The Quantum Computing Race Is Shifting From Qubits To Infrastructure

The Quantum Computing Race Is Shifting From Qubits To Infrastructure

For most of its commercial history, quantum computing has been sold as a promise about the future. The new evidence points somewhere more familiar to anyone who buys industrial capacity for a living. Quantum is starting to behave like an infrastructure procurement market, with machines being bought outright, wired into supercomputers, subjected to sovereignty and integration conditions, and surrounded by internal teams built years before the hardware is expected to pay for itself. That shift, rather than any single technical milestone, is the development worth the attention of anyone who thinks about how large compute assets get specified, financed and operated.

The signal comes from State of Quantum 2026, the fourth annual industry study published by IQM Quantum Computers with research conducted independently by The Quantum Insider, a Resonance company. The headline tension is stark. Among industry respondents, 89% report hands-on quantum work, yet only 10% report limited production use and just 3% have reached scaled deployment. Engagement is close to universal while production remains rare, and the report argues that the organisations closing that gap now will hold an advantage later entrants find structurally difficult to close.

For a readership that procures roads, rail, ports and the digital systems that run them, the interesting part is not the physics. It is that the procurement logic looks increasingly like the logic of buying any other piece of national or industrial compute capacity.

Briefing

• Quantum computing has moved from a question of access to a question of capability, with 89% of surveyed organisations doing hands-on work but only 3% reaching scaled deployment, according to State of Quantum 2026.
• Quantum computing drew 8.3 billion US dollars of investment in 2025, close to five times the prior year, and the report attributes the rise to genuine procurement and larger deal sizes rather than a jump in deal count.
• On-premises and hybrid infrastructure is becoming the preferred model, with roughly 46% of buyers expecting on-premises infrastructure to form part of their access model within three years against 24% favouring public cloud alone.
• The binding constraints are now people and software rather than hardware, with skills shortages cited by 66% or more of large enterprises, universities and government buyers, and workforce training running two to five years.
• IQM completed its listing and began trading on the Nasdaq Global Select Market as IQMX on 2 July 2026, becoming the first European quantum computing company on a major US exchange, a detail the pre-listing source material had not yet caught up with.

The Market Can Finally Be Measured, and the Numbers Read Like Capital Spending

The clearest sign that quantum has entered a procurement phase is that the market can now be quantified in the language of contracts and capital rather than laboratory demonstrations. The report tracks 121 cumulative deals with an estimated combined value of 2.6 billion US dollars by the first quarter of 2026, with annual contract volume up more than sixfold since 2021.

Investment into quantum computing reached 8.3 billion US dollars in 2025, close to five times the 2024 figure, and the report is explicit that this was driven by larger deals rather than more of them. Deal count held broadly flat while cheque sizes grew, which is exactly the pattern seen when a technology moves from speculative funding into genuine buying by institutions with budgets to defend.

Alex Challans, chief executive of The Quantum Insider, framed the change in terms that will be recognisable to any infrastructure finance professional. “Markets mature when their questions do,” he said. “A year ago, people were still asking whether quantum investment had peaked. This year’s report closes that debate. Capital is arriving at scale, and it is going to the companies with demonstrated results behind their roadmaps.”

The point is not that money has arrived. It is that money is now following delivery rather than ambition, rewarding vendors who can show verifiable milestones over those selling a distant vision. That is the behaviour of a maturing procurement market, and it changes what suppliers have to prove and what buyers have to plan for.

To put a number on how prepared the buyer base actually is, the report introduces a Quantum Readiness Index, a composite score across workforce, innovation, investment and adoption. The global cohort scores 58 out of 100, placing it in a developing tier that has moved past awareness but is not yet ready for scaled deployment.

The revealing detail inside that score is that hiring, budget and pilots are running ahead of proprietary output and production use, with only 9% of surveyed organisations reporting a resourced intellectual property programme. Readiness is real but shallow, which is precisely the profile of a market that has started buying capacity faster than it has learned to exploit it.

Buyers Have Stopped Asking About Qubit Counts and Started Asking About Ownership

The most consequential behavioural change the report documents is what serious buyers now interrogate before they commit. They have largely stopped leading with qubit counts and now ask whether they can see into a machine, calibrate it, integrate it with the systems they already run, and retain the capability they build around it.

