Marin Ivezic
A rule change, not a headline
On January 2, 2025, the United States Treasury’s outbound investment rules took effect. From that date, an American investor putting money into a Chinese company that builds quantum computers, quantum sensors or quantum networking equipment either has to notify the government or is barred from the transaction outright, depending on what the company does. Outbound investment screening is the newer cousin of export control: instead of restricting what leaves the country as a product, it restricts what leaves as capital.
No qubit record was broken that week. Nothing was demonstrated in a laboratory. A category of investment simply became paperwork, and in some cases became illegal.
That is how quantum geopolitics usually reaches an organisation. It arrives as a licence application, a clause in a procurement template, a question from a client’s security team, or a deadline published by a national cyber security agency. The satellites and the qubit counts are the part that gets reported. The part that changes what you can buy, from whom, and by when, tends to be a fifteen-page rule with a comment period.
This guide is written for people who have to make decisions with that second category. It covers what the major national programs are actually doing, how to read them without being misled by announcement figures, and the four channels through which state competition becomes an operational constraint on ordinary organisations.
What the headline numbers hide
Almost every comparison of national quantum programs is a comparison of announcement figures, and announcement figures are close to useless for planning. We suggest four questions instead.
What is the money, exactly? National quantum funding gets quoted as a single number, and that number can mean an authorisation (a legislature permitting spending), an appropriation (money actually allocated), a multi-year envelope collapsed into one figure, or an analyst’s estimate assembled from partial disclosures. The widely quoted figure of roughly fifteen billion dollars for China’s public quantum investment falls into the last category. It comes from consultancy estimates rather than a published budget line, and it is repeated so often that its origin has been lost. That doesn’t make it wrong. It does mean you should never build a competitive argument on it, and you should be sceptical of anyone who does.
Who ends up owning the hardware? A program that funds a state laboratory produces a sovereign capability that citizens and companies may never touch. A program that subsidises cloud access produces a user base with no hardware. A program that funds domestic manufacturers produces an export industry, and an export industry eventually produces export controls. These three shapes have very different consequences for anyone trying to buy quantum capability five years from now.
What is being restricted, and who agreed to the restriction? Controls imposed by one country can be routed around. Controls adopted by several countries at once cannot. The distinction between a unilateral rule and a coordinated one is the single most practically useful thing to know about any export announcement.
What does the program require of buyers rather than builders? This is the question most reporting skips and the one that produces your deadlines. A national quantum strategy that also sets a cryptographic migration schedule for government suppliers has just written a line into your project plan, whether or not you ever buy a quantum computer.
A cryptographically relevant quantum computer, usually abbreviated to CRQC, is a machine large and reliable enough to break the public-key cryptography that currently protects internet traffic, payments, software updates and secure messaging. No such machine exists in public. Every government timeline discussed below is a bet on when one might.
The two poles
The United States runs a market-led program with federal coordination. The National Quantum Initiative Act, signed in December 2018, authorised roughly $1.2 billion USD over five years and set up research centres across the Department of Energy, the National Science Foundation and the National Institute of Standards and Technology (NIST). The important detail is proportion: American public funding is modest relative to what IBM, Google, Microsoft and a long list of venture-backed firms spend on the same problem. The state’s most consequential interventions have not been cheques at all. They have been standards and restrictions.
NIST ran the competition that produced the first post-quantum cryptography standards, published in August 2024 as FIPS 203, 204 and 205, specifying ML-KEM (previously CRYSTALS-Kyber), ML-DSA (previously CRYSTALS-Dilithium) and SLH-DSA (previously SPHINCS+). A fourth algorithm, FN-DSA, was selected but had not been finalised alongside the first three. Those documents now govern cryptographic procurement in a great many countries that had no part in writing them, which is a fair description of how American technological power actually operates.
China runs a state-directed program of a different character. Its most cited achievements are real and verifiable: the Micius satellite, launched in 2016, demonstrated quantum key distribution between space and ground stations; the Zuchongzhi superconducting processor and the Jiuzhang photonic experiments, both from the University of Science and Technology of China under Pan Jianwei, produced sampling results comparable to the American claims of the same period; Origin Quantum brought a 72-qubit superconducting system online in 2024.
