Alice

Alice, rebranded from ActiveFence in January 2026, builds a safety and security layer for AI models and the platforms built on them. The company started in 2018 as a content-moderation and trust-and-safety data provider for social platforms, then pivoted its core technology toward securing AI models directly: red-teaming foundation models, monitoring for jailbreaks and misuse, and feeding real-world attack pattern data (its “Rabbit Hole” dataset) back into model safety training.

The platform is used by AI labs and large consumer platforms to catch abuse patterns before they scale, drawing on years of trust-and-safety data collection across billions of user interactions. Its differentiator is that depth of real-world abuse data, rather than a lab-only red-teaming exercise, which the company argues produces more realistic coverage of how models get misused in production.

Alice raised a $140 million round in August 2026 led by Apax Digital, with participation from SentinelOne and Samsung, valuing the company between $700 million and $800 million and bringing total funding to roughly $280 million. The company says it is approaching $100 million in annual recurring revenue and protects eight of the world’s ten leading AI model-development labs.