That is not a research question. It is the same set of questions an HPC centre or a large infrastructure operator asks about any major compute acquisition, and it explains why on-premises and hybrid models are pulling ahead of pure cloud access. Across hybrid and standalone models, roughly 46% of buyers expect on-premises infrastructure to form part of their access model within three years, compared with 24% favouring public cloud alone.

The reason ownership matters comes down to timing, and the report makes a case that will resonate with anyone who has planned a long-lived asset. Quantum advantage is not delivered at the point of installation. It is built incrementally through trained people, algorithms written for specific problems, and operational experience accumulated over successive cycles, none of which can be assembled quickly later. Vendor roadmaps across the major hardware modalities now converge on a 2029 to 2031 window for fault-tolerant systems, which turns the intervening years into the period in which capability has to be built.

Jan Goetz, co-founder and chief executive of IQM, put the strategic choice bluntly in the report’s foreword. “The quantum future is closer than it looks,” he wrote. “The work of being ready for it starts now. The organisations holding out for a clear signal tend to find the signal and the deadline show up on the same morning.”

For infrastructure owners the analogy is direct. A quantum computer is not a server that starts delivering value the moment it is switched on. It is closer to a specialised piece of plant whose return depends on the workforce trained to operate it and the operational knowledge accumulated across successive cycles. Cloud access remains the sensible entry point for exploration because it offers low upfront cost and hardware variety, but organisations building durable capability are increasingly treating operated infrastructure as the anchor and cloud as the flexible layer around it.

The strategic question, as the report frames it, is no longer whether to own or to rent, but what to operate directly and what to leave in the cloud.

The Real Bottlenecks Are People and Software, Not Machines

One of the more useful corrections in the report, and one that maps almost exactly onto the construction and infrastructure sector’s own experience of digital transformation, is that the binding constraints are no longer mainly about hardware. Skills shortages are the most consistently cited barrier, named by 66% or more of large enterprises, universities and government buyers. The second constraint is algorithm design rather than qubit immaturity, the difficulty of mapping real-world problems onto quantum systems and the absence of efficient algorithms to do it. The third is return-on-investment uncertainty at the commitment threshold, concentrated among commercial buyers weighing a multi-year quantum horizon against near-term budget pressure where generative AI already delivers demonstrable results.

The workforce point deserves emphasis because it carries a hard timing constraint that no amount of capital can shortcut. Quantum workforce training takes two to five years, and the pipeline does not grow unless organisations begin now. This is the same structural problem the infrastructure sector knows from every skilled trade and every wave of new technology: the people needed to run tomorrow’s assets have to be trained against today’s, well before the payoff is visible.

The report’s blunt implication is that an organisation which waits for fault-tolerant hardware before building its team will arrive at the starting line just as the race finishes. Software maturity compounds the challenge, with the report noting that quantum software ecosystems, like the AI ecosystems before them, tend to take five to seven years to reach operational maturity, lagging the hardware they are meant to support.

There is a further constraint that will be familiar to anyone who has integrated proprietary industrial systems: the absence of cross-vendor standards. Without a portability layer comparable to the standards that let classical high-performance computing move workloads between machines, each quantum vendor integration requires bespoke engineering, algorithms do not transfer cleanly between platforms, and buyers who commit to a non-interoperable vendor face rising switching costs over time. For procurement teams, that raises the same lock-in questions that shape any major systems purchase, and it is pushing openness and interoperability up the list of formal selection criteria.

Openness and Sovereignty Are Becoming Procurement Gates, Not Preferences

As quantum buying has matured, two conditions have moved from nice-to-have to gating requirement, and both will be instantly recognisable to public infrastructure buyers. The first is openness. HPC centres and technically advanced enterprises increasingly insist on the ability to interrogate a system, understand its calibration state and access low-level controls, because a closed black-box machine is incompatible with a buyer whose goal is to build capability of its own rather than consume a finished service. The report frames the relationship between an open-system vendor and an engaged buyer as a co-development one rather than a standard customer-supplier transaction, in which the buyer surfaces bugs and feeds requirements back into the product and the quality of that relationship shapes how quickly the technology matures for everyone.

