Quantum Computing: what is getting real adoption now?

What should we actually call real quantum computing adoption?

Real quantum computing adoption today means companies are paying for the technology and using it repeatedly inside a serious workflow, with production use sitting at the strongest end of that spectrum.

That distinction changes the whole article. A company signing a quantum partnership gives us weak evidence. A research team paying for cloud access and running workloads every week tells us more. A company embedding a quantum-assisted solver into an operational process gives us much stronger evidence again.

We also need to separate adoption from quantum advantage. Quantum advantage asks whether a quantum system can outperform the best practical classical alternative on speed, accuracy, cost or another useful measure. Adoption asks whether people are actually buying and using the technology.

The two have started moving at different speeds.

McKinsey's latest Quantum Technology Monitor identified more than 300 organizations actively working with quantum computing. Among the large companies it examined in more detail, one-third spent more than $10 million on quantum initiatives during 2025, while 7% spent more than $50 million.

Those companies have clearly started spending real money. Much of that money still goes into application development, integration and internal expertise rather than proven quantum-powered production.

For this article, that gives us a useful scale.

What is happening?	How we should describe it
Partnership, paper or one-off demo	Exploration
Paid access and repeated experiments	Early adoption
Quantum used inside a real workflow	Operational adoption
Repeated production use with measured business impact	Real production adoption
Better result than the best practical classical alternative	Quantum advantage

If you want more recent data on this point, please see our latest quantum computing market report.

Why is quantum computing suddenly being talked about like a real business?

Quantum computing looks much more commercial now because enterprise budgets, provider revenue and actual deployment are all growing at the same time.

The spending shift is probably the strongest evidence. According to McKinsey's 2026 study, 72% of the quantum-computing activity it examined now came from majority-private organizations. Universities, national laboratories and government programs used to dominate the field much more heavily.

Money is also starting to flow through actual quantum companies. McKinsey estimates that quantum-computing providers collectively generated more than $1 billion of revenue during 2025.

IonQ gives us the clearest recent example of how quickly that commercial layer is expanding. The company reported $80.1 million of revenue in the second quarter of 2026, up 287% from a year earlier. About 60% came from commercial customers, according to IonQ, while the company said Tempo system deployments and cloud usage were important drivers.

D-Wave looks very different financially but gives us something IonQ's headline revenue cannot: a breakdown of production usage. In the first half of 2026, 37.3% of D-Wave's quantum-computing-as-a-service revenue came from production applications, compared with 9.8% one year earlier.

Quantinuum adds another useful angle. Its latest quarterly results showed revenue rising 279% year over year to $8 million, while 180 organizations were using its Nexus cloud platform. Its full-year revenue guidance remains only $28 million to $32 million, which tells us how early this market still is despite the rapid percentage growth.

So the quantum business is becoming real at a surprisingly fast pace, but we are still talking about a small emerging computing market rather than anything remotely comparable with cloud computing, GPUs or enterprise software.

How many companies are genuinely using quantum computers today?

Hundreds of organizations now use quantum computing in some serious form, while the group running recurring production applications remains much smaller.

McKinsey tracked more than 300 companies working with quantum-computing providers. IBM says its own quantum client and partner network now exceeds 340 organizations. Quantinuum recently reported 180 Nexus users.

We cannot add those figures together because many organizations will appear in more than one ecosystem. Even with that overlap, the order of magnitude is clear: enterprise quantum adoption has moved well beyond a handful of specialist research groups.

The harder question is how many of those users have reached production.

D-Wave provides one of the few useful financial clues because it separates production quantum-cloud revenue from the rest. Production applications generated $1.3 million during the first half of 2026. That represented 37.3% of its QCaaS revenue, up from $0.3 million and 9.8% a year earlier.

The dollar amount is still tiny. The fourfold increase in production revenue and sharp rise in production's share of cloud revenue are more interesting.

Publicly disclosed production cases remain rare enough that we can still name many of the strongest ones individually. That makes “hundreds of companies are using quantum computing” a fair statement today. Saying that hundreds are already running quantum computing in production would go much further than the evidence supports.

Is optimization the first quantum computing use case that actually works in production?

Optimization currently has the strongest evidence of real quantum-computing production use, especially through D-Wave's hybrid annealing systems.

Here the story gets concrete.

Pattison Food Group has been running a quantum-assisted scheduling application in production since 2022. The company previously needed about 80 person-hours per week to prepare e-commerce driver schedules across more than 100 stores. The automated hybrid system cut that workload to roughly 15 hours while continuing to meet at least 95% of e-commerce demand.

