Silverfort acquires Fabrix Security, a one-year-old AI access-decision engine

Silverfort has acquired Fabrix Security, an AI-native identity security company founded in 2025 by chief executive Raz Rotenberg, a founding engineer at Run:ai, and chief technology officer Ofir Yakovian, previously at Orca Security and Microsoft Entra. The value was not disclosed; Calcalist put it in the tens of millions. Fabrix uses an identity knowledge graph and AI agents to analyse access activity, organisational context, and intent, then make access decisions for both human and non-human identities. Silverfort pairs that with its Runtime Access Protection layer, which enforces decisions across systems that cannot be modified to speak modern protocols. Most joint capabilities are due in the second half of 2026.

Why it matters

Silverfort's whole business is being an enforcement point where one is otherwise impossible: Active Directory, service accounts, legacy applications, the things that never got an MFA prompt because there was nowhere to put one. What it has lacked is a decision engine sophisticated enough to say yes or no at agent speed rather than by static policy.

That is a coherent buy, and a useful contrast with the year's other deals. Most acquirers bought discovery and posture. Silverfort bought decisioning to attach to enforcement it already owns, which is the harder half.

Judge it on latency and failure mode. An inline enforcement point that consults an AI decision engine has a new dependency in the authentication path, so ask what happens when that engine is slow or unreachable. Our ITDR guide covers the evaluation.

Source: Silverfort