Two features distinguish the Chinese effort. First, quantum communication receives investment on a scale seen nowhere else, reflecting a judgement that hardware-based key distribution is worth building at national scale. Second, the program is designed for self-sufficiency, partly by choice and increasingly by necessity, since several Chinese quantum firms and institutes have been added to the United States Entity List, the trade blacklist that requires American exporters to obtain a licence before shipping to a named organisation.
The two poles are frequently described as a race, and the framing does some damage. A race implies a single finish line and a single winner. What the evidence supports is something less tidy: two large programs with different strengths, different definitions of success, and an increasingly small overlap in the components they can legally buy from each other.
The coordinators
The European Union competes through coordination rather than concentration. The Quantum Flagship, launched in 2018, committed €1 billion over ten years across computing, communication, simulation and sensing. EuroHPC has been installing quantum systems alongside supercomputers at sites in Czechia, France, Germany, Italy, Poland and Spain, and EuroQCI is building a secure communication infrastructure spanning member states. In July 2025 the Commission published a Quantum Europe Strategy, signalling a Quantum Act to follow, with the explicit aim of pulling national programs into something that behaves like one program.
Individual member states fund at serious levels. Germany and France have each announced multi-billion-euro national plans, and the Netherlands, Austria, Finland and Italy run substantial programs of their own. The persistent European difficulty is not money or science. It is that twenty-seven procurement regimes, twenty-seven research funding bodies and twenty-seven security agencies produce coordination costs that a single national program does not pay.
Europe also moves earliest on rules. EU member states agreed a coordinated post-quantum migration roadmap in June 2025, with high-risk systems to be migrated by the end of 2030 and the bulk of the work finished by 2035. If you sell into European critical infrastructure, that roadmap is a more immediate commercial fact than anything happening in a European laboratory.
The United Kingdom started earlier than almost anyone. The National Quantum Technologies Programme began in 2014 with an initial £270 million and built a network of research hubs that seeded most of the country’s quantum companies. The National Quantum Strategy, published in March 2023, committed £2.5 billion over ten years. Britain also participates in AUKUS Pillar 2, which covers quantum among other advanced capabilities, and which produced something more useful than a research budget: on May 1, 2024 the United States brought into force defence trade exemptions for Australia and the United Kingdom, removing licence requirements for a large class of transfers among the three. The relief is not total, and a list of sensitive items remains outside it. Export control relief between allies is a form of industrial policy, and it is one of the few forms that costs nothing.
The National Cyber Security Centre has published a migration timeline of its own, with discovery and planning expected by 2028, the highest-priority migrations by 2031, and completion by 2035.
The middle tier
India approved its National Quantum Mission in April 2023, with ₹6,003.65 crore (roughly $730 million USD) allocated across eight years and four thematic hubs covering computing, communication, sensing and materials. The stated targets include intermediate-scale quantum computers and satellite-based key distribution over long domestic baselines. Whether those targets are met on schedule is less consequential than the capacity being built underneath them.
Canada launched a National Quantum Strategy in January 2023 with C$360 million, a figure that looks small next to the poles and considerably less small relative to the size of the domestic research community it supports.
Russia has a state quantum program run largely through Rosatom and a small number of institutes, with reported results including a 50-qubit neutral-atom prototype in 2024. Its structural problem is procurement. Dilution refrigerators, ultra-stable lasers, specialised control electronics and high-purity materials come overwhelmingly from suppliers in countries that will not sell to Russia, and no amount of domestic funding fixes that quickly.
The pattern across the middle tier is worth noticing. None of these countries is attempting to build every layer of the stack. Each is choosing a subset, and the choices reveal what each government believes it can defend: research capacity, a specific hardware modality, secure communication infrastructure, or simply enough expertise to be an informed buyer. For most organisations, and for most countries, being an informed buyer is the achievable goal.
Four channels to your desk
The part most likely to appear in a contract you sign this year is also the least dramatic.
Export controls and investment screening
Quantum technologies are now treated as dual-use, meaning goods with both civilian and military applications, which places them inside the same control machinery that governs encryption software, advanced semiconductors and machine tools. In September 2024, the United States Bureau of Industry and Security issued an interim final rule imposing controls on quantum computing items, and did so in deliberate alignment with several allied countries that had adopted or were adopting equivalent measures. Coordination was the point. Unilateral controls on a component that three countries manufacture achieve very little.
