.

Signal to inspect	What it proves	What it hides
Total annual revenue	Buying is now normal across hardware, cloud, software and services	Whether the money recurs or was one-off exploration budget
Share coming from enterprises	Commercial demand rather than grant funding	Whether any workload beat a classical baseline
Forecast versus delivery	Management's ability to estimate its own market	Nothing about technical quality - the two are separate questions
Logical qubit totals	Movement toward error-corrected machines	Error models, test conditions and post-selection behind each result
Headline valuation versus cash	Access to capital markets	Whether several more hardware generations can be funded

Which roadmap numbers can legitimately be compared?

Physical qubit counts made roadmaps easy to draw and hard to falsify. The conversation has moved toward logical qubits and fault-tolerant systems, which is an improvement - but a logical-qubit total is not automatically comparable from one result to the next.

An error-detected result differs from an error-corrected result. Some demonstrations lean on post-selection. Architectures carry different error models, and experiments run under different conditions. The disciplined reading starts from what has been demonstrated, then the conditions of that demonstration, then the engineering still needed for the next milestone.

Recent work illustrates the difference. On 9 December 2024, Google Quantum AI reported in Nature that surface codes operated below threshold on its Willow processors; the authors Michael Newman and Kevin Satzinger described it as the first demonstration of error suppression improving exponentially as the code grows. On overhead, IBM's published hardware roadmap argues that qLDPC codes can cut physical-qubit overhead by as much as 90 percent, with the fault-tolerant Starling system scheduled for 2029 and specified as roughly 200 logical qubits running 100 million gates. Encouraging - and still architecture-specific, which is exactly the caution required.

Larger logical systems are expected later this decade. For credibility, the gap between what has been demonstrated and what has been promised is now the single most useful measure available.

Why does cash, rather than valuation, decide which plans survive?

For hardware companies, financing and engineering are one problem rather than two. Scaling demands specialist teams, fabrication, packaging, control electronics, cryogenic or optical infrastructure, testing and repeated hardware iterations - each round absorbing budget before a customer invoice appears.

That is why cash raised deserves more attention than headline value. A SPAC deal can display a large valuation while far less money reaches the business, because shareholders may redeem before the merger closes. Where multiple hardware generations still have to be built, a roadmap can be technically plausible and still impossible to finance. At that point financial due diligence becomes part of technical due diligence.

Capital supply has not been the binding constraint so far. McKinsey's 2026 edition of the monitor reports investment in quantum technology start-ups of $12.6 billion in 2025, around 6.3 times the 2024 level. Abundant money buys time; it does not shorten the engineering path, and it makes the revenue-versus-promise test above the figure worth watching.

Is there spending that grows even without quantum advantage?

This is the most under-covered part of the story, and it changes how a non-specialist should read "quantum is late". Security budgets are moving on a calendar rather than waiting for a physics breakthrough. NIST finalised its first three post-quantum cryptography standards in August 2024; US National Security Memorandum 10 set 2035 as the primary target for completing migration across federal systems, and NIST's transition guidance proposes deprecating quantum-vulnerable public-key cryptography after 2030 and disallowing it under NIST standards after 2035.

Practical reading: a substantial slice of quantum-adjacent spending is defensive and deadline-driven. It requires no machine that beats a classical one, only a credible expectation that such a machine will eventually exist. That is a real market, but it is not the same evidence base as computational advantage, and the two should never be counted as the same dollar.

When will quantum computing be broadly useful?

Treat any confident date with suspicion. The field has no reliable calendar, and the threshold for useful advantage depends heavily on the application: chemistry, materials, optimisation and other workloads may reach it at quite different system sizes.

The comparison point keeps moving. GPUs continue to improve and specialised classical approaches advance, so a workload a roadmap treats as "solved once the machine is big enough" may become harder in the interim.

What has changed is the evidence base. Technical demonstrations can now be measured against roadmaps, customer spending can be examined, financing histories reveal whether development plans are funded, and old forecasts can be checked against reality. The judging standard has moved accordingly: the question is no longer whether a firm can publish an ambitious roadmap, but what has been demonstrated, what customers are paying for, and which engineering steps still separate the two.

Common misconceptions about the $1 billion headline

• "A billion in revenue means the technology works." It means somebody paid. A large share of that payment buys access, pilots and learning.
• "The missed forecasts prove the field is a fraud." They show promises were made before buyers, milestones and scaling costs were measurable.
• "More qubits means a better machine." Comparability depends on error handling, definitions and conditions, not on the count.
• "A big valuation means a funded roadmap." Redemptions and actual cash receipts decide whether the plan can be built.
• "If hardware runs late, spending stops." Deadline-driven cryptography work keeps a separate budget line moving.

A six-step framework for judging any quantum claim

1. Ask for the demonstrated result and its date, not the roadmap target.
2. Ask whether the result is error-detected, error-corrected, or partly post-selected.
3. Ask which classical baseline the workload beat, and who measured it.
4. Split revenue into exploratory and recurring; a single invoice is not a market.
5. Compare cash actually received by the company with the headline valuation.
6. Test the claim against a deadline you control - the 2030 and 2035 cryptography migration dates are examples - rather than against the hype cycle.

Readers without a technical background will get the most value from steps two and three: that is where confident-sounding numbers conceal the real uncertainty.

Frequently asked questions

Does crossing $1 billion mean quantum computing is ready?

No. It means customers are willing to pay during the build-out. Readiness would require reproducible advantage on a named workload, which the record does not yet show at broad scale.

Why did the three companies miss their own forecast by so much?

Their published targets assumed faster adoption, earlier hardware milestones and lower scaling costs than the market delivered. Roughly 12 cents of each forecast dollar is a measurement of optimism, not of technical merit.

Should a normal business buy quantum access now?

A small, time-boxed pilot and an internal owner are reasonable. A multi-year strategic commitment is not, unless the company can name the workload and the classical comparison it intends to beat.

What should I do even if quantum hardware stays late?

Start post-quantum cryptography planning. Inventory which systems still depend on today's public-key algorithms, and set a migration schedule aligned to the 2030 deprecation and 2035 completion milestones published by NIST.

Which single number matters most next year?

Recurring enterprise revenue, plus documented advantage on a specific workload. Those two together, rather than any total market-size projection, decide whether the billion-dollar year was a beginning or a peak of expectations.

Sources cited

• McKinsey & Company, The Year of Quantum: From concept to reality in 2025 - https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/the-year-of-quantum-from-concept-to-reality-in-2025
• McKinsey & Company, Quantum Technology Monitor 2026: A commercial tipping point - https://www.mckinsey.com/capabilities/mckinsey-technology/our-insights/mckinsey-quantum-technology-monitor-2026-a-commercial-tipping-point
• The Wall Street Journal, Businesses Are Spending Big on Quantum - https://www.wsj.com/tech/businesses-are-spending-big-on-quantum-c73060ad
• Nature, Quantum error correction below the surface code threshold (9 December 2024) - https://www.nature.com/articles/s41586-024-08449-y
• IBM Quantum hardware roadmap - https://www.ibm.com/roadmaps/quantum/
• NIST, post-quantum cryptography project and transition guidance - https://csrc.nist.gov/projects/post-quantum-cryptography
• IBM Institute for Business Value, quantum computing readiness research - https://www.ibm.com/thought-leadership/institute-business-value/en-us/report/2025-quantum-computing-readiness

Analysis by James Hartley for ItInsightsWorld. Figures are drawn from publicly reported sources; nothing here constitutes investment advice.