Key Takeaways

  • 1Password crossed $500M in ARR, posting 90% year-over-year growth in new ARR for Q3 and 150% growth in new B2B logos over Q3 of the prior year.
  • Autonomous AI agents operating in development cycles frequently hold permanent API tokens and credentials, creating massive attack surfaces.
  • 1Password for Agents was developed alongside major AI labs to enforce just-in-time, scoped secret delivery using confidential computing.
  • Secrets must never touch model context windows or local config files where bad actors or rogue agent loops can read them.
  • Because the financial cost to launch automated software attacks is collapsing toward zero, every developer and company is now an active target.

Standing Access Is a Ticking Time Bomb

When developers wire up autonomous AI agents to build software, run tests, or manage cloud infrastructure, they usually take the path of least resistance: they generate a long-lived API key, paste it into an environment file, and let the agent run. That habit creates immediate enterprise risk.

1Password CEO David Faugno argues that this approach cannot survive the transition to agentic workflows. As Faugno noted: “especially in the development cycle agents need access and that access when it's left standing around is a ticking time bomb.” Permanent credentials sitting in files or persistent sessions invite disaster. If an agent gets hijacked via prompt injection, hallucination, or compromised tool calls, any standing secret in its path belongs to the attacker.

This security problem is converging with explosive business adoption. 1Password reached half a billion dollars in ARR on the back of rapid enterprise and developer growth. Faugno pointed out that “new ARR for Q3 was 90% growth year-over-year. And we're adding new logos on the B2B side at 150% growth over quarter over last year, third quarter.” As companies scale their agent fleets, securing non-human identities has become the primary operational hurdle.

Secrets Must Stay Out of Model Context

Traditional credential management assumes a human sits behind the screen, reviews permissions, and copies passwords into an input box. Autonomous agents break that assumption. They run thousands of requests an hour, spin up ephemeral instances, and read wide swaths of code and system memory.

To counter this risk, 1Password built 1Password for Agents in partnership with top AI research labs. The core thesis is minimal exposure. As Faugno explained: “one password for agents is really built around the premise of just give the agents what they need. Just just when they need it, just enough with full visibility.”

Mechanically, this requires zero-knowledge architecture and confidential computing. When an agent requests a secret to complete an API call or database migration, the credential must decrypt only inside a protected runtime. It cannot sit in plaintext in a context window or log file. Faugno described the technical barrier: “We do a lot with confidential computing to make sure that when those secrets get decrypted if they need to be decrypted for a reset or otherwise it's never in a place where a bad actor or an agent can find it and do something with it.”

The urgency is driven by unit economics. Threat actors no longer need large teams to probe enterprise perimeters; they can spin up autonomous agent swarms to run continuous vulnerability scans. According to Faugno: “having that end to-end encryption and having that just in time delivery where the secret is never exposed or can't be read from a file is going to be really important because the cost to attack is going to zero. That means everybody is a target.”

What to Do With This

Audit your development and staging environments tomorrow morning. Identify every persistent API key or service account assigned to local coding assistants and automated testing bots, revoke tokens with standing write permissions, and replace them with short-lived, scoped credentials that expire after the task finishes.