Why Are Companies and Governments Buying Quantum Computers That Can’t Do Anything Yet?

The global quantum computing market reached $1.4 billion in 2025, according to the QED-C’s State of the Global Quantum Industry 2026 report. Governments worldwide have committed over $56 billion to quantum research and development. IonQ crossed $130 million in annual revenue. The UK just launched a £2 billion quantum programme with an explicit advanced procurement component. Japan declared 2025 “the first year of quantum industrialization” and backed it with billions.

All of this spending is flowing into machines that, by any honest measure, cannot yet outperform a good laptop for any commercially relevant task. As I mapped in detail in my Quantum Utility Map series, practical quantum advantage remains years away even for the five industries where the evidence is strongest. So why is anyone writing checks?

The answer is more rational than it appears, and more varied than most analyses suggest. Through my work at Applied Quantum, I spend a lot of time helping enterprises and governments cut through quantum computing marketing claims and vendor hype. We are always very honest with clients about what current quantum computers can do (not much) and what they cannot (nearly everything they are marketed for). And yet, after those conversations, clients consistently come back with valid reasons why they want to invest anyway. The pattern is familiar. The organizations buying quantum computers today are not, for the most part, deluded about what they’re getting. They are making calculated bets on strategic, institutional, and industrial logic that operates on a different timeline than “can this machine do useful computation right now?”

Here is my breakdown of the real reasons behind the buying.

The Procurement Clock Runs Slower Than the Physics Clock

The single most underappreciated reason for buying quantum computers today has nothing to do with qubits, gate fidelities, or algorithm performance. It has to do with procurement cycles.

In large enterprises and government agencies, the path from “we should evaluate this technology” to “we have an operational system integrated into our environment” takes three to seven years. Security accreditation alone can consume 18 months. Facilities modifications for cryogenic systems require planning, budgeting, environmental review, and construction. Integration with existing HPC infrastructure demands custom engineering. Vendor evaluation, competitive bidding, contract negotiation, and approval chains add more.

The organizations buying quantum hardware today are not betting that current machines will deliver value. They are ensuring that when the next generation arrives (the one that might actually matter) they are not starting from zero. The facilities exist. The security accreditation is done. The integration patterns are established. The vendor relationship is mature. The internal team knows the technology stack.

This is the same logic that drove early cloud computing adoption in enterprises circa 2006-2008. The technology was not mature enough for mission-critical workloads, but organizations that waited for maturity found themselves five years behind those who started building muscle memory early.

Sovereign Capability Is Not Optional

The second major driver is national sovereignty, and it operates on an entirely different logic than commercial return on investment.

When the UK commits £1 billion specifically to procuring large-scale quantum computers through its ProQure programme, it is not primarily asking whether those machines will generate economic returns in the near term. It is asking whether the United Kingdom will have onshore quantum computing capability when it matters, or whether it will depend on foreign providers for a technology class that touches cryptography, defense, materials, and pharmaceutical research.

This calculation has become urgent for reasons I have covered extensively in my China’s Quantum Ambition series and in Quantum Sovereignty more broadly. The nations currently investing most aggressively in quantum computer procurement are the same ones that learned painful lessons from semiconductor dependency. The pattern is consistent across programs:

Canada launched its Canadian Quantum Champions Program with $92 million backing four domestic quantum hardware companies. Japan invested ¥50 billion ($335 million) specifically to back more than 10 quantum companies, with Fujitsu and RIKEN building a 256-qubit system that is as much about industrial capability as scientific output. Singapore signed a strategic partnership with Quantinuum to install a Helios system on its territory. Even the U.S. government has been exploring equity stakes in quantum companies like IonQ, Rigetti, and D-Wave, a move that would mark a significant shift toward direct government ownership in a technology sector.

The sovereign logic is straightforward: classified, defense-related, and regulated workloads cannot run on foreign cloud platforms. When quantum computing matures enough to affect cryptographic security, national intelligence, drug development, or materials design, countries that lack onshore capability will face an unacceptable dependency. Buying early, even before the machines are useful, is the cost of avoiding that dependency.

As I have argued in Quantum Sovereignty, the nations that control quantum computing infrastructure will hold asymmetric advantages in the same way that semiconductor leadership has shaped geopolitical power. The time to build that infrastructure is before it becomes critical, not after.

The Workforce Cannot Be Hired Into a Vacuum

Recruiting and retaining quantum talent is among the most competitive hiring challenges in deep technology. Only one qualified candidate exists for every three specialized quantum positions globally, according to McKinsey estimates. The researchers, engineers, and application developers who can operate quantum hardware, write quantum algorithms, and integrate quantum-classical hybrid workflows are rare and in demand from Google, IBM, Microsoft, Amazon, and every well-funded startup.

