O que aconteceu
Citybiz reports that Alice raised $140 million in a funding round led by Apax Digital Funds, bringing the company’s reported total funding to $280 million. The New York- and Tel Aviv-based company plans to expand its AI security platform, research team and go-to-market operations.
Citybiz reports that Alice, the AI trust, safety and security company formerly known as ActiveFence, raised $140 million in a round led by Apax Digital Funds. MoreTech and Phoenix Financial reportedly participated, alongside existing investors including Resolute Ventures, Grove Ventures, CRV, Highland Europe, Vintage Investments, Norwest, NFX and Claltech. Citybiz says Apax Digital will join Alice’s board and that the financing brings the company’s total funding to $280 million. The supplied source does not independently confirm the transaction, its closing date beyond the article’s publication context, the round’s valuation or any investor commitments beyond Citybiz’s account.
Citybiz reports that Alice plans to use the capital to develop its AI security platform, expand the research team behind its proprietary Rabbit Hole dataset and enlarge its go-to-market organization for foundation-model developers and enterprises deploying AI. According to the outlet, Alice is approaching $100 million in annual recurring revenue and its AI business has grown by more than 500% over the past two years. Those figures are attributed to Citybiz and are not independently verified in the supplied material. The report does not provide audited financial statements, a revenue breakdown, customer-level spending data or a definition of how the reported growth was calculated.
The outlet describes Alice’s platform as covering multiple stages of the AI lifecycle. Before deployment, Citybiz says Alice researchers simulate malicious prompts and agentic tasks to look for jailbreaks, prompt injection and other unpredictable behavior. After deployment, the company reportedly offers tools for organization-specific policies, attack simulation, compliance and data-security checks, and real-time monitoring of model inputs and outputs. Citybiz says Alice operates an AI security research lab with more than 150 researchers, works with eight of the 10 leading AI model labs and protects more than 3 billion people through technology used on major online platforms. The supplied source does not independently confirm those reach, staffing or customer claims.
Leia a fonte primária: citybiz.co ↗
Por que isso importa
The financing reflects growing demand for tools that test and monitor AI systems, particularly systems connected to corporate data, software tools and autonomous workflows. Citybiz says Alice works with major AI labs and enterprises, but those claims are not independently confirmed in the supplied source.
Citybiz frames the investment as a response to a shift in how AI systems are being used. As models move beyond conversational interfaces and gain access to company data, software tools and autonomous capabilities, the possible failure modes become broader than inaccurate answers. The article identifies prompt injection, data leakage, policy violations and systems acting outside their intended parameters as risks. These are practical security concerns because an AI system connected to tools can potentially expose information or take actions across the systems to which it has access. The source, however, does not document a specific incident caused by Alice-protected systems or provide independent testing of its defenses.
The reported business model is significant because it places AI security between model developers and the organizations that deploy models. Citybiz says Alice works with foundation-model developers before release and with enterprises after deployment, where local policies and data protections may differ from the safeguards built into a general-purpose model. That distinction matters in practice: a model that is acceptable for one use may require different controls when connected to regulated data, internal software or autonomous task execution. Citybiz describes Alice’s feedback loop as a process in which attacks observed in production inform new tests, while test results help refine safeguards and monitoring. The report does not establish how often that loop improves outcomes, how false positives are handled or how Alice compares with competing security products.
The funding also signals that investors see AI security as a growing commercial category. Citybiz compares the opportunity with the security market that developed around cloud computing and reports that Apax Digital views autonomous agents as a driver of a widening attack surface. That investment thesis is plausible as a description of market direction, but it is not evidence that Alice’s products prevent attacks at scale. The article cites the International AI Safety Report 2026 and research organization METR’s documentation of agents exceeding their assigned scope, yet it does not reproduce the underlying findings or independently connect them to Alice’s performance. The practical public impact will depend on whether companies adopt effective controls, maintain human oversight and disclose failures rather than treating security tooling as a complete solution.
O que assistir a seguir
Watch how Alice converts the funding into measurable security capabilities, customer deployments and independently verifiable results. Important unknowns include the round’s valuation and terms, the identities and spending of customers, and whether the company’s reported revenue and reach figures can be externally confirmed.
The first question is how Alice uses the new capital in practice. Citybiz says the company will spend on testing, defense, monitoring, the Rabbit Hole dataset and commercial expansion. Useful follow-up reporting would identify specific product releases, deployment environments, customer commitments and independently reproducible evaluations. It would also clarify whether the company’s security tools operate across different model providers and agent frameworks, and whether customers can inspect or audit the decisions made by the monitoring system.
The second question is whether Alice’s reported scale translates into demonstrated protection. Citybiz says the company works with eight of the 10 leading AI model labs, serves platforms including Google, Meta, TikTok and Amazon, and protects more than 3 billion people. The supplied source does not include contracts, customer statements, deployment boundaries, incident data or independent validation of those claims. Future coverage should distinguish between a company’s technology being used somewhere within a large platform and the company being responsible for protecting all users of that platform. It should also examine how Alice measures successful detection, missed attacks, false alarms and the effect of defenses on model usefulness.
The third question concerns the limits of adversarial testing as AI capabilities and attack methods change. Citybiz reports that Alice’s Rabbit Hole dataset is built from real-world harmful and adversarial content and that the company uses it to identify patterns and develop defenses. A dataset based on observed attacks may improve coverage, but the article does not say how representative it is, how quickly it is updated, what data rights apply or whether models can adapt to the tests. The report also does not specify the security controls around Alice’s own research data. The round’s valuation, detailed financial terms, customer concentration, independent performance results and any material security incidents remain unknown.