This coordination happened outside the traditional multilateral forum. The Wassenaar Arrangement includes Russia, which makes consensus on controls aimed partly at Russia unworkable, so the like-minded states simply moved without it. Anyone forecasting future controls should watch the coalitions, not the treaty bodies.
The clearest precedent is American and about thirty years old: through the 1990s, the United States regulated strong cryptography as a munition. Exportable software shipped with deliberately weakened key lengths, and the resulting split between domestic and export builds shaped commercial products long after the controls were relaxed in 2000. Two consequences outlived the policy. The weakened cipher suites stayed in codebases for years and turned up as exploitable flaws more than a decade later. And the controls pushed a generation of cryptographic engineering offshore, which is one reason the field is as internationally distributed as it is today. Export controls do not simply slow an adversary. They redirect where capability accumulates, and the redirection lasts longer than the rule.
For an organisation, the practical effects are ordinary. Research collaborations across borders acquire a compliance review. Cloud access to quantum hardware may carry nationality restrictions on who can use the account. Acquisition and investment activity in the sector triggers screening in several jurisdictions at once. Hiring a specialist who is a national of a controlled country may require a licence in the country where the work happens. None of this is exotic. All of it takes time that project plans rarely include.
Standards are diverging faster than hardware
Post-quantum cryptography, or PQC, means classical algorithms designed to resist attack by a future quantum computer. It runs on ordinary hardware and it is available now. QKD, or quantum key distribution, is the different approach: purpose-built hardware and dedicated optical links that distribute keys using the physics of single photons.
Governments disagree about the second one, and the disagreement is not subtle. China has invested heavily in QKD infrastructure and the European Union is building EuroQCI. The United States National Security Agency and the United Kingdom’s National Cyber Security Centre have both published guidance advising against QKD for securing national security systems, on the grounds that it protects only one link in a chain, requires trusted intermediate nodes over distance, and introduces hardware you then have to secure. Two blocs, opposite conclusions, identical physics.
The divergence extends into PQC deployment. Germany’s Bundesamt für Sicherheit in der Informationstechnik (BSI) and France’s Agence nationale de la sécurité des systèmes d’information (ANSSI) both favour hybrid deployment, meaning a classical algorithm and a post-quantum algorithm used together so that a weakness in the newer scheme does not by itself break the connection, and ANSSI has been explicit about expecting hybrid deployment for years to come. NSA guidance under the Commercial National Security Algorithm Suite 2.0 (CNSA 2.0) sets a different path, with milestone dates through the early 2030s and exclusive use of the approved post-quantum algorithms in national security systems by 2033. China has run its own algorithm selection through domestic cryptographic bodies rather than adopting the NIST set.
If you operate in one jurisdiction, pick the local guidance and follow it. If you operate in several, you are going to be running more than one cryptographic configuration for a decade, and the cost of that is a design decision to make deliberately now rather than an accident to discover during an audit. This is the practical case for crypto-agility: build systems where the algorithm is a configuration choice rather than something welded into the protocol, because you will be changing it more than once.
Supply chains are narrower than the map suggests
The geopolitical map shows a dozen countries with quantum programs. The supply chain shows something considerably more concentrated.
Superconducting quantum processors need dilution refrigerators, which are the machines that cool a chip to a few thousandths of a degree above absolute zero. Two suppliers dominate the market, Bluefors in Finland and Oxford Instruments in the United Kingdom, with a small number of others competing at the edges. Those refrigerators run on helium-3, an isotope produced almost entirely as a decay product of tritium in nuclear weapons stockpiles, which means the supply is controlled by a handful of governments and has been rationed before. Trapped-ion and neutral-atom systems depend on ultra-stable lasers and optical components from a similarly short list of specialist manufacturers. Control electronics, cryogenic cabling and specialist test equipment follow the same pattern.
