OpenAI Acquires AI Security Platform Promptfoo: Red-Teaming Tool Used by 350K Developers Joins Frontier

OpenAI announced on March 9 its plan to acquire Promptfoo, an AI security testing platform, integrating its technology into the enterprise-grade Frontier product. Founded in 2024 by Ian Webster and Michael D'Angelo, Promptfoo has been used by over 350,000 developers with 130,000 monthly active users, serving more than 25% of Fortune 500 companies. The platform specializes in adversarial testing of AI systems, systematically detecting vulnerabilities like prompt injections, jailbreaks, and data leaks.

The acquisition reflects accelerating consolidation in the enterprise AI security space. As agentic AI transitions from experimentation to production deployment, demand for AI system security testing is growing exponentially. Promptfoo's open-source red-teaming suite will continue to be maintained while its core technology brings automated security testing, red-teaming capabilities, and compliance monitoring to Frontier.

This deal also represents the latest move in the OpenAI-Anthropic competition for enterprise security. Anthropic previously launched Claude Code Security, while OpenAI responded with Codex Security. Acquiring Promptfoo signals OpenAI's shift from relying solely on internal R&D to rapid capability acquisition through M&A, preparing for the large-scale deployment of AI agents.

OpenAI Acquires Promptfoo: A Critical Step in AI Security Infrastructure

Acquisition Context and Strategic Intent

OpenAI officially announced the acquisition of AI security testing platform Promptfoo on March 9, 2026. While the specific transaction amount was not disclosed, the strategic significance of this deal extends far beyond its financial dimensions. Promptfoo will be integrated into OpenAI's enterprise AI platform Frontier—the company's solution for building and operating AI Coworkers.

Promptfoo's core value lies in its comprehensive AI system security testing infrastructure. Since its founding in 2024, the platform has accumulated over 350,000 developer users, 130,000 monthly active users, and adoption by more than 25% of Fortune 500 companies. Its technology stack covers the complete lifecycle of AI security testing: from static code scanning to dynamic red-teaming, from prompt injection detection to jailbreak defense, and from data leak prevention to compliance monitoring.

The Enterprise AI Security Landscape

This acquisition comes at a critical juncture. As AI agents rapidly transition from proof-of-concept to production deployment in 2026, enterprise AI security challenges are expanding from 'model security' to 'system security'—organizations need to not only ensure LLMs don't produce harmful outputs, but also defend against various attack vectors that could exploit AI agents during multi-step task execution.

Promptfoo's competitors include Robust Intelligence (acquired by Cisco), CalypsoAI, and HiddenLayer. However, Promptfoo holds clear advantages in open-source community influence and enterprise penetration, with its red-teaming framework becoming the de facto industry standard.

OpenAI vs Anthropic: The Security Arms Race

This acquisition continues the OpenAI-Anthropic competition in enterprise AI security. The timeline clearly illustrates their offensive and defensive moves:

  • February 2026: Anthropic releases Claude Code Security
  • February 2026: OpenAI responds with Codex Security
  • March 2026: OpenAI acquires Promptfoo, upgrading from tool-level to platform-level

OpenAI CTO Srinivas Narayanan stated that Promptfoo brings 'deep engineering expertise in evaluating, securing, and testing AI systems at enterprise scale.' This suggests OpenAI recognizes that building security capabilities solely through internal teams can no longer match the growth rate of AI agent deployment.

Impact on the Developer Ecosystem

Promptfoo co-founder Ian Webster committed that the open-source testing suite will continue to be maintained, covering red-teaming, static scanning, and evaluation across multiple AI models and applications. This means Promptfoo won't become a closed OpenAI tool but will operate on a dual-track model of open-source plus enterprise.

For development teams building AI agents, this acquisition sends a clear signal: AI security testing is no longer optional—it's a prerequisite for production deployment. Just as CI/CD became standard in software development, Continuous AI Security Testing is becoming the new industry norm.

Outlook

With Promptfoo joining Frontier, OpenAI's enterprise AI platform evolves from 'model services + development tools' to a complete stack of 'model services + development tools + security infrastructure.' For enterprise decision-makers evaluating AI platforms, this could become a significant differentiating factor.

Sources:

  • [OpenAI Official Announcement](https://openai.com/index/openai-to-acquire-promptfoo/)
  • [Promptfoo Blog](https://www.promptfoo.dev/blog/promptfoo-joining-openai/)
  • [PYMNTS Coverage](https://www.pymnts.com/artificial-intelligence-2/2026/openai-plans-to-acquire-promptfoo-to-secure-agentic-ai/)
  • [Seeking Alpha Analysis](https://seekingalpha.com/news/4562413)