Cybersecurity startup Pallma AI lands $1.6 million to protect enterprise AI agents

Although the adoption of generative AI has accelerated rapidly, the security infrastructure for AI agents has struggled to keep pace with the unique risks that these workflows introduce.

Founded in 2025 by Pavlos Mitsoulis and Dionysis Varelas, Pallma AI provides a specialised security and control platform tailored to protect AI agents. The Pallma platform provides real-time oversight of agents' internal decision trees, intent and multi-step plans. This approach enables organisations to enforce dynamic runtime policies and establish clear boundaries for agents operating within business-critical environments. The platform helps to prevent issues such as prompt injection and unauthorised tool misuse before they can affect production systems.

The London-based start-up recently closed a $1.6 million pre-seed funding round, led by Marathon Venture Capital and with participation from a range of strategic investors. The company plans to use this initial capital to further develop its AI-native protection platform and expand its technical team.

Sources: Pallma | Cruncbase

Founders: Pavlos Mitsoulis, Dionysis Varelas