A country can fund quantum research generously and still be one export decision away from a stalled program. So can a company. The question to put to a hardware vendor is not where the company is headquartered. It is where the refrigerator comes from, where the isotope comes from, and what happens to the delivery schedule if either origin becomes subject to a licence requirement. Vendors who have thought about this have an answer ready. The answer itself is informative either way.
Talent and research security
Every national program identifies the same shortage, and none of them can buy their way out of it quickly, because the constraint is the number of people with several years of relevant training rather than the number of funded positions.
Two things follow. Competition for experienced people is intense and will stay that way, which raises the value of building internal capability rather than planning to hire it. And the movement of those people is being watched more closely than it used to be. Visa scrutiny in quantum-adjacent fields has tightened in several countries, university collaborations now come with disclosure requirements, and national laboratories apply vetting that would have seemed excessive ten years ago. Research security has become a line item, and organisations doing joint work with foreign institutions should expect to document it.
The Q-Day question, deflated
Most writing on quantum geopolitics reaches for a scenario: one country builds a CRQC first, reads everyone’s traffic, and the balance of power shifts overnight. It is a memorable story and a poor planning assumption, for three reasons.
A CRQC does not arrive as a switch being thrown. Fault-tolerant machines get built incrementally, and the engineering milestones along the way are visible in the published literature and in what suppliers are being asked to deliver. The intelligence services of the major powers watch each other’s progress closely enough that a genuine surprise would be remarkable.
The defence is already standardised and already deployable. The FIPS documents exist. Implementations exist. The remaining work is discovery, planning and migration across estates that nobody has fully inventoried, which is difficult, expensive and unglamorous, but is not blocked on any scientific unknown.
And the real exposure is not future traffic. It is data captured today and stored against a future capability, a practice usually called harvest now, decrypt later, or HNDL. That exposure is a function of one thing you can calculate without any view on quantum timelines: how long your data needs to stay confidential. Encrypted records with a twenty-five-year confidentiality requirement, moving across networks today, are exposed under almost any plausible timeline. Session data that becomes worthless in a week is not. The migration priority follows from that calculation, and it does not require anyone to guess a date.
We put this plainly because the alternative is worse in both directions. Fear-driven urgency produces panic buying and abandoned projects. Dismissal produces organisations that start their inventory work in 2032 with a 2035 obligation. The regulatory deadlines are the honest driver: they are published, they are dated, and they will be enforced by auditors and clients regardless of where the physics stands.
Reading the map from where you sit
For most organisations, the useful posture is neither participation in the race nor indifference to it. It is informed dependence, managed deliberately.
Know which jurisdictions write your rules. If you sell to European critical infrastructure operators, the 2030 and 2035 dates in the EU roadmap are yours. If you supply the American federal government, the 2035 target from the 2022 national security memorandum and the CNSA 2.0 milestones apply to you through your customers. Most multinationals inherit several sets of dates and should reconcile them on one page.
Treat cryptographic inventory as the first deliverable. Nothing else can be scheduled until you know where cryptography appears in your estate, which algorithms and key lengths, in which products, under whose support contract. This is the work that always takes longer than the estimate.
Ask supply chain questions of quantum vendors, not just capability questions. Component origin, isotope dependency, licence exposure, and what the vendor’s own contingency looks like.
Build a small internal capability that can read the primary sources. Two or three people who can distinguish an authorisation from an appropriation, physical qubits from logical qubits, an announcement from a demonstration, and a unilateral control from a coordinated one. That capacity is what turns geopolitical news into a decision instead of an anxiety.
Assume the rules will change again. Every element of the current arrangement is younger than five years. Design for the version after this one.
Where to build the capability
The competition between national quantum programs is not something most organisations can influence. What they can control is whether they read it accurately, and whether the people making procurement, architecture and compliance decisions understand quantum technology well enough to tell a constraint from a headline.
That understanding is trainable, and it does not require a physics background. It requires knowing what the technology can and cannot do, what the standards actually say, how the control regimes work, and where the supply chain narrows. Our programs at Quantum Academy are built for exactly that audience: professionals who need working fluency rather than research-level depth, and who have to make decisions before the science settles.
For the technical detail behind the migration work described here, PostQuantum.com carries deeper treatments of the cryptographic side, and pqcframework.org sets out migration methodology in a form you can hand to a project team.