The Cold War’s “Mutually Assured Destruction” no longer works. You fire a nuclear missile, the other side sees it in thirty minutes, retaliates, and those thirty minutes kept diplomacy alive. Hypersonic weapons shattered that arithmetic: five times the speed of sound, unpredictable trajectory, five to ten minutes of warning. Existing defense systems cannot see them. Cannot stop them. This is not the end of war. This is the ignition of a new arms race.
A hypersonic weapon travels above Mach 5, above 6,125 kilometers per hour. Two types exist. The Hypersonic Glide Vehicle launches via rocket carrier to 50 to 100 kilometers altitude, the motors cut, and the vehicle glides, not straight but maneuvering, shifting trajectory in real time. China’s DF-ZF and Russia’s Avangard belong here. The Hypersonic Cruise Missile uses a scramjet engine that inhales air at Mach 5 and burns it continuously, sustaining propulsion inside the atmosphere at 20 to 30 kilometers altitude. Russia’s Zircon and India’s BrahMos-II operate this way. A conventional ballistic missile draws a parabolic arc and radar reads that arc, calculates the endpoint, and sends an interceptor. A hypersonic weapon maneuvers: climbs, drops, changes direction. Radar cannot answer the only question that matters: where is it going?
The idea of hypersonic flight is not new, it dates to the 1960s. Three barriers blocked it. Materials science solved the first: at Mach 5 and above, aerodynamic friction generates temperatures exceeding 2,000 degrees Celsius. Conventional metals melt. Ultra-high temperature ceramics hold. The scramjet engine solved the second: supersonic combustion was theoretical in the 1990s and is now operational. Computing power solved the third: hypersonic airflow is chaotic and previously unsimulatable, but AI-assisted computational fluid dynamics now makes the design tractable. Remove three barriers and the weapon exists. That is what happened.
Russia leads, and it proved this on an active battlefield. Avangard has been operational since 2019 at Mach 27 with nuclear capability. The Kinzhal has flown combat missions in Ukraine since 2022, striking command centers, ammunition depots, and energy infrastructure. Russia claims above 80 percent strike accuracy. Ukraine has shot down some Kinzhals with Patriot systems, but barely, and with enormous effort. The Zircon naval missile reaches 1,000 kilometers at Mach 9. China shocked Western intelligence in 2021 with a fractional orbital bombardment test: a vehicle entered orbit, then executed a hypersonic dive. Nobody predicted it. The DF-ZF has been operational since 2020. The United States is behind. The AGM-183A ARRW program failed in testing and was cancelled in 2023. The HAWC logged successful tests but remains non-operational. The Dark Eagle is still in development. The Army, Navy, and Air Force each run separate programs, fragmenting the budget, the precise opposite of Russia and China’s centralized approach.
Defense is not possible, or more precisely, it is possible only in the weakest sense of the word. The Patriot can intercept a straight-trajectory Mach 5 target. It cannot track a maneuvering hypersonic vehicle. THAAD operates at high altitude against ballistic missiles and hypersonic glide vehicles fly below its effective envelope. The S-400 and S-500 are theoretically the most capable systems on earth, but Russia has never tested them against its own hypersonic weapons. Best-case intercept probability sits at 50 to 60 percent. Three compounding problems make even that figure optimistic. Radar detects the low-altitude vehicle late. The trajectory cannot be projected. The interceptor cannot close the distance in time. Three solutions are under development. Laser weapons move at the speed of light and could theoretically catch anything, but their range tops out at 50 to 100 kilometers and their energy requirements are enormous. The U.S. Navy’s HELIOS system targets 2025 operational status. Low Earth orbit satellite constellations can detect a launch in real time, and the U.S. Hypersonic and Ballistic Tracking Space Sensor program plans its first launches in 2025. Machine learning attempts to predict hypersonic trajectories and compress the decision window. Total defense remains impossible. But a 30 to 50 percent intercept rate still functions as deterrence, because that is how the mathematics of deterrence work.
The strategic consequences are structural, not incidental. A hypersonic weapon can carry a nuclear warhead. The adversary cannot determine whether the incoming vehicle is nuclear or conventional. That ambiguity makes a false alarm a potential trigger for nuclear retaliation. The Cold War gave decision-makers thirty minutes, time to wake the president, convene advisors, and deliberate. Hypersonic compresses that window to five minutes, realistically to two or three. No human decides in two minutes. Systems automate. An algorithm can start a war. China can threaten U.S. carrier groups in the Taiwan Strait in real time. Russia can strike NATO bases in Poland and Romania within minutes. The concept of American force projection requires a complete redefinition. And once the race begins, it generates its own fuel: every new hypersonic weapon demands new defense systems, which demand more hypersonic weapons. The 1960s nuclear race, replayed at higher speed.
No treaty covers this. The INF Treaty banned intermediate-range missiles but excluded hypersonic systems, and the United States withdrew from it entirely in 2019. New START limits strategic nuclear weapons but does not count hypersonic delivery vehicles, leaving the category legally ambiguous. No international hypersonic arms control agreement exists. None is plausible. The United States, Russia, and China each believe the technology is too decisive to surrender. When every major power holds that belief simultaneously, negotiations produce nothing.
The autonomous dimension is the most dangerous question of all. AI-integrated hypersonic missiles can select targets without human input. The scenario is not theoretical: a missile launches with the instruction to find and strike enemy vessels, the AI analyzes radar data and chooses the target, and no human participates in that choice. A civilian vessel gets struck. A friendly vessel gets struck. The identification system makes an error. The military logic is internally coherent: a human cannot decide in five minutes, so automation is necessary. The peace activist’s objection is also coherent: life-and-death decisions cannot be delegated to an algorithm. Both are correct. That is precisely what makes it an irresolvable tension, not a policy problem awaiting a clever solution.
The optimistic scenario has hypersonic weapons functioning as nuclear weapons did, as deterrents that are never used, eventually constrained by treaty, eventually countered by defense systems reaching above 50 percent intercept probability. The pessimistic scenario has a regional crisis, in Taiwan or elsewhere, triggering hypersonic use, a false alarm escalating to nuclear exchange, and the arms race spiraling until small states and eventually non-state actors acquire the technology. The realistic scenario is that hypersonic becomes the new normal, as ballistic missiles became normal in the 1960s, major powers construct a rough mutually assured hypersonic deterrence equilibrium, and the instability never fully resolves because the decision window never reopens. Technology does not change war. It accelerates war. And acceleration eliminates the margin for error. One wrong decision now produces catastrophe in minutes, not hours.
This is not legal advice. This is analysis.
THINKLINIC 
Leave a comment