OpenAI’s Strategic Acquisition of Promptfoo: Fortifying the Future of Secure AI Agents
On March 9, 2026, OpenAI announced the acquisition of cybersecurity startup Promptfoo. The deal is aimed at strengthening the security of large language models (LLMs) and autonomous AI agents. Promptfoo’s technology will be integrated into the enterprise platform OpenAI Frontier to improve the reliability and protection of AI systems used in business environments.
The development of AI agents capable of performing complex tasks autonomously creates major opportunities for businesses. However, it also introduces new security risks. Malicious actors may attempt to exploit such systems to access corporate data, manipulate automated workflows, or disrupt operations. As a result, security has become one of the key requirements for enterprise adoption of artificial intelligence.
Promptfoo was founded in 2024 by security experts Ian Webster and Michael D’Angelo. The company developed tools designed to identify vulnerabilities in language models and test AI systems for potential security risks. According to the company, more than 25% of Fortune 500 companies already use its solutions to stress-test their AI implementations. Despite strong market adoption, the startup raised about $23 million in funding and was valued at approximately $86 million. The financial terms of the acquisition were not disclosed.
Promptfoo’s technology focuses on identifying vulnerabilities in AI systems. Its tools allow automated security testing, simulate potential attacks, and analyze complex AI-agent workflows for hidden risks. The platform will also enable continuous monitoring of security and compliance during live operations.
OpenAI also confirmed that it will continue supporting Promptfoo’s open-source tools, helping to establish broader security standards across the AI industry.
The acquisition reflects a broader industry trend: security is becoming a fundamental component of artificial intelligence development. For industries such as finance, healthcare, and government, the reliability and protection of AI systems are now critical requirements.
OpenAI’s move highlights that the future of AI will be defined not only by the power of models, but also by how effectively they can be secured and safely deployed.