Cybersecurity startup NewCore exited stealth mode this week with $66 million in seed funding, betting that enterprises will soon need to manage AI agents as first-class digital employees with their own identity lifecycle, permissions, and revocation controls. The round, led by Cyberstarts with Index Ventures and Evolution Equity Partners, values the company at $300 million post-money.

The core thesis: existing identity platforms—built for human users and static service accounts—won’t scale when thousands of autonomous agents start executing code, querying databases, and triggering workflows across an organization. Goldman Sachs tested this last year with Devin, an AI coding agent treated as a new hire. McKinsey now runs 25,000 AI agents alongside its human workforce. NewCore’s founders argue that traditional IAM systems treat machine identities as second-class citizens, relying on shared secrets and long-lived tokens that become audit nightmares.

NewCore’s platform unifies human and AI-agent identity management in a single control plane. Agents get their own permissions, lifecycle policies, and revocation mechanisms. A key architectural decision is a “split-key” system: critical identity credentials are divided between the customer’s environment and NewCore’s infrastructure, eliminating any single point of compromise. The platform also ships with an “Agentic Skill” integration for coding assistants like Claude Code, Codex, and Cursor, letting them authenticate as managed identities rather than via manually distributed API keys. A mobile app gives human operators a lightweight approval layer for granting or revoking agent access.

CEO Zohar Alon, who previously founded Dome9 (acquired by Check Point), says the idea crystallized in 2023 during a budget review at a company using an incumbent identity provider. The customer was paying millions but wasn’t happy—a symptom of a stagnant market with little competitive pressure. NewCore, co-founded with CTO Amihai Neiderman (ex-Unit 8200, Nym Health) and CRO Erez Yarkoni (ex-CIO at T-Mobile and Telstra), now has fewer than 10 paying customers and over 10 design partners. It expects to start charging this summer.

Alon predicts that within a few years, AI agents could outnumber human employees at many tech companies. The engineering challenge isn’t just building authentication—it’s designing revocation at scale, auditing agent behavior across distributed systems, and ensuring that a compromised agent can’t escalate privileges. “The question is whether we’ll build the guardrails in time,” he said.

Source: TechCrunch