Cloud Security Startup Upwind Raises $400M at $3.8B Valuation to Secure AI Agents

The rapid deployment of autonomous AI agents and complex cloud architectures has created a visibility gap for enterprise security teams. Traditional static scanning tools often generate thousands of isolated alerts without the context needed to prioritize them, leaving DevOps and security teams struggling to identify which vulnerabilities pose actual risks in a live environment.

To address this bottleneck, San Francisco-based Upwind Security has raised $400 million in a Series C funding round, catapulting its valuation to $3.8 billion. Led by Bessemer Venture Partners, the round includes participation from Greylock Partners, Craft Ventures, and Salesforce Ventures. The capital injection follows a year of rapid growth for the startup, which reported that its annual recurring revenue (ARR) more than doubled to over $50 million in 2026.

Unlike conventional security posture management tools, Upwind operates by analyzing runtime data directly from cloud workloads, application programming interfaces (APIs), and AI agents. By unifying these signals, the platform provides a shared view for developers and security teams. When an anomaly or threat is detected—such as an AI agent exhibiting unusual behavior—the system allows engineers to trace the finding directly back to the underlying code.

This massive funding round highlights a broader industry shift toward runtime-centric security as organizations demand tools that can protect AI-driven applications. Founded by Amiram Shachar, who previously sold Spot.io to NetApp, Upwind is aggressively positioning itself against market heavyweights like Wiz, CrowdStrike, and Palo Alto Networks. As cloud environments become increasingly autonomous, the ability to secure AI agents in real time is transitioning from a niche feature to a core enterprise requirement.