Argo: IA and mathematics to test satellite network security
Argo combines AI with formal methods to verify the resilience of satellite networks. Here is its functioning, its potential and its limitations.
A cyber attack can interrupt communications far beyond its original target. In February 2022, the attack on Viasat's KA-SAT satellite network had deactivated thousands of modems in Ukraine and elsewhere in Europe.
Atalanta wants to help critical infrastructure operators better measure this type of risk with Argo, a commercial platform unveiled on August 18, 2026. It combines artificial intelligence, mathematical systems modelling and methods to check whether certain safety or resilience properties are actually respected.
The technology is promising, but evidence of a specific model does not guarantee that a complete system will withstand all attacks.
Check behavior rather than just looking for faults
Traditional cybersecurity tools can scan the code, monitor the network or simulate attacks. They find known weaknesses, but do not necessarily demonstrate how a system as a whole will react when several components fail or a link is blurred.
Argo takes an approach inspired by formal methods. These mathematically describe the expected behavior of a software or system, and then check whether specific properties remain true in the analysed scenarios.
The National Institute of Standards and Technology defines these methods as mathematical techniques for specifying and verifying properties. For example, they can confirm that an essential link is still available despite certain disturbances or that data does not cross a security border.
Atalanta calls its approach "software understanding". The company claims that Argo is the first commercial platform in this category. This claim was made by the manufacturer and no full independent market comparison was published.
A first concrete case with Viasat
Atalanta has been working with Viasat for about a year, according to information provided to the Associated Press.
In their joint announcement, companies explained that they mathematically modelled a satellite communications system. Argo would then have checked whether the improvements to the links met operational thresholds when they were exposed to interference and hostile interference.
This demonstration must be distinguished from the attack of 2022. Argo did not repair the KA-SAT network after the incident and Atalanta did not claim to have reproduced exactly the same intrusion.
Instead, the partnership seeks to provide evidence of the resistance of some current systems to emerging disturbances and attacks. Viasat states that the platform helps measure improvements that were previously difficult to demonstrate.
(https://apnews.com/article/cybersecurity-ai-new-technology-hacking-742e4946bc20e421fd0d19f0c5a36e0d) confirms the marketing of Argo and its use in Viasat's resilience work. However, there is no detailed technical report, costed result or independent audit accompanying the announcement.
Why the attack of 2022 remains important
On 24 February 2022, about an hour before the Russian invasion of Ukraine began, a cyberattack disrupted the KA-SAT network. It has affected authorities, businesses and users in several countries.
This incident showed that an attack on a communications infrastructure could produce cross-border civilian consequences. It also explains why operators now want to assess the resilience of a network before it is deployed, rather than after an interruption.
Beyond satellites
Atalanta also presents argo versions for energy and defence systems.
The platform could be used to verify that an industrial system respects its security constraints, that sensitive data remains isolated or that mandatory human interventions are not circumvented.
According to the Associated Press, Atalanta technology is also involved in work related to the Genesis mission of the US Department of Energy, which is studying more autonomous nuclear reactors.
This participation does not constitute a general approval of Argo. Rather, it shows that the approach is evaluated in areas where software error can have significant consequences.
Prices and availability in Canada
Atalanta does not publish any prices or ranges for Argo. The platform is available on request to organizations operating critical infrastructure.
No specific Canadian availability, facility or trade agreement has been announced. Costs may vary depending on the system being analyzed, its complexity, integration services, security requirements, the support contract and the region.
Limits or Considerations
A mathematical proof always depends on the model, the properties chosen and the assumptions of the beginning. If a component, configuration or threat is not properly represented, the evidence may not cover the actual behaviour.
NIST also recalls that traditional testing is still necessary: certain assumptions used in evidence may no longer be valid when the specification is converted into code and linked to external libraries.
Argo therefore does not replace operational monitoring, updates, access controls, intrusion testing or staff training.
Atalanta also does not detail the exact role of the AI, the duration of the analyses, the resources required, the rate of false alerts or the proportion of a system that the platform can automatically examine.
Conclusion
Argo tries to pass cybersecurity from a logic where one looks for known faults to a logic where one mathematically checks certain essential properties.
The work done with Viasat is a concrete example, but still little documented publicly. It does not prove the invulnerability of a satellite network or the elimination of all cyberattacks.
If the results can be replicated and audited, this combination of AI and formal methods could nevertheless become an important tool for satellites, energy networks and other infrastructures whose operation cannot depend solely on conventional testing.
