An AI cyberattack could trigger a satellite apocalypse in the next 2 years. Are we prepared?

The margin for a catastrophic collision cascade in low Earth orbit has shrunk to 5.5 days. That is the finding of new research from Sarah Thiele and colleagues at Princeton, published on arXiv in December 2025, and it is one of the more alarming numbers to come out of orbital mechanics in recent years.

The researchers developed a metric they call the CRASH Clock: Collision Risk from the Accumulation of Space Hardware. It measures how long humanity would have before a catastrophic collision cascade becomes unavoidable if satellites lost their ability to maneuver or situational awareness collapsed. In 2018, that window was 164 days. Today it is 5.5 days. The collapse tracks the deployment of megaconstellations, primarily Starlink, which has placed thousands of satellites in low Earth orbit and generates the highest rates of collision avoidance maneuvering of any operator.

Let's be clear about what this means and what it does not mean. The 5.5-day figure is not a prediction that a catastrophe is imminent. It is a measurement of how little slack remains before a triggering event could cascade into sustained, unrecoverable collisions. A cascade requires two simultaneous failures: loss of maneuver capability and loss of situational awareness. Get both wrong at scale, and the window to respond collapses to days rather than months.

That context matters because Space.com published a story this week headlined: "An AI cyberattack could trigger a satellite apocalypse in the next 2 years." That framing is not supported by either paper cited in that article. Thiele et al. do not claim an AI cyberattack is imminent or planned. The CRASH Clock measures the gap between a manageable debris environment and something close to unrecoverable under worst-case conditions. The second paper, by Ernesto López-Morales of New Mexico State University, models satellite ransomware economics using game theory. It is a separate analysis on a separate threat model. Conflating orbital mechanics with cybersecurity economics produced the 2-year apocalypse headline, which neither paper supports. The actual findings are more compelling without the hype.

The authors note that even a single breakup producing 10 critical debris fragments could drop the CRASH Clock below one day. At that point, coordinating responses across dozens of operators and national agencies in real time becomes impossible. A large collision at the wrong altitude could trigger a Kessler cascade that renders entire orbital shells unusable for decades.

The policy implications are significant. Current norms around space traffic management were designed when 164 days was the buffer. At 5.5 days, the burden on operators to maintain situational awareness, execute maneuvers rapidly, and coordinate across borders is qualitatively different. The authors argue this requires new international frameworks for rapid response, not the slow bilateral diplomatic channels currently in place.

For the commercial space industry, the number is a direct operating risk. Starlink, Amazon Kuiper, OneWeb, and other constellation operators all depend on the same orbital volume and tracking infrastructure. A cascade that rendered LEO unusable would destroy the business cases of every operator with hardware in the affected altitude bands. The CRASH Clock belongs in the same conversation as collision liability, insurance pricing, and regulatory timelines for debris remediation.