The second gate is sovereignty, and here the parallels with critical national infrastructure are exact. In Europe and the Gulf, local presence, data residency and host-country capability are increasingly written into formal procurement requirements. A vendor without a local entity or in-country service model may not reach technical evaluation at all, regardless of hardware quality, which makes geographic footprint part of go-to-market strategy rather than a regional sales decision. This is reinforced at the policy level.

The European Commission presented its European Technological Sovereignty Package on 3 June 2026, aimed at reducing structural dependencies across semiconductors, AI, cloud and open source, while Saudi Arabia frames quantum as a matter of national sovereignty through its National Quantum Alliance. The United States has moved in parallel: in May 2026 the Department of Commerce signed letters of intent worth 2.013 billion US dollars with nine companies under the CHIPS and Science Act, including foundry incentives for IBM and GlobalFoundries, explicitly on the grounds that a strong domestic quantum ecosystem is essential to national security and long-term strategic leadership.

Governments, in other words, are not simply funding hardware. They are funding the ability to build, operate and retain quantum competence inside their own borders, which is the defining characteristic of infrastructure investment rather than technology procurement.

National Compute Programmes Are Where the Buying Is Concentrated

The clearest evidence that quantum has become an infrastructure category lies in where the machines are actually going. Public-sector and research-infrastructure buyers remain the main source of recorded procurement, and their behaviour has shifted from isolated pilots to staged capability building along multi-year ladders. Finland’s VTT is scaling from 5 to 20, 150 and 300 qubits with the same vendor across successive cycles.

Germany’s Leibniz Supercomputing Centre is moving from 20 to 54 qubits with a 150-qubit system planned, and IQM has already launched an integration service that lets its machines operate as nodes inside a supercomputing environment there. When an anchor institution buys from the same supplier across four successive upgrades, the pattern is closer to staged repeat procurement from a major customer than to experimental one-off pilots, even if broad commercial recurrence across the wider market remains unproven.

More than 30 national HPC-quantum programmes have launched since 2025, spanning 14 countries and treating quantum processors as specialised accelerators sitting alongside classical supercomputers. Examples run from EuroHPC procurements across Germany, Finland, Spain, France and Italy to India’s National Quantum Mission and South Korea’s national programme, where the Korea Institute of Science and Technology Information is integrating a 100-qubit system with its flagship HANKANG supercomputer as part of a broader push, backed by a national plan worth roughly 2.3 billion US dollars, to build domestic superconducting capability and scale toward thousand-qubit systems.

The report’s transaction data places IQM at the front of this national HPC-quantum activity, leading all tracked vendors with nine deployments across six countries between 2025 and the first quarter of 2026, three times the figure of the next closest vendor. The structural reason EMEA’s deployment count runs so far ahead of its share of global capital is the EuroHPC Joint Undertaking, a coordinated European co-funding model that has produced a capability ladder across member-state supercomputing centres rather than scattered individual purchases. That is procurement architecture, and it is doing for quantum what coordinated national programmes have long done for other categories of strategic infrastructure.

The Enterprise Layer Is Building the Case Before the Returns Arrive

Beyond the national programmes, enterprise buyers present a more cautious but still telling picture. Among industry respondents, 89% report hands-on engagement, yet most activity sits in pilots, proofs of concept and use-case discovery rather than production. A large majority frame their spend as strategic positioning and benchmarking against classical methods, with only about a third pointing to an early production use case. The market is not short of interest. It is short of proven quantum advantage, and buyers are essentially building the internal case, through pilots, benchmarking and workforce development, so that when validated use cases emerge they can scale rather than start from scratch.

The application areas concentrating attention will not surprise a construction and materials audience. Machine learning and AI augmentation lead at 79%, followed by optimisation problems such as scheduling, routing and portfolio work at 66%, with security, simulation and cryptography behind. The report identifies materials science as the most mature use case today, spanning batteries, fuel cells and advanced materials, where AI handles pre- and post-processing and quantum methods target the core computations.

For an industry whose future depends on better materials, lower-carbon binders and smarter asset optimisation, that is the corner of the quantum map worth watching, even though the timing of commercial payoff remains genuinely uncertain. The enterprise message is one of patient positioning: build literacy among engineers now, run pilots for the organisational learning rather than the computational output, and keep options open across vendors and modalities while no single architecture has demonstrated a durable lead.

