Explore Founder OS: AI tools to help you build faster
Founder OS Founder OS
As enterprises give AI agents access to real systems, data, and the ability to take action, Grove Ventures-backed Alice is raising fresh capital to close the gap between a model’s built-in guardrails and what enterprises actually need.
Widely covered across major media, Alice, a Grove Ventures portfolio company focused on AI trust, safety, and security, has raised $140 million in a funding round led by Apax Digital, with participation from Samsung, SentinelOne, Maj Invest, MoreTech and Phoenix Financial, alongside existing investors including Grove Ventures. The round brings Alice’s total funding to $280 million, with Apax Digital joining the company’s board.
Founded in 2018 as ActiveFence, Alice has spent nearly a decade tracking fraud, abuse, manipulation, cyberattacks and other adversarial behavior across major digital platforms. Today, the company applies that intelligence to AI trust, safety, and security. Grove Ventures invested in Alice early, with Lotan Levkowitz working closely with the team throughout the years.
Alice’s approach is built on years of observing how malicious actors exploit technology at scale.
That work produced Rabbit Hole, Alice’s proprietary dataset of real-world adversarial and harmful content. The company uses patterns identified through this dataset to help predict and prevent attacks on AI systems.
Alice works across the lifecycle of an AI system. Before models are released, its researchers work with frontier AI labs to simulate malicious prompts and agentic tasks, surface unpredictable behavior, and harden models against jailbreaks and prompt injection.
Once models are deployed, Alice works with enterprises to address the gap between a model’s built-in safeguards and an organization’s own requirements. Enterprises can define company-specific policies, simulate attacks to uncover compliance risks and data leaks, and monitor inputs and outputs in real time.
Alice works with 8 of the 10 leading AI model labs, including Anthropic, Google, and Cohere.
The technology developed through Alice’s work tracking real-world adversarial behavior also protects more than 3 billion people across platforms including Google, Meta, TikTok, and Amazon, with coverage spanning 117+ languages.The company has approximately 400 employees, including more than 150 AI security researchers, with teams in Tel Aviv, New York, London and Hanoi.Alice is also approaching $100 million in annual recurring revenue, while its AI business has grown by more than 500% over the past two years.
As enterprises give AI systems access to real data, tools and the ability to take actions, the consequences of systems operating outside their intended scope increase. Alice focuses on identifying how AI systems can be manipulated or fail, and how to stop it. This distinction matters as AI agents move from answering questions to executing tasks with real-world consequences.
The new funding will support Alice’s platform, Rabbit Hole, its research team, and its work with both frontier AI labs and enterprises deploying AI in production. The company describes its mission simply: to make sure AI does exactly what it is supposed to do.
Read more in TheMarker and Globes.
FAQs
A: Alice is an AI trust, safety, and security company that applies nearly a decade of intelligence on real-world adversarial behavior to protect AI systems.
A: Alice raised $140 million in a funding round led by Apax Digital, bringing its total funding to $280 million.
A: The round was led by Apax Digital, with participation from Samsung, SentinelOne, Maj Invest, MoreTech and Phoenix Financial, alongside existing investors including Grove Ventures, Resolute Ventures, CRV, Highland Europe, Vintage Investments, Norwest, NFX and Claltech.
A: Alice works with 8 of the 10 leading AI model labs, including Anthropic, Google, and Cohere. Its technology also protects people across platforms including Google, Meta, TikTok, and Amazon, with coverage spanning more than 117 languages.
A: Rabbit Hole is Alice’s proprietary dataset of real-world adversarial and harmful content, built from years of tracking fraud, extremism, coordinated manipulation and other adversarial behavior across the open web.
