Alice (formerly ActiveFence)

From ActiveFence to Alice: the January 2026 rebrand

Alice is an Israeli-American security company founded in 2018 under the name ActiveFence and rebranded as Alice on 14 January 2026. According to the company's announcement, the old name no longer described the business: after eight years focused on protecting content platforms from harmful content, fraud and coordinated manipulation, its AI model security business had grown faster than the original business and become the core of its activity. According to the company, at the time of the rebrand it was already working with 7 of the world's 10 largest AI labs and protecting more than 3 billion people.

What Alice actually does

Alice offers three layers of protection for AI models and AI agents: pre-launch testing that finds security gaps in a model before it goes into use; runtime protection that intercepts input or output with malicious content, such as a prompt injection or jailbreak attempt; and continuous monitoring after launch. The technology relies on a proprietary database called Rabbit Hole, which, according to the company, collects and analyzes billions of real attack examples gathered from platforms around the world, and on an additional open-source tool called Caterpillar.

Founders, early funding and two acquisitions in 2023

The company was founded by Noam Schwartz (CEO and co-founder), Iftach Orr (CTO), Alon Porat (customer lead) and Eyal Dykan. In July 2021 it revealed $100 million raised in two rounds: a previously unannounced Series A, and a Series B led by CRV and Highland Europe, at a valuation then estimated at about $500 million. In 2023 it made two acquisitions that expanded its AI capabilities: in March it bought Rewire, a London company founded in 2021 that detects harmful content in text; in September it bought Miami-based Spectrum Labs, which had developed context-based detection of harmful content and was then valued at about $137 million.

A $140 million round, August 2026

On 25 August 2026 the company announced a $140 million funding round led by Apax Digital, with participation from Samsung, SentinelOne and earlier investors such as CRV, Grove Ventures and Norwest Venture Partners. The round brings total funding since founding to $280 million. According to the company, its annual recurring revenue (ARR) is approaching $100 million, and its AI security business grew by more than 500 percent in the two years before the round. In the same announcement Alice reported that it now works with 8 of the world's 10 leading AI labs, up from the 7 of 10 it reported in January. Globes and TheMarker reported that the company's valuation after the round is estimated at about $800 million, up from about $500 million in the previous round in 2021.

The stated approach: you can't defend against what you haven't seen

"There are infinite ways to break an AI, and you can't defend against something you've never seen," Schwartz said in the August 2026 funding announcement. The quote explains the company's main investment in the Rabbit Hole database: according to Alice, the volume of real attack examples it collects from the world, and not just a list of rules written in advance, is the basis on which it tests and protects new models.

Customers, employees, and what has not been externally verified

Its reported customers include AI labs such as Anthropic, Google and Nvidia, alongside large platforms such as Amazon. As of the August 2026 round, the company employs about 400 people, most of them in Israel and New York. As with most company announcements about customer numbers and scale of protection, these are figures Alice itself publishes without independent external review. No third party has confirmed the number of users protected or the list of AI labs that work with it.