IQM’s Public Listing Confirms the Direction of Travel

A detail in the supplied material has since been overtaken by events, and the update reinforces the report’s central argument. The source described IQM as nearing its Nasdaq listing through a merger with Real Asset Acquisition Corporation. That transaction has now completed. IQM began trading on the Nasdaq Global Select Market under the ticker IQMX on 2 July 2026, becoming the first European quantum computing company listed on a major US exchange, with a pro forma cash position of around 337 million euros and, on its own account, more quantum computers sold worldwide than any other manufacturer. Shares in the underlying company also began trading on Nasdaq Helsinki the following day.

The listing matters for reasons beyond IQM itself. It confirms a structural shift the report tracks, in which quantum companies are moving into public markets to fund the capital demands of building and shipping physical machines, with seven SPAC mergers completed since 2021 and a further wave through 2025 and into 2026. Public capital is a currency for scale and for acquisitions, and it signals that the infrastructure layer of the quantum market is consolidating around companies that can demonstrate delivery.

The timing of surrounding events makes the point look current rather than speculative. On 11 August 2026, Quantinuum and Oracle announced a multi-year partnership to deploy Quantinuum’s Helios system inside an Oracle Cloud Infrastructure data centre, integrating quantum with the same high-performance computing, GPU and governance environment enterprise customers already use. That is the hybrid quantum-classical-AI model the report describes, arriving as a commercial cloud service from one of the largest infrastructure providers in the world.

The direction of travel is consistent across the report’s evidence and the developments that have followed it: quantum is being absorbed into the industrial and national compute stack, and the organisations that use the years before fault tolerance to build talent, integration experience and proprietary learning will be better placed than those that wait for a signal that arrives too late to act on.

Key Industry Questions

1. Is quantum computing actually being bought as infrastructure, or is this still mostly research funding? Both are true, but the balance is shifting. Paid hardware procurement remains concentrated among governments, national laboratories and supercomputing centres, and enterprise activity is still dominated by pilots rather than production. What has changed is the buying behaviour. Public and research-infrastructure buyers are now planning purchases as staged, multi-year capability ladders with the same vendor across successive upgrades, wiring machines into supercomputers and writing sovereignty and integration terms into procurement. That is the pattern of infrastructure acquisition rather than experimental grant spending. The report is careful to note that broad commercial recurrence across the wider market remains unproven, so this is an infrastructure market in formation rather than a fully self-sustaining commercial one.
2. Why does the report emphasise on-premises ownership when cloud access is cheaper to start with? Because the decision depends on usage volume and time horizon, not on cost alone. Cloud access is the sensible entry point for exploration, offering low upfront cost, hardware variety and automatic upgrades. For institutions supporting sustained use by many researchers, reserved or dedicated cloud access can become expensive, and one interview cited costs of the order of 100,000 US dollars per day to support roughly ten scientists. More importantly, on-premises ownership supports workforce development, intellectual property control, HPC integration and sovereignty, which are the foundations of durable capability. Roughly 46% of buyers expect on-premises infrastructure to form part of their access model within three years, against 24% favouring public cloud alone. The mature answer is a hybrid one, with operated infrastructure as the anchor and cloud as the flexible layer.
3. What are the real barriers to adoption if not the hardware? The report identifies three, and only one is technical in the conventional sense. The most consistent is skills, cited by 66% or more of large enterprises, universities and government buyers, with training taking two to five years to reach working proficiency. The second is algorithm design and software accessibility, the difficulty of mapping real-world problems onto quantum systems and the immaturity of the software stack, compounded by the absence of cross-vendor standards that would let workloads move between platforms. The third is return-on-investment uncertainty at the point of scaling, concentrated among commercial buyers weighing a multi-year quantum horizon against generative AI, which already delivers demonstrable returns. Hardware immaturity still matters most for the technically advanced buyers whose workloads test the limits of current systems.
4. When is fault-tolerant quantum computing actually expected, and how firm is that date? Vendor roadmaps across the major hardware modalities converge on a 2029 to 2031 window for fault-tolerant systems. The report treats this as a genuine convergence rather than marketing, but it is careful about certainty. The timing of useful quantum advantage remains uncertain, no single hardware modality has demonstrated a decisive and durable lead, and any commitment to a single architecture today carries material technical risk. The practical takeaway is not the precise date but its consequence: the years before fault tolerance are the window in which workforce, integration experience and proprietary learning have to be built, because those cannot be acquired quickly once the technology matures.
5. What does the IQM Nasdaq listing signify for the wider market? IQM began trading on the Nasdaq Global Select Market as IQMX on 2 July 2026, becoming the first European quantum computing company on a major US exchange, following its business combination with Real Asset Acquisition Corporation and leaving it with a pro forma cash position of around 337 million euros. The significance is structural. Building and shipping cryogenic quantum machines is capital-intensive, and public markets offer both a deeper funding pool and a currency for acquisitions. The listing sits within a broader wave of quantum companies entering public markets, with seven SPAC mergers completed since 2021, and it signals that the hardware layer is consolidating around firms able to demonstrate delivery rather than roadmap ambition.
6. How relevant is any of this to construction, infrastructure and materials businesses today? Directly relevant as preparation, not yet as production tooling. The most mature quantum use case today is materials science, spanning batteries, fuel cells and advanced materials, which bears on lower-carbon binders, better composites and asset optimisation. Optimisation problems such as scheduling and routing are among the most active areas of exploration and map onto logistics and project planning. The report’s honest position is that quantum advantage in a commercially meaningful use case has not yet been demonstrated, so the sensible posture for most infrastructure businesses is to build literacy, watch the materials-science work, and understand how national compute programmes are integrating quantum, rather than to procure hardware.
7. Why is sovereignty becoming a procurement condition rather than a preference? Because governments have concluded that quantum capability is strategic infrastructure that has to be retained domestically. In Europe and the Gulf, local presence, data residency and host-country capability are increasingly written into formal procurement, and a vendor without a local entity may not reach technical evaluation regardless of hardware quality. This is reinforced by policy: the European Commission’s European Technological Sovereignty Package of 3 June 2026, Saudi Arabia’s National Quantum Alliance, and the US Department of Commerce’s 2.013 billion US dollar CHIPS-Act commitments to nine quantum companies in May 2026. The common thread is that these programmes fund the ability to build, operate and retain quantum competence within national borders, which is the logic of critical infrastructure rather than ordinary technology purchasing.

