Exein, the Physical AI cybersecurity company, announced $270 million in new funding at a $1.7 billion valuation, making it the most valuable cybersecurity startup in Europe. The funding will accelerate Exein’s US and APAC expansion and drive further impacts in Physical AI security: protecting intelligent machines that use AI to perceive, make decisions and act in the physical world, from robots and drones to autonomous vehicles. Exein is also developing a proprietary foundation model specifically to secure these systems.
The round is led by Headline, with participation from Sofina, Goldman Sachs, European Investment Bank Group ETCI, KfW Capital and T.Capital, with involvement from previous investors Balderton, HV, Intrepid Growth Partners, 33N Ventures, Lakestar, Supernova Invest, Blue Cloud Ventures and Geodesic Capital. The round was significantly oversubscribed.
This is alongside an upsizing of Exein’s existing revolving credit facility led by J.P. Morgan, while KfW joins as an additional lender.
Exein’s valuation has increased thirty-fold in two years, reflecting a period of rapid commercial and international growth. The company's H1 2026 ARR was up 4x year-on-year.
As AI moves into machines that perceive, make decisions and act in the physical world, a new security challenge is emerging. Exein has been building for this transition for years, combining deep expertise in embedded and runtime security with technology that operates where code actually runs. Earlier this year, the company launched Photon, its preemptive runtime security architecture, which operates at the kernel level to block malicious execution before an attack can run, an approach particularly suited to autonomous machines where security must respond at machine speed.
Exein is now combining this runtime expertise with its machine data advantage to develop a proprietary foundation model designed specifically for Physical AI security. In development for two years, the model is being trained on machine telemetry generated through Exein technology distributed across more than two billion devices. Unlike general-purpose AI models trained largely on human-generated data, Exein’s model learns from data showing how machines actually operate, a dataset built through years of real-world machine activity that would be difficult for competitors to replicate.
