Alice, an AI trust, safety and security business that shields billions of internet users from online manipulation, has closed a $140m funding round as enterprises race to contain the risks of giving AI systems real autonomy.
The round was led by the Apax Digital Funds, with fresh backing from Samsung, SentinelOne, Maj Invest, MoreTech and Phoenix Insurance. Existing backers Resolute Ventures, Grove Ventures, CRV, Highland Europe, Vintage Investments, Norwest Venture Partners, NFX and Claltech also took part, lifting Alice’s total funding to $280m.
The raise lands as pressure builds across the industry to keep pace with adversaries who are exploiting AI at speed. Alice built its foundations by tracking how bad actors abuse the internet’s biggest platforms, amassing a proprietary record of real-world attacks that it now applies to forecasting and blocking threats against AI systems specifically.
That focus has translated into rapid commercial traction: the company is nearing $100m in annual recurring revenue, with its AI-focused business expanding more than fivefold over the past two years.
Alice’s research arm, one of the largest dedicated to AI security anywhere, employs upwards of 150 researchers probing how models can be tricked or manipulated. The firm now safeguards more than three billion people online and counts eight of the world’s ten leading AI model labs among its clients.
The urgency behind the raise is underscored by independent findings. The International AI Safety Report 2026, compiled by over 100 experts, concluded that even heavily protected models remain vulnerable to a high rate of successful attacks, with new exploitation methods surfacing faster than defences can be updated. Separately, research body METR has documented numerous cases of AI agents operating beyond their intended boundaries, including instances where they attempted to hide this from human supervisors.
Alice positions itself across the entire AI lifecycle. Ahead of a model’s release, its researchers collaborate with frontier labs such as Anthropic, Google and Cohere to stress-test systems against adversarial prompts and agentic misuse, hardening them against jailbreaks and prompt injection.
After deployment, the company helps enterprises translate a model’s built-in protections into policies tailored to their own risk requirements, running simulated attacks to expose compliance gaps and data leakage before monitoring live inputs and outputs on an ongoing basis.
This capability draws on nearly ten years spent mapping fraud, extremism and coordinated abuse across the open web, work that produced Rabbit Hole, described as the largest existing dataset of real-world adversarial and harmful content. That same intelligence already underpins protections for users of Google, Meta, TikTok and Amazon.
Alice CEO and co-founder Noam Schwartz said, “There are infinite ways to break an AI, and you can’t defend against something you’ve never seen. We spent nearly a decade learning exactly how people abuse technology at the scale of billions, first on the world’s largest platforms and now on the models those same people are probing. We are very quickly democratizing capabilities before we’ve democratized the defenses. This round is about closing that gap.”
The new capital will go towards deepening Alice’s platform across testing, defence and monitoring; growing the team behind Rabbit Hole so its dataset keeps up with fast-evolving attack methods; and scaling the go-to-market function serving both foundation model labs and enterprise customers.
For those tracking where capital and strategy are converging next, stay ahead of the moves shaping the industry’s future. Subscribe to the FinTech Global newsletter today.
Copyright © 2026 FinTech Global
