Somewhere, a hard drive is filling up with secrets no one can read yet, and its owner is waiting for the machine that opens them all at once. On March 30, 2026, that wait got shorter.

Two independent research teams lowered the public estimates for breaking the encryption that secures banking, communications, and classified traffic. One showed that Shor’s algorithm, the quantum method for factoring the large numbers behind public-key encryption, could run at cryptographically relevant scale with as few as 10,000 reconfigurable neutral-atom qubits. Google Quantum AI, with the Ethereum Foundation and Stanford, put the threshold for breaking elliptic-curve encryption below 500,000 physical qubits on a superconducting machine. Google is planning as if the estimates will keep falling. Days before the papers appeared, it set a deadline to move its products onto quantum-resistant encryption by 2029.

Skeptics have long called quantum computing decades away. Nvidia’s Jensen Huang told analysts in January 2025 that very useful quantum machines were 15 to 30 years away. By March he was walking that back.  Within a year, Google had unveiled its Willow processor, Amazon its Ocelot, IBM its Nighthawk and Loon, and Quantinuum, now on the Nasdaq, delivered Helios, a machine already running error-corrected logical qubits, the reliable building blocks a full-scale computer needs. Road maps that were once vague now come with dates, climbing toward the million-qubit scale most vendors assume a fault-tolerant machine will need to run long calculations without errors piling up. The most detailed of them aims at fault tolerance before the decade is out, backed by IBM’s commitment of more than $10 billion.

Yet the United States is preparing for the quantum era one office at a time, while the threat is arriving all at once. Cryptography, artificial intelligence, networks, sensing, and the supply chain are being handled as five separate problems on five separate schedules. They share one deadline and one prize: “Quantum First,” before China. Responsibility should sit with the National Security Council, with the first year of work inexpensive and concrete, and allies offered a stake in the science.

Sign Up for Our Newsletter

A Post-Quantum Future Will Arrive

Quantum policy has to deal with at least five fronts at once. So far, the government has answered each separately. Each front follows the same pattern: a real problem, some policy on the books, and a gap.

Cryptography

A quantum computer big enough to run Shor’s algorithm at scale would break the public-key encryption that secures most of digital life, including online banking, software updates, virtual private networks, and much government and military traffic. That is the problem “Q-Day” names: the day such a machine exists and everything encrypted under the old assumptions becomes readable.

The government is not ignoring it. The National Security Agency, the Cybersecurity and Infrastructure Security Agency, and the National Institute of Standards and Technology have issued joint guidance telling critical infrastructure owners to inventory their systems and press their vendors. The standards institute finalized its post-quantum encryption standards in 2024, and a June 22, 2026, executive order formalized them into deadlines. The Group of Seven’s Cyber Expert Group has made a similar case for financial systems. The gap is coverage: The deadlines bind federal systems, but the banks, utilities, and hospitals running the same vulnerable encryption are only being asked politely.

Q-Day is not necessarily imminent, and one can disagree about how and when. The case for moving now rests on lead times. Cryptographic migrations historically take a decade or more, and the slowest systems to change are the ones that matter most, such as weapons platforms, industrial controls, and satellites. If the machine arrives in 2035, the migration still has to start now.

Artificial Intelligence

Artificial intelligence makes the quantum cryptography problem meaner. Intelligence services have long collected material they could not yet read, and “harvest now, decrypt later” applies that old bet to encrypted traffic. AI sharpens it twice over, as machine-speed discovery shortens the time between a flaw and its exploitation, and machine triage raises the value of clinical, laboratory, and industrial data stolen today.

This spring showed how fast that threat is moving. In April 2026, Anthropic disclosed that its Claude Mythos Preview model had autonomously surfaced thousands of zero-day vulnerabilities. On June 12, days after the finished model launched, the Commerce Department used national-security export controls to bar transfers of the company’s two most capable models to any foreign national, and Anthropic shut them off worldwide to comply. Commerce lifted the controls on June 30, after the company agreed to stronger safeguards and expanded pre-release government testing.

Policy is beginning to reach this front too. A June 2, 2026, executive order on artificial intelligence pushes agencies toward AI-assisted cyber defense and opens a voluntary channel for government review of frontier models before release. It does not yet connect the AI file to the quantum file, even though the data harvested today will be decrypted by one technology and sorted by the other.

Quantum Networks

China is building quantum networks faster than anyone else, from the Beijing-Shanghai backbone it has run for years to recent work on quantum repeaters and device-independent key distribution over more than 100 kilometers, key exchange that stays secure even on untrusted hardware. What that build-out buys Beijing depends on a distinction policy documents tend to blur.