Strategic Takeaways

1. Quantum has crossed from an access question to a capability question, and the organisations building workforce, integration experience and proprietary learning during the 2029 to 2031 pre-fault-tolerance window will hold an advantage that later, better-resourced entrants find structurally difficult to close.
2. The binding constraints are people and software, not machines, so the decisions that matter most in the near term are workforce training that takes two to five years and internal capability building, not chasing headline qubit counts.
3. On-premises and hybrid infrastructure is becoming the preferred model precisely because durable capability requires ownership, integration and control, mirroring how national and industrial compute assets have always been procured and operated.
4. Sovereignty, openness and interoperability have hardened into procurement gates rather than preferences, which means vendor selection increasingly turns on local presence, transparency and co-development capacity as much as on raw hardware performance.
5. Capital and public listings are now following demonstrated delivery rather than roadmap ambition, and events since the report, from IQM’s Nasdaq debut to the Oracle-Quantinuum cloud partnership, confirm that quantum is being absorbed into the mainstream industrial and national compute stack.

SEO Package

Main title

Quantum Stops Being a Science Project and Starts Behaving Like Infrastructure Procurement

Five alternative title options

Quantum Computing Becomes an Infrastructure Buy: Inside the State of Quantum 2026
From Access to Ownership: How Quantum Turned Into a National Compute Procurement Market
Why Quantum Buyers Have Stopped Asking About Qubits and Started Demanding Sovereignty
The Capability Race Has Begun: Quantum Computing Enters Its Infrastructure Phase
Machines, Sovereignty and Skills: The Quiet Industrialisation of Quantum Computing

Primary SEO keyword

quantum computing infrastructure procurement

Keyphrase synonyms

quantum computing market 2026, State of Quantum 2026 report, HPC quantum integration, quantum readiness index, on-premises quantum computing, national quantum programmes, fault-tolerant quantum computing timeline, IQM Nasdaq listing IQMX

SEO tags