Above Security is an Israeli cybersecurity startup founded in mid-2025 by Aviv Nahum and Amir Boldo, both alumni of elite branches of Israel's Unit 8200 signals-intelligence corps. The company is building an AI-native managed insider-risk platform designed to surface insider threats — compromised employees, malicious actors, and increasingly the misbehavior of internal AI agents — without rules, policies, or manual tuning.
The platform deploys a fleet of specialized AI investigators that continuously model both human and machine behavior across an organization's systems. Rather than firing on signature matches, the agents reason about intent: who has anomalous access patterns, which AI agents are exfiltrating data at machine speed, and where insider blast radius is concentrated. This positioning treats AI agents themselves as a new class of insider — one that operates 24/7, has broad credentials, and is largely invisible to existing insider-risk programs designed for human users.
Above Security closed a $7M seed round led by Merlin Ventures and Norwest at founding, then added a $43M Series A within six months, bringing total funding to approximately $50M in just eight months — with a headcount of around 10 people at the time of the Series A announcement. Ballistic Ventures also participated. The capital-efficient profile reflects investor enthusiasm for AI-agent governance as an emerging security category and the founders' track record: Aviv Nahum previously built Ctrl, an enterprise data-management startup acquired by Israeli AI company Sana.
The company was selected for the 2026 Cybersecurity Startup Accelerator and has not yet disclosed pricing publicly, instead working directly with design-partner enterprises on managed-service engagements. Early customers span financial services and AI-forward enterprises deploying autonomous agents in production.
Above Security differentiates from incumbents like Proofpoint Insider Threat Management, DTEX, and Code42 by being AI-native from day one and explicitly modeling AI agents as insiders — a category most legacy DLP and UEBA platforms have not yet adapted to.