In a significant move that underscores the growing urgency of AI security, Alice—formerly known as ActiveFence—announced on Tuesday that it has successfully closed a $140 million funding round. This capital injection brings the firm’s total venture backing to an impressive $280 million, positioning it as a cornerstone player in the race to protect foundational machine learning models from adversarial attacks and systemic vulnerabilities.
The round was led by Apax Digital Funds, with a robust consortium of participants including MoreTech, Phoenix Financial, Resolute Ventures, Grove Ventures, CRV, Highland Europe, Vintage Investments, Norwest, NFX, and Claltech. This financial influx is earmarked for three primary objectives: accelerating the technical development of Alice’s AI defense platform, expanding its elite team of data scientists and security researchers, and scaling global go-to-market operations.
The Chronology of Evolution: From ActiveFence to Alice
The narrative of Alice is one of deep-rooted expertise in digital safety. Operating for nearly a decade under the name ActiveFence, the company built its reputation by tracking and neutralizing digital fraud, ideological extremism, and large-scale manipulation campaigns. This foundational work allowed the firm to curate a proprietary intelligence database known as "Rabbit Hole."
As the landscape of technology shifted toward generative AI, the leadership team at Alice recognized a critical parity gap. While the tools to create powerful AI were becoming democratized at an exponential rate, the tools to defend those systems remained fragmented and reactive. The transition from ActiveFence to Alice represents more than a rebranding; it signifies a strategic pivot toward applying years of experience in content moderation and threat intelligence specifically to the architecture of large language models (LLMs) and agentic AI systems.
Decoding the Technical Defense: How Alice Protects AI
Alice’s approach to security is bifurcated, focusing on the pre-release "stress-testing" phase and the post-deployment "runtime" protection phase.
Pre-Deployment: Red-Teaming and Stress-Testing
Before a foundational model is released to the public, Alice partners with developers to perform rigorous stress tests. This involves a research division that subjects models to mock malicious inputs and complex, agentic tasks designed to trigger erratic behavior. By simulating prompt injections, jailbreak attempts, and logic-bomb scenarios, Alice helps developers harden their models, ensuring that the software behaves according to its intended parameters before it encounters real-world bad actors.
Post-Deployment: Dynamic Runtime Guardrails
When models enter production, the challenge changes. Alice provides continuous, real-time monitoring through dynamic runtime guardrails. This allows organizations to:
- Configure custom internal guidelines: Tailoring AI responses to specific organizational policies.
- Simulate breaches: Constantly testing the system against new attack vectors as they emerge.
- Real-time traffic monitoring: Intercepting hostile requests before they can influence the model’s outputs.
The backbone of this capability is the "Rabbit Hole" dataset. By leveraging a decade’s worth of data on how bad actors manipulate digital platforms, Alice’s AI can predict and intercept novel attack patterns that other security systems might overlook.
Official Perspectives: Closing the Defense Gap
Noam Schwartz, CEO and co-founder of Alice, views the current state of AI security as an urgent, asymmetric challenge. "There are infinite ways to break an AI, and you can’t defend against something you’ve never seen," Schwartz noted in a statement following the funding announcement.
He emphasized that the current speed of AI development has outpaced the development of security protocols. "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."
By bridging the divide between offensive research and defensive implementation, Alice aims to provide a safety net that is as dynamic and adaptive as the models it protects.
Supporting Data and Industry Context
The rise of Alice comes at a time when the broader cybersecurity sector is increasingly focused on the unique risks posed by generative AI. Unlike traditional software, AI systems can be "tricked" into bypassing safety filters, leaking training data, or performing malicious actions through sophisticated prompt engineering.
The funding environment for AI security remains highly active. Recent industry activity includes:
- Xpander: Recently raised $7.5 million, highlighting the growing demand for AI management and governance platforms.
- Mindgard: Secured $30 million to focus specifically on the protection of AI systems against adversarial machine learning.
- Team8: A venture firm that recently secured an additional $365 million, signaling significant institutional belief in the long-term viability of the AI security and infrastructure sector.
Alice’s 150-strong team of specialists—housed in research facilities in New York and Tel Aviv—acts as a specialized force, collaborating with major model developers to anticipate failures. This human-in-the-loop approach is vital, as the "infinite ways to break an AI" require constant human insight to interpret and counter.
Broader Implications for the AI Ecosystem
The infusion of $140 million into Alice has several profound implications for the future of the artificial intelligence industry.
1. The Rise of "Security-by-Design"
Alice’s business model reinforces the industry shift toward "security-by-design." By embedding testing into the foundational model development lifecycle, companies are moving away from treating security as a post-release patch. This trend will likely become the gold standard for any enterprise-grade AI deployment.
2. The Commercialization of Threat Intelligence
The success of the "Rabbit Hole" dataset proves that historical data regarding human behavior—specifically regarding fraud and extremism—is one of the most valuable assets in the AI era. As AI becomes more integrated into the global economy, the value of companies that possess "pre-AI" threat intelligence will continue to rise.
3. Regulatory Preparedness
As governments worldwide begin to implement frameworks for AI regulation (such as the EU AI Act), companies like Alice provide the technical infrastructure that allows organizations to comply with these rules. By offering dynamic guardrails, Alice enables corporations to demonstrate that they are taking proactive steps to mitigate risks, a requirement that will likely become mandatory in the coming years.
4. The Tel Aviv-New York Corridor
The success of Alice further cements the strategic importance of the Tel Aviv-New York technology corridor. By leveraging the deep cybersecurity and research expertise often found in Israeli technical circles, combined with the market access and capital depth of New York, Alice is well-positioned to serve a global client base.
Conclusion: The Path Forward
The path to secure, reliable, and trustworthy AI is long and complex. As models grow in size and capability, the "surface area" for potential attacks grows in tandem. Alice’s $140 million round is a clear signal that the investment community recognizes that the utility of AI is directly tied to its security. Without robust defense mechanisms, the widespread adoption of AI in critical infrastructure, healthcare, and finance remains a high-risk endeavor.
By shifting the focus from reaction to anticipation, Alice is setting a new precedent for how technology companies approach the "infinite" threats posed by adversarial actors. As the firm scales its operations and expands its research footprint, it stands as a sentinel at the gates of the AI revolution, tasked with ensuring that the progress of machine learning does not come at the cost of digital safety.
In the coming months, industry observers will be watching to see how Alice integrates its new funding into its research lab and whether its "Rabbit Hole" dataset can continue to stay ahead of the curve in an environment where the methods of abuse are as fast-evolving as the AI models themselves. For now, the company has secured the resources necessary to attempt the monumental task of hardening the world’s most powerful software against the most persistent of human threats.