Ford Otosan provides a different industrial example. Its production problem involves sequencing Ford Transit vehicles through manufacturing while dealing with different configurations, parts availability and factory constraints. Its hybrid quantum application can schedule 1,000 vehicles in under five minutes, compared with roughly 30 minutes using the previous process.

Telecom has now produced an even fresher example. NTT DOCOMO recently put its second D-Wave-powered application into production. The system optimizes tracking-area lists across a mobile network serving more than 93 million subscriptions. DOCOMO reported a 65.3% reduction in peak location-registration signals and a 7% reduction in paging signals.

Three unrelated industries have therefore reached operational use through the same broad problem class: scheduling or allocating resources under lots of interacting constraints.

BASF sits one step earlier and helps show how the pipeline is expanding. In a completed proof of concept involving a liquid-filling facility, its hybrid quantum approach reduced projected scheduling calculation time from around ten hours with an industrial classical solver to roughly five seconds. BASF has yet to give us the same production evidence as Ford Otosan or DOCOMO, so we should keep the cases separate.

What seems to be working now is quite specific: companies already have difficult optimization problems, quantum-hybrid solvers plug into those workflows, and the resulting output can immediately be compared with an existing scheduling or allocation process.

Company	Problem	Current maturity
Pattison Food Group	Driver scheduling	Production
Ford Otosan	Vehicle production sequencing	Production
NTT DOCOMO	Mobile-network optimization	Production
BASF	Factory scheduling	Proof of concept

If you want more recent data on this point, please see our latest quantum computing market report.

Why is quantum optimization getting adopted before most other applications?

Quantum optimization is getting adopted first because companies can test it against an existing business problem and know fairly quickly whether the result is useful.

A factory already needs a production schedule. A telecom operator already needs to allocate network resources. A retailer already schedules drivers. These companies do not need quantum computing to invent a new workflow before they can use it.

Optimization also plays reasonably well with today's hardware. Quantum annealers can handle large combinatorial search spaces, while classical software manages the surrounding data, constraints and processing. The quantum processor only needs to contribute to one difficult part of the calculation.

The recent examples reinforce that pattern. Ford Otosan is using quantum for sequencing rather than designing an entire vehicle. DOCOMO is optimizing network configuration rather than running its telecom network on quantum hardware. Pattison uses quantum-assisted computation inside one scheduling process.

AT&T offers another useful data point without requiring us to count it as full production adoption. In work disclosed with D-Wave, the company reduced one field-technician dispatch optimization from around an hour to roughly 15 seconds. AT&T has since explored related problems such as outage response, technician routing and network planning.

We can already see where early adoption clusters: bounded problems, existing operational data, obvious constraints and an output that the company already knows how to use.

That is why optimization currently feels much more concrete than sweeping claims about quantum transforming entire industries.

Is quantum annealing doing more useful work today than gate-based quantum computing?

Quantum annealing currently has better public evidence of production business applications, while gate-based quantum computing is further ahead in many of the scientific problems that could eventually create much larger value.

The difference comes partly from what each architecture is trying to do.

D-Wave's annealing machines specialize in optimization. That narrower target has produced the production cases we see at Pattison, Ford Otosan and NTT DOCOMO.

Gate-based machines from IBM, Quantinuum, IonQ and others aim at a much broader class of computation. They could eventually become useful for molecular simulation, materials, financial algorithms and many other problems. Today's machines still face limits around errors, circuit depth and the amount of useful computation they can sustain.

Recent gate-model progress is impressive. Quantinuum says its Helios system has demonstrated logical operation at close to five-nines fidelity under one error-correction approach. IBM continues to target its first demonstrations of quantum advantage using quantum computing together with high-performance computing. IonQ is deploying successive generations of trapped-ion machines to customers.

Those achievements make commercial use easier and bring harder applications closer.

Production evidence still favors the narrower annealing route.

That may eventually reverse. For now, if somebody asks us where quantum processors themselves are already touching day-to-day business operations, optimization on annealing and hybrid systems is the clearest answer.

If you want more recent data on this point, please see our latest quantum computing market report.

Are drug companies actually using quantum computing for drug discovery?

Pharmaceutical companies are genuinely using quantum computing in research today, but we still cannot point to a commercial drug that owes its discovery to a quantum computer.

Moderna's work with IBM is a good example of how far the research has moved. The two companies used an IBM Heron processor to study mRNA secondary structure, initially working with as many as 80 qubits and sequences up to 60 nucleotides. Their subsequent work expanded the approach to problems using as many as 156 qubits and 950 non-local gates.