Organizations that purchase quantum hardware create a gravitational pull for this talent. Researchers who would otherwise leave for a tech giant or a well-funded startup will stay at a national laboratory, a defense contractor, or a pharmaceutical company if it offers access to real quantum systems. The hardware is as much a retention tool as a research platform.

There is a deeper point here as well. The institutional learning curve for quantum computing is steep, non-linear, and cannot be accelerated by reading papers. Simulators do not teach you about noise, calibration drift, cross-talk between qubits, the quirks of specific vendor stacks, or the operational realities of cryogenic systems. The only way to learn these things is to operate real hardware, and the organizations that start now will have five to ten years of accumulated operational knowledge when the technology reaches commercial relevance.

Hardware-in-the-Loop Development Is Not a Luxury

Quantum software development against simulators hits a wall. Classical quantum simulators can handle roughly 40-50 qubits before they become intractable, and even within that range, they model idealized noise rather than the messy reality of physical devices. Algorithms that perform beautifully on simulators can fail catastrophically on real hardware due to device-specific noise profiles, calibration errors, and connectivity constraints.

Organizations building quantum applications for chemistry, materials science, finance, or logistics need access to real hardware to develop the hybrid quantum-classical workflows that will form the basis of early practical applications. This is particularly true for error correction research: the quantum error correction revolution that is compressing timelines faster than hardware roadmaps anticipated depends on experiments running on physical devices, not simulations.

Companies like Amgen, BMW, JPMorganChase, and SoftBank are already running workloads on Quantinuum’s Helios system. They are not doing this because Helios delivers quantum advantage for their problems. They are doing it because understanding real-device performance, developing application-specific circuits, and calibrating resource estimates against actual hardware behavior will give them a meaningful head start when the machines become capable enough to matter.

Anchor-Customer Economics

Many government quantum purchases operate as anchor-customer procurement, a model borrowed from defense contracting and aerospace. The government buys the first units not because it needs the immediate capability, but to seed a domestic vendor base, provide revenue that allows companies to survive the valley between research prototype and commercial product, and signal to private investors that the market is real.

The UK’s ProQure programme is explicit about this. Its first phase funds companies up to £14 million each to develop, build, and validate integrated quantum computing systems. The programme is designed to “inform a future public procurement of large-scale quantum computers beyond 2030” while supporting “the growth of enterprise operations and industrial capabilities.” The Japanese model is similar: METI’s backing of Fujitsu, KDDI, and quantum startups is as much about creating a domestically competitive industry as it is about the machines themselves.

This logic makes the spending look very different than the simplistic narrative of “governments buying useless computers.” They are investing in the industrial capacity to produce quantum computers, not merely buying what currently exists.

The Security Dimension Is Unique

Unlike almost every other emerging technology, quantum computing carries an asymmetric security implication that makes early engagement a matter of risk management rather than opportunity capture.

The Harvest Now, Decrypt Later (HNDL) threat is active today. Nation-state actors are already collecting encrypted data with the expectation of decrypting it when cryptographically relevant quantum computers (CRQCs) become available. For organizations that handle data with long confidentiality lifetimes (government, defense, banking, healthcare, intellectual property), the quantum threat timeline is not “when will a CRQC arrive?” but “how long does my data need to stay secret?”

This creates a rational basis for acquiring quantum hardware specifically to conduct empirical research on how close current systems are to cryptographic relevance. Running small-scale versions of Shor’s algorithm, testing error correction codes that would underpin a future CRQC, and calibrating resource estimates against real device performance all feed directly into the risk assessment that drives PQC migration decisions.

Several government purchasers I am aware of cite this as a primary justification: understanding the threat timeline from the inside, with access to the same class of hardware that adversaries are developing, rather than depending on external assessments that may be years out of date.

As I have argued extensively, the debate over when Q-Day arrives is increasingly academic. Regulators, insurers, investors, and customers are setting their own quantum deadlines. But for the subset of buyers who are defense and intelligence agencies, empirical access to quantum hardware remains a direct input to national security planning.

Ecosystem Access and Standards Influence

Owning and operating quantum hardware grants a seat at tables that are otherwise inaccessible. Participation in vendor advisory boards, national quantum research centers, standards bodies, and multi-stakeholder consortia is materially easier for organizations that have deployed hardware than for those with only cloud access agreements.

This matters because quantum computing standards, from error correction benchmarks to application programming interfaces to interconnect protocols, are being set now. The organizations that shape those standards will have structural advantages when the technology matures. Those that show up late will adopt standards designed around someone else’s architecture and someone else’s use cases.

The Competitive Fear Factor

This is the reason clients mention most often, and it is the hardest to dismiss: the fear of being left behind.

