Frontier AI startup Twin Prime has raised $10 million in pre-seed funding as it looks to build next-generation AI models designed specifically for defence and national security applications.
The round was led by Expeditions, with additional backing from American and European venture capital firms, family offices, and angel investors connected to companies including Palantir, Anduril Industries, and Quorum.
The funding comes amid growing global interest in AI systems capable of operating in high-stakes defence environments where real-time decision-making, sensor integration, and edge deployment are becoming increasingly important.
Twin Prime Is Building AI for Real-Time Threat Detection
Founded in 2025, Twin Prime is developing AI models designed to reason across large numbers of sensor modalities in the physical world, particularly within defence and national security operations.
The company says its models are focused on compressing the “perception-to-decision” layer, enabling faster and smarter responses to threats in real time.
Unlike traditional AI systems that primarily process text or isolated datasets, Twin Prime’s technology is being designed to combine and interpret multiple streams of sensor data simultaneously.
This approach could improve situational awareness and operational decision-making across military and security environments.
Founding Team Brings Experience From AI, Finance, and Defence
Twin Prime was founded by George Lentzas, Stephane Sezer, Drew Calcagno, and Michael Leite-Garcia.
The multinational founding team combines expertise across frontier AI research, quantitative finance, and national security.
Collectively, the founders have previously worked at institutions including:
- Hudson River Trading
- Google Research
- Lawrence Livermore National Laboratory
- Columbia University
- The White House
- United States Department of Defense
The founders also bring experience from various branches of both the US and European armed forces.
Investors See Growing Demand for AI-Native Defence Systems
Mikolaj Firlej said current AI systems are often poorly suited for modern warfare and defence operations.
According to Firlej:
“Current AI models are not specialised enough to capture the complexity of modern war and wider security challenges.”
He added that many existing systems struggle with edge deployment, sensor integration, and real-time data processing.
“Twin Prime is developing models purpose-built for high-stakes environments where size, fusion and speed are critical.”
Firlej also argued that defence contractors have faced difficulties integrating advanced AI into legacy systems, creating opportunities for startups building AI-native infrastructure for defence applications.
Twin Prime Plans Joint Venture With Theon
Twin Prime also announced plans to form a joint venture with Theon, a major European defence prime contractor.
The partnership aims to develop, commercialise, and deploy AI solutions built on Twin Prime’s proprietary models.
The joint venture is expected to expand Theon’s deployment of defence sensors and support its broader Theon Next initiative focused on next-generation defence technologies.
The collaboration highlights increasing interest from established defence firms seeking to integrate advanced AI capabilities into military systems and operations.
AI Defence Startups Continue to Attract Investor Attention
Twin Prime’s funding reflects growing investor appetite for AI companies operating at the intersection of national security, defence technology, and frontier artificial intelligence.
As governments and military organizations increasingly prioritize autonomous systems, real-time intelligence, and sensor fusion, startups building AI-native defence infrastructure are attracting significant attention from both venture capital firms and strategic industry players.
The company’s focus on multi-modal AI reasoning and real-time operational intelligence positions it within one of the fastest-growing sectors of the AI industry.
