Alice secures $140M to harden frontier AI model security

Alice’s 140 million dollar raise is a marker that AI safety and security is maturing into its own infrastructure layer rather than a niche add on. The company sits in the critical path between frontier labs and real world deployment, stress testing models for jailbreaks, prompt injection, data leaks and unsafe behaviors before and after release. That it is approaching nine figure recurring revenue while working with most of the top labs shows that the demand for adversarial testing and monitoring is not theoretical anymore, it is a line item in big AI budgets.

For the AGI race, this capital does two things. It gives a specialized safety player the resources to build deeper tooling and a larger research team just as labs are rolling out agents with real tools, data and autonomy. It also externalizes part of the safety function into a vendor whose incentives are tied to being the one that finds the failure first. That can improve coverage compared with fragmented in house red teams, but it also creates dependency on a single commercial chokepoint.

Strategically, the round signals that investors believe AI safety can be both a mission and a high growth business. If Alice succeeds, that raises the bar for what “responsible deployment” looks like across the ecosystem, and could become a de facto standard for enterprise buyers pushing labs to prove they have done serious abuse testing.