Oak Raises $60M to Build Identity Operating System for AI Agents and Human Users

Oak raises $60M for an AI-native identity platform

Oak, an Israeli startup, has raised $60M for an AI-native identity platform that governs every user, machine, and AI agent from one place. Most companies still cannot say who, or what, has access to their systems at any given moment. Oak, an Israeli startup, has raised $60M to fix that, and it is betting AI agents make the problem urgent. The company came out of stealth with the seed round, it said. Its aim is an "identity operating system": a single control plane that governs every identity in a firm, whether a human, a machine, or an AI agent. Why identity is the target Identity is the front door to any company, which makes it the top target for attackers. Yet most firms run a patchwork of old tools built for human staff and slow-moving systems. The rush of machine and agent accounts has left them behind. Oak wants to replace that patchwork. Its software connects to any system, builds a live map of every identity from how it actually behaves, and strips out access that is no longer used. It does this in real time, not in a once-a-year review. A serial founder's next act The bet is really a bet on the team. Chief executive Shai Morag has built and sold three security companies, including Ermetic, bought by Tenable for $265m in 2023, TechCrunch reported. His exits total around $500m. "Our vision is to be born as a giant," Morag said. He told his wife Oak would be his last company. "I will go big or go home." The round was co-led by Accel, Greylock, and CRV, and first revealed by Calcalist. It is one of the largest seed rounds ever for an Israeli cyber firm. Oak already has about 50 staff and paying enterprise customers. An identity land grab Oak is not alone. Identity has become one of the hottest corners of security, and the giants are circling. Palo Alto Networks recently agreed to buy CyberArk, a sign of how much the category is worth. The trigger is AI. As agents start acting on their own, the web is scrambling to work out how to give them an identity. Researchers have already tricked agents into leaking private code and even running a ransomware attack. Firms are racing to control the bots before that spreads. Oak's wager is that all of it, from staff logins to Alexa-style assistants, will finally sit under one roof. Morag thinks the winners here will be worth "tens and even hundreds of billions."

Identity management startup Oak launches with $60M in funding

Oak Inc., a startup that helps companies prevent unauthorized access to their applications, launched today with $60 million in seed funding. Accel, Greylock Partners and CRV jointly led the round. They were joined by Hetz Ventures, AlphaDrive Ventures and several unnamed angel investors. Tel Aviv-based Oak is led by Chief Executive Officer Shai Morag, who earlier founded three other cybersecurity startups. The first was acquired by Mellanox, a network equipment supplier that was itself sold to Nvidia Corp. in 2020. The other two companies were bought by Palo Alto Networks Inc. and venture-backed cybersecurity provider Tenable Inc. Each employee at a large enterprise has access to multiple business applications. Furthermore, many workers use artificial intelligence agents that themselves use several programs. That makes it difficult to track who accesses what and when, which can lead to cybersecurity issues. Oak provides a platform that automatically maps out the access permissions in a corporate network. According to the company, its software identifies every employee's application accounts and AI agents. It then uses that information to find potential risks. Many enterprises implement a cybersecurity best practice known as separation of duties, or SoD. It specifies that a single employee shouldn't have access to multiple sensitive components of the same system. For example, a company may wish to ensure that workers who can download files from a database don't have permission to modify the database's user access logs. Oak can identify accounts that don't implement SoD correctly. It also spots other issues, such as AI agents that have access to services they don't strictly require to perform their work. Unnecessary access permissions can become a source of risk if hackers compromise an AI agent. Oak displays the issues that it finds in a ChatGPT-like interface. Administrators can use natural language prompts to collect more information about an incident and request remediation suggestions. Furthermore, Oak can fix some issues automatically. Before the platform implements a fix, it generates a step-by-step overview of its remediation plan and requests approval. Oak collects the telemetry that it uses to spot cybersecurity gaps via prepackaged integrations with popular software products. In some cases, setting up such connectors can take months of work. Oak says that its platform enables customers to complete the task in a few days. Furthermore, it can collect telemetry from systems that don't provide an application programming interface for connector developers. "I've built several companies in this space, so I understand why identity has stayed broken for so long," Morag said. "The tools were never built to work as one, and adding more of them was never going to fix it." Oak will use the proceeds from its funding round to hire more engineers.