That is substantial quantum hardware usage inside a real pharmaceutical research program. It still sits upstream from discovering, validating and commercializing a medicine.

Pfizer has recently moved into a similar layer with Quantinuum and NVIDIA. The companies combined AI with quantum simulation to investigate molecular properties on Quantinuum hardware. Quantinuum described the work as a way to push pharmaceutical quantum simulation toward larger and more useful chemistry problems.

Japan Tobacco has explored another route using D-Wave technology. Its pharmaceutical division combined quantum computation with AI during early-stage drug discovery and reported generating molecular structures that looked more promising for later drug development than those produced by its classical approach. The result remains an experimental research outcome rather than evidence of a quantum-created approved drug.

The repeated engagement from Moderna, Pfizer, Japan Tobacco and other pharmaceutical groups makes drug discovery one of the strongest examples of serious quantum R&D adoption.

The remaining gap is huge, though. Drug discovery is valuable when better computation eventually changes which molecules enter the pipeline, how quickly researchers reach them or how often those candidates succeed. We have very little public evidence of that downstream impact yet.

Is quantum chemistry closer to real adoption than quantum drug discovery?

Quantum chemistry currently looks closer to becoming a useful enterprise workflow because companies are already running repeated molecular and materials calculations on real quantum hardware.

BMW has worked with Quantinuum for several years and recently extended that relationship through a new multi-year agreement. The research covers areas such as catalytic reactions, material behavior and chemical simulation.

BMW and Airbus previously worked with Quantinuum on the oxygen reduction reaction, a chemically difficult process relevant to hydrogen fuel cells. Other industrial users have explored similar techniques for batteries, catalysts and advanced materials.

The important change is repetition. These companies keep coming back to chemistry rather than performing one experiment and disappearing.

Software is also becoming more usable. Quantinuum's InQuanto gives computational chemists tools designed specifically for quantum chemistry. Amazon Braket now supports third-party chemistry workflows such as Kvantify's Qrunch, which can send jobs to different quantum processors through the same cloud environment.

That removes some of the friction that used to make every quantum chemistry project feel custom-built.

Still, we should be precise about where adoption has reached. Scientists can use quantum machines inside chemistry R&D today. We have much weaker evidence that quantum computation has already produced a commercially important molecule or material faster or cheaper than the best classical methods.

Quantum chemistry therefore looks like real research adoption with plausible industrial value ahead, rather than mature production computing.

Are banks really using quantum computing, or are they just experimenting?

Large banks are taking quantum computing seriously enough to build permanent research capabilities, although their actual trading, pricing and risk systems remain overwhelmingly classical.

JPMorgan Chase is the clearest example. The bank has spent years developing quantum algorithms around portfolio optimization, option pricing, machine learning and other financial problems. It has worked with IBM, Quantinuum, AWS, QuEra and other providers rather than tying its research to a single hardware architecture.

That level of repeated investment tells us quantum has become an institutional capability inside the bank.

Recent work still shows how early the production side is. JPMorgan and AWS developed a hybrid optimization pipeline that used a neutral-atom quantum processor from QuEra as a co-processor. For some test problems, classical preprocessing reduced the original portfolio problem by roughly 80% before the harder remaining portion was sent into the quantum workflow.

JPMorgan has also started building a dedicated Quantum-AI research environment in London with OQC and AMD. The project will combine quantum hardware, AI and high-performance classical computing so researchers can test problems such as portfolio optimization in a secure enterprise environment.

That is considerable infrastructure for something supposedly stuck at the conference-demo stage.

We still have no strong public evidence that a major bank has moved a core portfolio optimizer, derivatives pricing engine or daily risk calculation onto quantum hardware because it already performs better there.

Finance currently represents deep organizational adoption without much disclosed production replacement.

Is cloud access becoming the main way companies adopt quantum computing?

Quantum cloud access is now one of the most established products in the entire quantum-computing market.

The reason is fairly simple. Most companies have no interest in installing dilution refrigerators, ion traps or the other infrastructure needed to operate a quantum processor themselves. They want access to computation.

McKinsey's latest research found a clear split: private companies were more likely to use hosted quantum applications through cloud providers, while public institutions were more inclined to buy machines for on-premise deployment.

Amazon Braket already lets customers access processors from multiple hardware vendors through one service. IBM operates its own cloud quantum platform. Quantinuum's Nexus now has 180 organizations using it. IonQ sells cloud access alongside direct hardware deployments.

