WorkOS raises $100M Series C at $2B valuation, betting on AI agent identity infrastructure

Summary

WorkOS has raised $100 million in a Series C at a $2 billion valuation, marking its first outside capital in over four years. Founded in 2019, the company builds enterprise authentication and identity infrastructure that lets software companies including OpenAI, Anthropic, Perplexity, and Cursor sell into large corporate accounts without building compliance, SSO, and permissions tooling themselves.

The AI tailwind

Founder and CEO Michael Grinich credits the AI wave with accelerating WorkOS's adoption. Enterprise security scrutiny of AI tools runs higher than it ever did for previous SaaS generations. Where Dropbox had years to figure out enterprise go-to-market, today's AI companies are being pushed upmarket within their first year. That compression leaves no time to build identity infrastructure in-house, and WorkOS has become the default plug.

The free tier supports up to 1 million users, with pricing kicking in only when customers close enterprise deals. This Stripe-like model aligns WorkOS revenue with customer revenue. Integration time has dropped to 7–8 minutes through an AI-powered CLI installer that analyzes a customer's codebase, runs an architecture review, and deploys the relevant WorkOS components automatically. Grinich now kicks off sales calls by asking prospects to run the installer in the background while he walks through the dashboard, and the environment is ready before the call ends.

Agent identity

The next growth layer is AI agent infrastructure. As companies deploy agents that need to take actions across internal systems and customer environments, those agents require scoped permissions, access controls, and audit trails. WorkOS is building what Grinich calls an identity fabric to sit across agent deployments, giving teams building agents the connectivity and security guardrails that enterprise buyers will demand. A recent incident where a Claude instance reportedly went rogue and deleted emails underscores the urgency of this problem.

Within WorkOS itself, Grinich runs AI across every function: sales reps using Claude Code, finance teams running AI workflows, company-wide hackathons. The AI installer is framed as the early version of what a forward-deployed engineer does, analyzing a client's architecture and prescribing the right integration path, compressed into minutes rather than weeks of solutions-engineering cycles.