Check Point Acquires Deepchecks as It Builds Out Agentic Security Platform
Cybersecurity firm Check Point announced it has signed a definitive agreement to acquire the team and intellectual property of Deepchecks, an Israeli AI startup focused on AI testing and reliability.
This will mark the fourth acquisition for Check Point so far this year. The deals show Check Point adding automation and AI capabilities to its security platform, including exposure management, AI governance, tools for managed service providers, and now technology for testing and monitoring AI agents.
Deepchecks gives Check Point a validation layer. The technology is meant to help determine whether AI agents are behaving reliably before and after they are deployed, a key requirement for Check Point’s new Agentic Network Security Orchestration platform. The platform, announced alongside the Deepchecks purchase, is designed to automate manual network security policy work.
That validation layer matters as enterprises begin giving AI agents access to critical tools and data. Check Point said Deepchecks will help teams evaluate, observe, test and monitor agents in production. Ofir Korzenyak, Check Point’s VP of AI technologies, said the technology will help the company tune and improve multi-agent systems over time, including adapting them to specific customer needs.
These kinds of safeguards are becoming increasingly important in cybersecurity, where agentic systems are being promoted as a way to reduce manual work, accelerate response times, and manage complex environments. But security teams are so far remaining cautious about giving AI systems too much autonomy without strong validation and guardrails.
Check Point’s answer to that concern is to give its agents more context about the environments they are acting on. A core part of the new platform is what the company calls a Network Knowledge Graph, or a live model of a customer’s network environment. According to Check Point, the graph is continuously updated with information about network topology, traffic flows, asset dependencies and configuration data. The idea is to give AI agents current context about how a customer’s network is operating in real time, rather than reasoning on static training data or security logs. Check Point said this context helps the agents interpret existing firewall policies, understand the intent behind older rules, and recommend or carry out changes with more awareness of how those changes could affect the whole environment.
For the 33-year-old Check Point, the acquisition adds talent and technology to a roadmap focused on automating network security operations and modernizing the company’s portfolio beyond the network security and firewall market. The company said the Deepchecks team includes large language model experts and graduates of Talpiot, an elite Israeli technology program. Check Point did not disclose financial terms. Israeli business publication CTech estimated the deal at between $10 million and $20 million, citing industry sources.
Financially, the Deepchecks deal appears to fit Check Point’s preference for smaller, targeted acquisitions. The startup had raised about $14 million before the acquisition, according to PitchBook. The purchase follows earlier 2026 deals for Cyclops Security, Cyata and the Rotate team, which CTech estimated at roughly $150 million combined. That pattern lines up with comments from CEO Nadav Zafrir, who told CTech that Check Point is not pursuing large acquisitions but is looking for strategic technologies as it works through weaker firewall demand and a slightly lowered annual revenue outlook.
