OpenAI Acquires Promptfoo for Frontier Security

Key Takeaways

• OpenAI is acquiring Promptfoo to integrate AI security testing directly into its Frontier agent platform.
• Promptfoo’s red-teaming tools will help enterprises detect prompt injections, jailbreaks and data leaks.
• Built-in governance and evaluation tools aim to make large-scale AI agent deployment safer for enterprises.

OpenAI announced on March 9, 2026, that it will acquire Promptfoo, an AI security platform used by over 25% of Fortune 500 companies. According to company officials, the acquisition will integrate Promptfoo's vulnerability detection and red-teaming capabilities directly into OpenAI Frontier, the company's platform for building enterprise AI agents.

The Promptfoo team, led by co-founders Ian Webster and Michael D'Angelo, developed an open-source CLI and library for evaluating and testing large language model applications. OpenAI said it will continue building the open-source project while advancing integrated enterprise capabilities within Frontier.

Table of Contents

• New Features Planned for OpenAI Frontier
• AI Agents Force a Rethink of Enterprise Security Testing
• OpenAI's Aggressive Expansion Strategy
• OpenAI Background

New Features Planned for OpenAI Frontier

The Promptfoo acquisition will expand Frontier with built-in security and evaluation tooling. The initial integration focuses on several core security functions designed to help enterprises test, monitor and validate AI agents at scale.

Feature	How It Works
Platform-native security testing	Automated red-teaming to detect prompt injections, jailbreaks and data leaks
Workflow-integrated evaluation	Security checks embedded in development pipelines
Oversight and accountability	Integrated reporting and traceability for governance and compliance
Open-source continuity	Continued development of Promptfoo's CLI and evaluation library

AI Agents Force a Rethink of Enterprise Security Testing

Deploying AI agents at scale demands native security testing, governance tooling and continuous validation baked into the platform itself.

Enterprises increasingly recognize that AI agent frameworks present distinct testing challenges compared to traditional software:

• Non-deterministic outcomes requiring specialized validation approaches
• Integration complexity with existing systems and data
• Security vulnerabilities necessitating continuous red teaming
• Multi-modal interactions spanning text, image, audio and video formats

Governance frameworks must now blend risk management, compliance and ethics into systems that keep pace with agentic AI. Yet only 10% of deployments have moved beyond experimentation into production.

OpenAI's Aggressive Expansion Strategy

OpenAI made three strategic acquisitions in 2024 and 2025:

• It acquired the team behind Multi in June 2024, a startup focused on screen sharing and collaboration.
• It bought Statsig for $1.1 billion in September 2025, bringing founder Vijaye Raji on as CTO of Applications.
• It acquired Neptune for under $400 million in December 2025 to enhance model governance.

On the product front, OpenAI most recently released GPT-5.4, which featured a one-million token context window and surpassed human knowledge workers on the GDPval, which tests agents’ abilities to produce well-specified knowledge work across 44 occupations.

In Q2 of 2025, the company launched a ChatGPT App Store.

OpenAI also diversified its compute infrastructure through massive partnerships, including a $100 billion agreement for Nvidia to become its key AI chip supplier, a $38 billion cloud contract with AWS and a $50 billion investment from Amazon.

OpenAI Background

OpenAI, founded in 2015, develops AI solutions for enterprise technology leaders and developer teams. Key offerings include the GPT model family, ChatGPT, developer APIs and agent-building platforms such as AgentKit.