Quantum communication is point-to-point: the transfer of quantum states, or of an encryption key, between two endpoints. Quantum key distribution, its best-known product, is a hardware overlay the National Security Agency calls a partial solution, dependent on special equipment and rarely worth its cost. Quantum networking is the larger architecture that ties many such links together through entanglement, the quantum bond between particles whose states stay correlated at any distance, using repeaters, routing, and control planes.

For a war planner, the prize is the network. Assured command and control, networked sensing, and resilient timing all depend on a connected web of many links, but a single key-distribution device provides none of them. That web is what China is building: a carrier-grade network of more than 12,000 kilometers of fiber, 145 backbone nodes, and 20 metropolitan networks across 80 cities, with an integrated space-to-ground network written into the new five-year plan.

This is the front where American policy has the least to show. A June 22, 2026, executive order on quantum innovation requires five-year plans for sensing and networking, a start, but there is no networking equivalent of the cryptography migration timetable.

Sensing and the X-Ray City

The quantum sensing threat is easy to oversell, but it is already here. Gravimeters and magnetometers can register voids, tunnels, mineral deposits, submarines, and magnetic signatures. A British Ministry of Defence-backed team located a buried tunnel from the surface, at sub-meter resolution, outdoors. The Defense Advanced Research Projects Agency’s Robust Quantum Sensors program is moving such devices off the laboratory bench and onto military platforms, and its tensor-magnetometer effort aims to fix a magnetic object from a single measurement.

Both the Congressional Research Service and the Center for Strategic and International Studies treat quantum sensing as a defense problem. So far, the policy response is thin: a five-year sensing plan and a directive to field three next-generation quantum sensors by 2028, both under the innovation order. No document yet addresses what happens if these sensors shrink, network together, and fuse with artificial intelligence into a surveillance layer. In the future, a drone-borne magnetometer above a street protest could map vehicles, weapons, cabling, and movement, then fuse those signatures with video and phone-location data: an “X-Ray City.” The same sensor families are already being tested for brain imaging in ordinary rooms.  While those applications are speculative, they belong in any sensing rulebook written now.

The Supply Chain

Quantum systems rest on narrow, fragile inputs: dilution refrigeration, photonics, specialized lasers, enriched isotopes, helium-3, and single-photon detectors. These keep qubits cold, shielded, and readable. For several, only a handful of suppliers exist worldwide. No country builds this chain alone: the parts are too many and too specialized for one industrial base. To defense planners they are readiness inputs: Army navigation when GPS is denied, Navy undersea missions, and Air Force timing and electronic warfare all draw on the same short list.

In a Taiwan Strait contingency, those dependencies turn strategic. Deterrence by denial rests on convincing Beijing that an invasion would fail, a case built, in part, from capabilities: quantum-secure command and control, targeting when navigation satellites are jammed, and communications among the United States, Japan, South Korea, and Australia that stay up under attack. If Beijing moves faster in sensing, networking, or timing, the anti-access problem gets harder and allied decision-making gets more brittle.

The policy response has a first step. On June 30, 2026, the National Security Agency’s Laboratory for Physical Sciences and the Army Research Office launched the QuantumEAGLe initiative, with the domestic supply chain for specialized quantum components among its priorities. What is missing is visibility: a supply-chain dashboard showing which inputs are concentrated, which substitutions are plausible, and which stockpiles deserve attention sooner than the United States Geological Survey cycle. Domestic supply of helium-3, cryogenics, and detector-grade materials will take years to build, and policymakers should say so plainly.

The China Threat and Existing Policy

China belongs in this argument because it treats quantum as strategic infrastructure, aligning its labs, universities, and state firms through military-civil fusion. The 15th Five-Year Plan, adopted in March 2026, lists quantum first among seven future industries, and a national fund has placed roughly $17.5 billion behind three regional quantum funds. On March 31, Beijing issued Order Number 834 on supply-chain security, tying together export controls, investment screening, and data security. Whether or not Beijing breaks Western cryptography first, it is building for a lasting advantage.

Washington has broadened its effort beyond cryptography. The United States-China Economic and Security Review Commission has pressed for a “Quantum First” goal by 2030: strategically significant quantum capabilities, fielded before China gets there. The June 2025 cybersecurity executive order kept the core post-quantum steps while paring back standards diplomacy, and the June 2026 AI order added AI-assisted cyber defense. Then, on June 22, 2026, came the dedicated cryptography executive order: post-quantum key exchange on high-value federal systems by the end of 2030 and digital signatures by the end of 2031, a standards-institute pilot by 2027, a cryptographic bill of materials from every agency, and a State Department push for allied adoption, reversing the 2025 retreat. The pieces are on the board, but someone has to play them as one game and give allies a reason to join.

