OpenAI to acquire AI security startup Promptfoo

OpenAI is acquiring Promptfoo, an AI security platform that helps identify and remediate vulnerabilities in AI systems during development and will integrate Promptfoo’s technology into OpenAI Frontier after the deal closes.

OpenAI said the Promptfoo team has built tools trusted by over 25% of Fortune 500 companies and maintains an open-source CLI and library for evaluating and red-teaming LLM applications.

OpenAI will embed automated security testing into Frontier to detect issues such as prompt injections, jailbreaks, data leaks, tool misuse, and out-of-policy agent behaviors.

OpenAI said testing will be integrated into development workflows and that Frontier will add reporting and traceability to support governance, risk, and compliance expectations for enterprise AI.

OpenAI will welcome the Promptfoo team, led by Ian Webster and Michael D’Angelo.

The acquisition is subject to customary closing conditions.

🔗 Source: OpenAI

🧠 Food for thought

Implications, context, and why it matters.

Promptfoo had strong funding and a big-company focus before the OpenAI deal

Before the acquisition, Promptfoo raised an $18.4 million Series A led by Insight Partners, with participation from Andreessen Horowitz 1.
It used an open-source go-to-market strategy, offering a free community product to bring in developers while selling custom-priced plans to large teams 2.
As it geared up for its first large-company customers, Promptfoo added enterprise-grade authentication (sign-in controls used by big companies) in one week to support Fortune 500 buyers using FusionAuth (an identity and access management platform), according to a FusionAuth case study 3.

The acquisition adds momentum to the push for safer AI agents in business settings

The deal centers on AI agents (AI systems that can take actions on a user’s behalf) that can run code and reach sensitive data 4.
Agents expand the attack surface (the number of ways a system can be exploited), so OpenAI appears to be moving early on security before companies hand over high-stakes workflows 5.
This approach fits a growing view that AI security and evaluation belong in the AI infrastructure stack, similar to CI/CD (continuous integration/continuous delivery) pipelines in traditional software development 6.
With Promptfoo folded in, OpenAI plans a standardized, automated way to manage these risks, which could reduce a deployment bottleneck for more advanced AI systems 7.