On August 25, Alice — formerly ActiveFence — announced a $140 million round led by Apax Digital Funds. Samsung, SentinelOne, Maj Invest, MoreTech, and Phoenix Insurance joined. Existing backers including Resolute Ventures, Grove Ventures, CRV, Highland Europe, Vintage, Norwest, NFX, and Claltech came back. Total capital raised: $280 million. Apax takes a board seat.
What the company did not print is a valuation.
That omission is the tell. Alice’s own announcement says it works with eight of the ten leading AI model labs, is approaching $100 million in annual recurring revenue, and that the AI slice of the business grew more than 500% over two years. Those are strong numbers. They are also the kind of numbers that usually come with a headline price. The price leaked anyway, in three directions. The Next Web, citing Bloomberg’s Marissa Newman, reported that CEO Noam Schwartz put the round “close to $1bn.” Calcalist’s English edition put it at $700–800 million. Globes, via the Jerusalem Post, reported $800 million. The spread is roughly $300 million — more than twice the size of the round.
That is not a rounding error. It is a disagreement about what Alice is.
If you believe Alice is an AI-native security platform selling tests and guardrails into Anthropic, Google, and Cohere, with a path through the enterprise as agents get tools and data, then a near-billion-dollar mark on roughly $100 million of ARR is a growth round, not a fantasy. About 10× ARR is not 2021 SaaS. It is not cheap either. If you believe Alice is a trust-and-safety company that rebranded in January 2026 and is now pointing a content-moderation archive — Rabbit Hole — at models, then $700–800 million is the more honest print. SentinelOne taking a stake is interesting either way: a public cyber vendor buying a look at the same buyers it sells.
Compare the round that did print a price. On August 4, Obsidian Security closed an $85 million Series D at a $1.1 billion valuation, led by Crescent Cove Advisors, with Greylock and Menlo Ventures participating. Obsidian disclosed operating metrics: more than 100 customers spending over $100,000 a year, and more than 14 spending more than $1 million. That is a SaaS security company with a price and a cohort. Alice disclosed labs, ARR “approaching $100 million,” and no price.
The valuation question is not “is AI security real.” Obsidian’s $1.1 billion and Alice’s $140 million in the same August answer that. The question is whether a company that will not put a number on its own round should be marked at 10× ARR because eight frontier labs are customers. Labs are concentrated. Enterprises are the expansion. Alice says it will spend on the dataset, the platform, and go-to-market. That is the right use of capital. It is also an admission that the enterprise motion is not finished.
TAM for “AI trust, safety, and security”: Undisclosed. Alice cites the International AI Safety Report 2026: even well-defended models still break at a high rate, and new attack techniques emerge faster than defenses close them. That is a demand signal, not a market size. Until Alice or a filing publishes the post-money, treat “close to $1 billion” as the founder’s number and “$700–800 million” as the local-press number. The round is real. The $140 million is real. The ARR claim is the company’s. The price is still an argument — and in a tape that just marked Obsidian at $1.1 billion with a customer cohort, the argument is the story.
If you are allocating, do not underwrite the Bloomberg number and the Calcalist number as if they were the same deal. They are not. A $300 million gap on a $140 million round is the market telling you it has not decided whether Alice is a platform or a lab-services business wearing a security logo. Force the company to pick one before you pick a price.