An AI-powered cyberattack could trigger a “satellite apocalypse” within two years: Are we ready?

March 23, 2026  19:30

Cybersecurity experts are raising alarms: within the next one to two years, artificial intelligence could give hackers the ability to take control of satellites in orbit and deliberately crash them into one another. According to Space.com, this could trigger the Kessler syndrome—a cascading chain reaction of collisions that would render low Earth orbit unusable for satellites for decades. This warning comes from specialists at the CR14 Cybersecurity Center in Estonia and ETH Zürich in Switzerland.

How AI is changing the rules of cyberwarfare in spaceChristian Kesküla, head of the Space Cyber Range at CR14, explained that AI is evolving extremely quickly and that the biggest issue is its ability to act autonomously—making decisions, analyzing situations, and discovering new exploits.

The discussion centers on so-called agentic AI—autonomous systems built on large language models, similar to ChatGPT or Gemini, which can plan and execute complex tasks without constant human supervision. These systems are already used in defense to identify zero-day vulnerabilities (previously unknown flaws in code) so they can be fixed before attackers exploit them. However, the same technologies are also accessible to adversaries.

Clémence Poirier from ETH Zürich noted that in 2024, OpenAI and Microsoft revealed that the Russian group Fancy Bear had used large language models to conduct reconnaissance on satellite communications, radar systems, and other space technologies. She explained that AI significantly accelerates the reconnaissance and information-gathering phase before an attack and that the time required to exploit known vulnerabilities has been reduced dramatically.

Andrzej Olchawa from VisionSpace added that large language models have drastically lowered the barrier to understanding how spacecraft and communication protocols work. He explained that what previously required years of studying thousands of pages of documentation can now be achieved by instructing an LLM to generate a parser with minimal mission context—even by someone without prior experience in the space industry.

Why older satellites are easy targetsThe space industry largely ignored cybersecurity threats for decades. Many satellites still operating in orbit have little to no protection against hacking. According to Kesküla, they represent “low-hanging fruit” for attackers.

Possible attack scenarios include:

jamming or spoofing communication signals from Earth or space; complete takeover of a satellite’s control systems; turning a satellite into an “orbital anti-satellite weapon” by deliberately crashing it into others.

Kesküla pointed out that gaining control of even a single satellite could create serious problems. Over the past three years, around 8,000 satellites have been launched into orbit, creating high density. One destroyed satellite could generate thousands of debris fragments in low Earth orbit.

Simulations and training are already underwayCR14 is one of the world’s leading cybersecurity centers. Due to Estonia’s proximity to Russia, its specialists are on the front lines of cyber defense. They regularly conduct exercises in virtual environments using digital twins of satellites, where one team simulates attacks while another defends.

Kesküla noted that these exact scenarios are being modeled in their exercises.

Martin Hanson, head of communications at CR14, added that both the number and complexity of cyberattacks are growing exponentially. In Ukraine, thousands of attacks on critical infrastructure—including energy systems and satellite communications—are recorded daily. In Europe, phishing attacks have increased by 500% in recent years, with AI making them far more targeted and convincing.

Hanson explained that AI enables attackers to gather more information about victims and convincingly impersonate friends or colleagues, making attacks increasingly sophisticated.

In briefExperts from CR14 (Estonia) and ETH Zürich warn that within one to two years, agentic AI—autonomous systems based on large language models—could enable hackers to seize control of satellites and orchestrate deliberate collisions. This could trigger the Kessler syndrome, making low Earth orbit unusable for decades. Many older satellites lack cybersecurity protections. While training exercises are already underway, the industry is only beginning to grasp the scale of the threat.

Follow NEWS.am Tech on Facebook and Twitter