The Solution

One potential solution is naming the National Security Council to lead on quantum because it is best positioned to settle disputes when cyber resilience, export controls, allied diplomacy, and military planning collide. The president should name a senior director for quantum readiness with a one-year mandate, quarterly reviews at the deputy level, and an escalation path to the principals. The Department of Defense should hold its own networked programs to the same standard: Joint All-Domain Command and Control and Replicator, now in its second phase fielding counter-drone defenses, run on encrypted links, and links recorded today can be read the day the machine arrives.

The first year is relatively inexpensive, once scoped. Program management, procurement clauses, a supply-chain dashboard, and allied coordination should run in the tens of millions of dollars. Heavier follow-on investments in cryogenics, photonics, helium, and foundries belong in industrial-base policy. Plausible sources include the National Quantum Initiative reauthorization, Department of Energy laboratory lines, Defense Department research accounts, federal cyber-modernization funds, and the roughly $2 billion in Commerce letters of intent already extended to quantum firms.

The United States should not try to match China dollar for dollar, but should instead spend where public money unlocks private scale and allied adoption. The July 7 White House quantum summit said so out loud, casting the government as quantum’s first buyer: a chips-program funding line past $1 billion and a 2028 quantum computing goal framed as an open competition.

The National Security Council will also require support from other offices and agencies. The Office of Science and Technology Policy should help provide technical judgment. The Office of Management and Budget can help turn its decisions into necessary procurement documents, like cryptographic bills of materials, vendor migration road maps, and red-team drills that assume an attacker armed with artificial intelligence.

On networks, the spending priorities follow from the difference between communication and networking. Post-quantum cryptography is software, runs on today’s networks, and should be the universal floor. Quantum key distribution should be treated as a specific hardware product. Of the two, quantum networking is the better bet for public money, because future military capabilities will likely run on that layer.

Abroad, the rule should be to keep technical interfaces open with allies and restrict the hard-to-build hardware. This means agreeing on standards for how quantum network nodes talk to one another, so allied forces can eventually connect their national networks. Restriction means export controls on the components that take years to master, including repeater hardware, testbeds, and access to quantum network services, the pieces an adversary needs to shortcut its build-out. Each control should pass one test: it should deny a concrete military capability, in the least trade-restrictive way available, and leave civilian science alone.

America should not run the quantum race alone. In practice, this means mutual recognition. The United States should accept European and Japanese certification where their testing is rigorous enough to lean on. America can work closely with allies and partners to adopt the National Institute of Standards and Technology’s post-quantum standards. The North Atlantic Treaty Organization would handle military interoperability, and for Europe, European Union institutions would handle civilian infrastructure and industrial policy.

Denial alone will not hold a coalition together. Democracies should also offer partners something worth joining. Quantum-accelerated climate modeling, drug discovery, materials science, and resilient energy are obvious candidates. The innovation order points this way at home, steering effort toward a science-enabling quantum computer at a Department of Energy facility, which the department aims to deploy by 2028 under the Genesis Mission, its program pairing national laboratories with frontier AI. Opened to trusted allies, that machine would be the strongest reason to join the American system and stay out of the Chinese one.

Quantum security is a cabinet-level problem because its lead times outrun the political cycle. The danger is quietly losing the race to “Quantum First” due to half-finished migration, fragile supply chains, undefined sensing rules, and network standards written in Beijing. One office in charge, working with our allies, is what Germans mean when they call something Chefsache: the boss’s problem, handled personally. The country that spends the lead time best will hold the advantage for decades.

Write for Cogs of War

Mauritz Kop is the founder of the Stanford Center for Responsible Quantum Technology (Stanford RQT), a Senior Fellow and Principal Investigator at the Centre for International Governance Innovation, and a Guest Professor at the United States Air Force Academy.

Joseph Federici is a Senior Policy Analyst on the Security and Foreign Affairs team at the United States-China Economic and Security Review Commission.

The views expressed are the authors’ own and do not necessarily reflect those of their institutions or the United States government.

Image: U.S. Army Combat Capabilities Development Command – Army Research Laboratory via DVIDS.

**Please note, as a matter of house style, War on the Rocks will not use a different name for the U.S. Department of Defense until and unless the name is changed by statute by the U.S. Congress.