AWS has also been working on the boring engineering details that usually show a technology is becoming easier to consume. Its Program Sets feature bundles repeated quantum executions together and can reduce the overhead surrounding some workloads dramatically. Braket also integrates error-suppression tools and managed classical resources around the quantum computation.

Quantinuum is pushing this further through Oracle. Helios is being integrated into an Oracle Cloud Infrastructure AI data center, where customers are expected to access quantum computation alongside ordinary compute, networking, storage, identity and governance systems.

That architecture looks much closer to how an enterprise buys computing today.

Cloud quantum is already real adoption even while many of the workloads running through it remain experimental. Companies are increasingly consuming QPUs as remote specialized hardware instead of treating every quantum project like a physics collaboration.

Is hybrid quantum computing what companies are really buying right now?

Hybrid quantum-classical computing has become the practical model for most serious enterprise quantum work today.

Almost every credible near-term use case follows roughly the same logic. CPUs and GPUs handle the parts they already do well. The software isolates a smaller problem that might benefit from quantum computation. A QPU handles that section. Classical systems then process, verify or use the result.

D-Wave's production applications work this way. JPMorgan's recent portfolio-optimization experiments work this way. Amazon Braket explicitly provides managed hybrid workloads. IBM's roadmap for quantum advantage assumes quantum machines working with high-performance classical computing.

Quantinuum is building the same architecture into its commercial strategy. Its Oracle partnership puts Helios next to conventional cloud compute and AI infrastructure, while its collaboration with HPE is designed around combining quantum, HPC and AI for enterprise workloads.

The convergence here may be more important than any individual hardware announcement.

Quantum computers are starting to find their place in the computing stack as specialized accelerators. That model is much easier for an enterprise to absorb than replacing classical computers with a completely separate quantum environment.

It also gives the industry somewhere useful to go while quantum hardware improves. A company can build the classical pipeline, data flows, algorithms and orchestration today, then give the QPU a larger role when the machine becomes more capable.

Have quantum computers actually beaten classical computers on useful business problems yet?

Quantum computers have produced increasingly credible advantage results, but broad commercial quantum advantage remains unproven today.

This is one of the easiest areas to exaggerate because “quantum supremacy,” “quantum utility” and “quantum advantage” are often used as though they mean the same thing.

A quantum system can perform a calculation that is difficult to simulate classically without giving a company anything useful. The harder threshold is showing that the complete quantum approach beats every relevant classical method on a problem somebody cares about.

Classical competition also keeps improving. When researchers publish a promising quantum result, other teams often respond with a better GPU algorithm, tensor-network simulation, heuristic or approximation technique. Any claim of advantage therefore has to survive a moving benchmark.

IBM's current roadmap captures where gate-model computing sits. The company is targeting examples of quantum advantage using quantum processors together with HPC and is building an open benchmarking effort so proposed advantage workloads can be tested against classical competitors.

JPMorgan's research reaches a similar conclusion from the application side. Some of its work has found attractive theoretical scaling for quantum optimization algorithms, yet researchers still need bigger and better hardware before they can run those methods in the regime where the scaling advantage becomes commercially interesting.

D-Wave makes stronger claims around annealing and has published peer-reviewed work showing quantum behavior on problems that become extremely difficult to reproduce classically. Its production customers give the company unusually concrete business cases. Even there, proving that every operational improvement comes specifically from an irreducible quantum advantage is a tougher claim than showing that the hybrid product works well.

So we should resist collapsing “useful quantum application” and “proven quantum advantage” into one category. Useful applications now exist. Broad proof that QPUs systematically beat the best classical tools for enterprise workloads is still missing.

If you want more recent data on this point, please see our latest quantum computing market report.

Why are companies spending millions on quantum computing before that advantage is proven?

Big companies are spending on quantum computing now because they can afford to prepare early, and the skills required to use future machines cannot be built overnight.

McKinsey's spending data makes that strategy fairly obvious. Among the large companies it studied, one-third were already allocating more than $10 million annually to quantum initiatives. The biggest buckets were application development, integration with existing systems and internal capability building.

In other words, much of today's quantum budget buys preparedness.

JPMorgan has spent years building researchers, algorithms and intellectual property. BMW has maintained a multi-year relationship with Quantinuum. Moderna continues working with IBM. Pharmaceutical, chemical, automotive and financial companies are all learning which internal problems map well onto quantum hardware.

The economics also look different when we put a $10 million budget inside a giant company. For an automaker, bank or pharmaceutical group already spending billions on technology and R&D, $10 million can be a relatively small strategic option.

Waiting carries its own cost. If fault-tolerant or genuinely advantageous quantum computing arrives faster than expected, companies beginning from zero would need to find specialist staff, identify useful problems, redesign workflows and integrate completely unfamiliar infrastructure at the same time.