At a few hundred logical qubits, quantum computers will begin delivering genuine advantage in specific domains, particularly molecular simulation, catalyst design, and certain optimization problems. The organizations that have spent years learning how to formulate problems for quantum hardware, developing hybrid workflows, and training their teams will be able to extract value from those machines immediately. Their competitors who waited will be starting from scratch. At that scale, quantum computing creates competitive advantage.

At a few thousand logical qubits, the calculus changes from advantage to something closer to competitive extinction. In industries where quantum simulation can design better drugs, better catalysts, better battery chemistries, or better materials, the organizations with operational quantum capabilities will not merely be ahead. They will be operating in a different computational regime entirely. Their competitors without quantum access will not be able to close the gap through incremental classical improvements, because the problems quantum computers solve at that scale are ones that classical machines cannot solve at all, regardless of how much hardware you throw at them.

This progression from advantage to extinction is what makes the FOMO rational rather than emotional. A pharmaceutical company that waits until fault-tolerant quantum computers are commercially proven before beginning its quantum journey will find that its competitors have already used those machines to explore chemical spaces that classical simulation cannot reach. The resulting drugs, catalysts, or materials will represent a structural lead that cannot be replicated without equivalent quantum capability, equivalent algorithmic expertise, and equivalent operational experience. None of those can be acquired overnight.

The executives I speak with through Applied Quantum understand this arithmetic even when they fully accept my assessment that current machines deliver no practical advantage. They are not buying today’s quantum computer for today’s problems. They are buying the learning curve, the vendor relationships, and the institutional muscle memory that will determine whether they can exploit tomorrow’s quantum computer on the day it arrives, or whether they’ll spend three to five years catching up while their competitors pull away.

What the Buying Market Actually Looks Like

The quantum computer market in 2026 is segmented in ways that the headline numbers obscure. The vast majority of quantum computing revenue comes from cloud access, not hardware sales. IonQ’s $130 million in 2025 revenue was driven primarily by its cloud platform and government contracts, with the Forte Enterprise system available for on-premise data center deployment at multi-million-dollar price points. Quantinuum offers Helios through both cloud and on-premise arrangements, with early adopters in pharma, finance, and national laboratories.

Government buyers represent the largest single segment, followed by research institutions and then a growing number of private-sector enterprises. The QED-C report identified 7,418 quantum-engaged organizations worldwide at the end of 2025, including 556 pure-play quantum companies, a figure that understates the breadth of organizational engagement with quantum technology.

The pricing structure reflects the pre-commercial nature of the market. Cloud access ranges from fractions of a cent per shot on basic systems to $135,000 or more per month for premium trapped-ion access on Quantinuum’s platform. On-premise systems carry multi-million-dollar price tags plus significant facilities and operating costs. These are not prices that can be justified by near-term computational returns. They can only be justified by the strategic, institutional, and security considerations outlined above.

The Honest Assessment

Not all of these motivations carry equal analytical weight, but more of them are legitimate than a skeptic might assume.

Take the PR and first-mover signaling that many buyers openly acknowledge. Being the first company in your industry or the first in your country to deploy a quantum computer generates real business value: media coverage, client confidence, talent magnetism, board-level credibility on technology strategy. In regulated industries like banking, pharmaceuticals, and defense contracting, the perception of technological leadership directly affects client retention and contract competitiveness. Several of my clients have been candid that the press release announcing their quantum capability was part of the business case, and that the reputational return exceeded the cost of the hardware in its first year.

This is marketing value, not computational value, and it should be recognized as such. But dismissing it as vanity misses the point. In competitive markets, perception shapes capital allocation, talent flows, and partnership opportunities. The company that announces a quantum computing program attracts the researchers, the co-development deals, and the investor attention that reinforces the investment. The company that waits gets none of those compounding benefits. Whether that feedback loop is “real” value or “just” PR depends on your definition. The executives writing these checks do not find the distinction meaningful.

Similarly, some government purchases are driven as much by political positioning as strategic necessity: a minister announcing a quantum program generates favorable coverage and positions a country in the “quantum race” narrative. But even the most politically motivated procurement creates real infrastructure, trains real engineers, and builds real vendor relationships that outlast the news cycle.

But the core of the buying activity, the procurement-cycle preparation, the sovereign capability building, the workforce development, the hardware-in-the-loop R&D, the security research, and the anchor-customer economics, represents a rational set of bets on a technology whose timeline is uncertain but whose arrival is not.

As I mapped in the Quantum Utility Map capstone, quantum computing will transform five industries and disappoint twenty. The organizations buying quantum hardware today are, in many cases, the ones positioned in those five industries. Pharmaceutical companies running molecular simulations, materials science labs exploring catalyst design, financial institutions stress-testing portfolio optimization, and defense agencies evaluating cryptographic resilience all have domain-specific reasons to build quantum muscle now.