OpenAI to Sign EU Code of Practice, Launches European AI Rollout

OpenAI has announced it intends to sign the EU's Code of Practice for General Purpose AI, subject to formal approval by the AI Board during its upcoming adequacy assessment. The company paired the announcement with a broad commercial and infrastructure push across Europe it calls the "OpenAI for Countries European Rollout."

The Code of Practice is a framework for AI developers to comply with the EU AI Act, the bloc's risk-based regulatory regime for artificial intelligence. Its final approval rests with the AI Board, which represents Member States, making OpenAI's endorsement a high-stakes signal to both regulators and competitors as Europe finalizes how the AI Act will work in practice.

OpenAI framed the decision around Europe's economic stakes. "AI is a core technology for nation building—akin to electricity—that will transform economies and deliver much needed growth by driving productivity," the company said. It argued that unlike earlier digital-era technologies "that were extractive and resulted in economics being pulled out of Europe," AI can expand the economic pie if made broadly available.

EU Member States rank among OpenAI's top markets globally for subscribers, API developers and business customers. The company cited existing deployments across the continent: a partnership with Sanofi on drug development, research collaboration with Sciences Po in France and the Max Planck Society in Germany, a nationwide rollout of ChatGPT Edu to secondary schools in Estonia, and European startups such as Parloa and Pigment building on its platform.

The announcement carries weight because Europe is still debating how aggressively to regulate foundation models. OpenAI pressed that argument directly. "Too often in Europe, the limelight has been taken by regulation," the company said. "Now it's time to flip the script and use this moment to empower Europe's innovators to innovate and builders to build for Europe's future."

What the European Rollout includes

Over the summer and into the fall, OpenAI plans to meet with governments, businesses and stakeholders in Europe on five fronts:

• Data center capacity. OpenAI has submitted Expressions of Interest to join consortiums for the EU's AI Gigafactories initiative, which would build next-generation AI infrastructure and provide sovereign compute capacity. The projects are early-stage.
• AI-powered education. Building on the Estonia rollout, OpenAI says other EU governments have expressed interest in deploying ChatGPT Edu in schools.
• Strategic government partnerships. OpenAI offers EU data residency and localization of ChatGPT for local cultures and languages, and proposes a "new playbook for public-private partnerships fit for the AI-era."
• National AI Startup Funds. OpenAI is interested in helping raise country-level funds with willing governments to seed AI startups, providing local founders with early access to technology and technical support.
• AI uptake. In September, OpenAI will run an AI policy hackathon in Brussels with Allied for Startups, focused on unlocking AI adoption across every sector of the European economy.

OpenAI also urged the EU to assess how regulatory settings affect the pace of AI adoption and the availability of energy and training data, citing Europe's long-standing productivity gap with the U.S., documented in the Draghi report. The company called for the Code of Practice to be aligned with the AI Continent Action Plan announced in April to form a unified European AI strategy.

The safety record OpenAI is leaning on

Signing the Code, OpenAI says, reinforces safety and transparency measures it has pioneered, including its status as a signatory to the Bletchley Declaration (November 2023) and the Seoul Framework (May 2024). The company was one of the first to publish a comprehensive safety and security protocol, its Preparedness Framework, in 2023, and published an updated version in April 2025.

Its ongoing practices include detailed System Cards with major releases, a public Safety Evaluation Hub, an external Red Teaming Network, and the Model Spec describing how model behavior reflects human values and democratic norms.

OpenAI says it will share progress on the European initiatives throughout the summer and into the fall. Whether the AI Board approves the Code in its current form will determine the timeline for compliance obligations across the general-purpose AI industry in Europe.