Alice raises $140M Series D at $800M valuation to secure AI models
What's the deal? Israeli AI security company Alice has raised $140 million in a Series D round led by Apax DigitalDealroom has a profile for this one. Try Dealroom →, valuing the company at $800 million. New and existing investors joined, including SamsungDealroom has a profile for this one. Try Dealroom →, SentinelOne, Maj InvestDealroom has a profile for this one. Try Dealroom →, MoreTech, Phoenix InsuranceDealroom has a profile for this one. Try Dealroom →, Norwest, CRV, Vintage, Grove, and Highland Europe.
What's the endgame? Alice, formerly ActiveFenceDealroom has a profile for this one. Try Dealroom →, tests AI models for vulnerabilities and defends them against misuse. It works with labs including Anthropic, GoogleDealroom has a profile for this one. Try Dealroom →, and Nvidia during training to simulate malicious prompts and harden models against jailbreaks and prompt-injection attacks.
Why now? Demand has grown as companies and AI labs face pressure to secure models and agents against a rising threat from bad actors. Founded in 2018 to moderate content on social platforms, Alice began working with AI labs in 2022 and rebranded in January to reflect the shift.
By the numbers: The round brings Alice's total raised to $280 million, and its annual recurring revenue is approaching $100 million. The company runs an AI security research lab with more than 150 researchers and protects eight of the world's 10 leading AI model developers.
The Series D sits in the 92nd percentile for its sector among Israeli rounds of this stage, underlining investor appetite for AI security.
Where it's growing: Alice employs about 400 people, most in Israel, with offices in New York, London, and Hanoi. It will use the funding to advance its platform, expand the team behind its Rabbit Hole dataset of real-world attack patterns, and scale its go-to-market organisation serving foundation model labs and enterprises.
What could go wrong? Alice is one of several companies pitching tools to catch AI misuse, and the field is crowded. Chief executive Noam Schwartz played down fears of runaway models: "I don't think we're going to see models running around and hacking people anytime soon. But we do need to take these things incredibly seriously."
The signal: As high-profile incidents involving autonomous agents alarm regulators, a distinct AI safety layer is forming alongside traditional cybersecurity. "The only way to earn that trust is to stress test models and agents every day and guard them with defenses built from real attacks," Schwartz said.
Read more: calcalistech.com, en.globes.co.il, Bloomberg
Image credit: Alice