That helps explain the strange-looking market we have today: commercial adoption can accelerate several years before widespread commercial advantage.

What part of quantum computing is still mostly hype?

Claims that quantum computing is already transforming drug discovery, finance, AI or ordinary enterprise computing still run well ahead of the evidence.

We get a much cleaner picture when we look at what companies themselves say their applications are doing.

NTT DOCOMO can point to an application running inside network operations and measure a 65.3% reduction in one type of peak signaling. Ford Otosan can point to a production scheduler and compare five minutes with 30 minutes. Pattison can compare 15 hours of scheduling work with 80.

Those are unusually concrete quantum stories.

The language changes once we move into many gate-based chemistry and finance announcements. Companies talk about simulations, experimental workflows, algorithm development, future fault-tolerant machines and research toward advantage. Serious work is happening, but production impact is much harder to find.

Revenue can create another illusion. IonQ's latest quarterly revenue is growing extremely quickly, yet IonQ has also expanded into quantum networking, sensing, security and acquired businesses. Hardware deployments and large strategic contracts can make revenue jump before end customers have discovered a large number of production quantum-computing workloads.

The same issue appears with system sales. An institution buying a multimillion-dollar quantum computer is clearly adopting quantum infrastructure. That purchase tells us much less about how many commercially valuable calculations will run on the machine.

The hype begins when all of these categories get bundled together under one sentence: “quantum is here.”

Some parts really are here. Others remain expensive preparation for what companies hope comes next.

So what is getting real adoption in quantum computing now?

Quantum computing has real adoption today, with optimization, cloud access and hybrid quantum-classical infrastructure clearly ahead; chemistry and finance have reached serious enterprise R&D, while broad gate-based production advantage remains too early.

The strongest production evidence comes from optimization. We have recurring deployments in retail scheduling, automotive manufacturing and telecom networks, with companies publishing measurable operational outcomes rather than simply reporting experiments.

Cloud access is probably the broadest form of adoption. Hundreds of organizations now use quantum platforms, and providers are increasingly packaging QPUs the way customers already consume other specialized computing resources.

Hybrid computing may prove even more important. AWS, IBM, Quantinuum, D-Wave, JPMorgan and others are converging on an architecture where classical systems perform most of the work and quantum processors handle selected calculations. That model already fits existing enterprise infrastructure.

Chemistry, materials and finance sit one stage behind production optimization. BMW, Moderna, Pfizer, JPMorgan and other major companies have kept their quantum programs alive for years and continue pushing them toward harder problems. Those programs are real enough to count as adoption. The measurable end-product impact remains limited.

As we saw above, D-Wave's latest numbers give us one of the cleanest measures of how this transition is moving: production applications rose from 9.8% to 37.3% of its quantum-cloud revenue within a year. That is a meaningful jump, while the underlying production revenue of $1.3 million reminds us how small the market still is.

We can therefore answer the title much more precisely than “quantum is finally here.”

Quantum computing is already being adopted where the QPU can act as a narrow accelerator inside an existing workflow. Optimization has moved furthest into production. Quantum cloud and hybrid computing have become genuine infrastructure products. Chemistry and finance are becoming permanent research capabilities inside large companies.

The huge unresolved step is broad computational superiority. When gate-based quantum machines consistently make an important commercial calculation faster, cheaper or better than the best classical alternative, the adoption story will change completely.

We are not there yet.

Part of quantum computing	What is happening now	Our judgment
Optimization and annealing	Several measurable production deployments	Real production adoption
Quantum cloud access	Hundreds of organizations using platforms	Real adoption
Hybrid quantum + HPC/AI	Becoming the default enterprise architecture	Real infrastructure adoption
Chemistry and materials	Repeated industrial research on hardware	Real R&D adoption
Finance	Permanent teams and advanced experiments	Serious R&D adoption
Direct system purchases	More institutions installing QPUs	Real but niche
Gate-based quantum advantage	Stronger candidates and benchmarks emerging	Still early
General enterprise workloads	Very little production evidence	Premature

If you want more recent data on this point, please see our latest quantum computing market report.

Who is the author of this content?

NEW MARKET PITCH TEAM

We track new markets so founders and investors can move faster

We build living "market pitch" documents for emerging markets: AI, synthetic biology, new proteins, and more. Instead of outdated PDFs or hallucinated LLM answers, our clients get a clean, visual, always-updated view of what's really happening: key players, deals, regulations, and signals that matter. Learn more about us.