Novee: Reclaiming the Offensive Advantage with the Industry’s First Proprietary AI Hacker

I am thrilled to announce the official launch of Novee today with $51.5M in total funding. Novee has built the industry’s first proprietary AI Hacker: a purpose-built LLM trained specifically on exploitation datasets that is already being deployed by global enterprises to transform their offensive security posture. In a market where attackers now operate at machine speed using autonomous agents, Novee provides the only platform capable of matching that velocity with continuous, autonomous penetration testing that thinks and adapts like an elite security researcher. Founded by Ido Geffen, Gon Chalamish, and Omer Ninburg, veteran offensive security leaders from the Israel Defense Force’s most elite cybersecurity programs, the company has achieved one of the fastest funding trajectories in the offensive security sector, raising two rounds in under four months. We are privileged to support these founders alongside our VC partners, Canaan and Zeev Ventures.

The underlying dynamics between attack execution and defensive response have fundamentally changed. The recent Anthropic AI hack, known as the GTG-1002 campaign, revealed a Chinese state-sponsored group utilizing autonomous agents to execute 90% of an attack chain at machine speed with several requests per second. When the offense becomes agentic, traditional detection and response playbooks are simply not fast enough. While attackers use AI as an operator to chain exploits in seconds, defenders remain trapped in manual, periodic validation cycles. To survive this shift, offensive security must move beyond a scheduled exercise and become a continuous, autonomous capability that can anticipate an attacker’s next move.

When Ofer and I met Ido, Gon, and Omer at the Norman Bar in Tel Aviv, we immediately recognized a rare combination: elite operational experience paired with academy-level AI research and sophisticated data science engineering abilities. It is rare to find a founding team that has spent years operating at scale in critical cyber environments and possesses the technical rigor to build proprietary AI models from the ground up. The conviction was immediate, and we committed to partnering with them on the spot.

What sets Novee apart is technical substance. While much of the market relies on thin layers built atop general-purpose models, Novee developed a proprietary LLM trained specifically on exploitation datasets. In recent benchmarks, this model outperformed frontier LLMs by over 55%, achieving 90% accuracy in complex exploitation scenarios where general models typically plateau at 65%. Novee’s platform reasons through attack paths, uncovers novel vulnerabilities including complex business logic flaws that previously only the most advanced manual pentesters could find, and validates exploits with the logic of a veteran researcher.

The market’s response has been decisive. Novee’s platform replaces periodic manual penetration testing with autonomous, continuous validation, ensuring security posture becomes a living defense rather than a point-in-time snapshot. This represents a fundamental rebalancing of the modern threat landscape, and we believe it will disrupt legacy security categories entirely. We are proud to back Ido, Gon, Omer, and the entire Novee team as they lead this transformation in how enterprises defend against the next generation of threats.