NewCore emerged from stealth this week with a $66 million seed round, an unusually large one, at a $300 million post-money valuation. The round was led by Cyberstarts, the cybersecurity-focused firm, with Index Ventures and Evolution Equity Partners joining. The problem the company is selling against is a specific one: AI agents are starting to act as employees, and the identity systems that govern who can touch what inside a company were built for humans.
Chief executive Zohar Alon frames it bluntly, that the scale and complexity agents will add to identity platforms that are fifteen or twenty years old are going to break them. His read carries some weight. Alon previously founded the cloud-security company Dome9, which Check Point acquired, and his co-founders are not lightweights either: Amihai Neiderman led research at Israel's Unit 8200 and founded the healthcare AI company Nym Health, and Erez Yarkoni was chief information officer at T-Mobile USA and at Telstra.
The technical model is the reason this belongs in the security column rather than the funding one. NewCore treats AI agents as first-class identities, each with its own permissions, lifecycle controls, and revocation mechanism, managed in the same system as human staff. It uses a split-key architecture that divides critical credentials between the customer and the platform, so a compromise of either side alone does not hand over the keys. Its Agentic Skill integration lets coding assistants like Claude Code, Codex, and Cursor reach into enterprise systems as managed identities, and a mobile app lets a person grant, review, and revoke an agent's access. That design rhymes with the security lesson the field keeps relearning this week: keep the dangerous capability outside the model's direct reach, and bind what it can do rather than trusting it to police itself.
The bet is on a near-future that is not fully here. NewCore has more than fifty employees across the United States and Israel, but fewer than ten customers, ten-plus design partners, and does not expect to start charging until the summer. What gives the thesis its urgency is the scale already visible elsewhere: Goldman Sachs has tested the coding agent Devin as an employee, and McKinsey says 25,000 AI agents already work alongside its 60,000 people. Alon's prediction, that agents could outnumber humans at technology organizations within a few years, is the kind of claim that sounds aggressive until you do the arithmetic on how fast the agent count is climbing.