Important date: On July 10, 2025, the European Commission published the definitive General-Purpose AI Code of Practice (CoP). The deadline for signing is August 1, 2025.
The playing field of major AI players
On July 10, 2025, the European Commission published the definitive General-Purpose AI Code of Practice (CoP). The document is voluntary, but those who sign will get a fast-track route to comply with the AI Act starting August 2, 2025, with less risk of heavy audits or fines. Yet the major model builders are responding very differently. Below you'll find the current playing field and the motivation that companies themselves provide.
| Company | Status | Core of their own motivation |
|---|---|---|
| OpenAI | Signing (intention announced) | Sees the Code as a "springboard" for a European AI ecosystem and praises the simplicity of complying with the AI Act (OpenAI) |
| Anthropic | Signing | Believes the CoP strengthens transparency, safety, and accountability; aligns with their Responsible Scaling Policy (Anthropic) |
| Microsoft | "Likely" signing | Brad Smith: "We want to be supportive; direct contact with the AI Office is welcome" (Reuters) |
| Mistral AI | Confirmed signing | French scale-up reports via LinkedIn that it is joining the code (LinkedIn) |
| Meta | Refusing | Joel Kaplan calls the CoP "overreach" that hinders innovation and creates legal uncertainty (TechCrunch) |
| Alphabet / Google DeepMind | Still deliberating | Spokesperson: "We will review the code… Europeans should have access to first-rate, secure AI models" (The Wall Street Journal) |
| Amazon (AWS) | Still deliberating | Wants the CoP to "exclusively facilitate compliance" and not add extra rules (EU About Amazon) |
| Apple | No public position | Also absent from earlier EU AI Pact signing; silence so far (TechCrunch) |
| xAI (Elon Musk) | No public position | No official response found (status: unclear) |
Why sign?
Transparency & certainty
OpenAI, Anthropic, Microsoft, and Mistral emphasize that the CoP is a practical guide to comply with the AI Act. By signing now, they save audit stress later and can serve the entire EU market with one harmonized framework. For European players like Mistral, the signal of trust toward Brussels also counts.
Competitive advantage
OpenAI immediately links it to its "OpenAI for Countries" program: those who conform early can secure public and government contracts in Europe faster. (OpenAI)
Strategic advantage: Early signatories can use their "CoP label" as a selling point toward EU governments and enterprise customers.
Why refuse?
Meta's objection
Meta argues that the Code "goes beyond the AI Act," particularly through stricter obligations around dataset documentation and copyright filters. The company fears this "stifles" the pace and scope of frontier model development in Europe (TechCrunch).
Lobby dynamics: Meta's rejection could lead to renegotiations, especially if other Big Tech players still sign and give Brussels extra legitimacy.
Why no signature yet?
Companies still deliberating
- Google DeepMind / Alphabet: Studying the document; wants certainty first that the requirements align with their Gemini roadmap. (The Wall Street Journal)
- Amazon: Lobbying for "compliance-supportive, not extra regulatory" character; will only sign when this is clear. (EU About Amazon)
- Apple & xAI: have not made any public statement to date. Apple's absence from earlier EU AI initiatives suggests the company wants to get internal governance processes in order first (TechCrunch).
Deadline approaching: The AI Office will publish an official list of signatories starting August 1, 2025. Those who haven't signed by then will miss the "light" compliance route and fall under regular—stricter—AI Act enforcement (Digital Strategy Europe).
What does this mean for the ecosystem?
Regulatory asymmetry
Companies that do not sign may face heavier controls or fines (up to 7% of revenue). This creates a clear separation between "compliant" and "non-compliant" players in the market.
Market positioning
Early signatories can use their "CoP label" as a selling point toward EU governments and enterprise customers. This gives them a competitive advantage in a market that increasingly values compliance.
Opportunities for EU startups
A clear compliance guide lowers the barrier to entry; this can help local models (Aleph Alpha, Helsing AI, etc.) gain scale.
Benefits for EU startups
The Code of Practice offers European AI companies:
- Clear guidelines for compliance
- Level playing field with major international players
- Access to government contracts through early adoption
- Reputation advantage as "EU-compliant" providers
The new focal point of European AI strategy
The Code of Practice is not yet law, but it is the new focal point of European AI strategy. Those who sign benefit from fast AI Act certification and political capital; those who refuse gamble on later, possibly more lenient negotiations.
The coming weeks - until August 1 - will determine whether the "we sign" camp becomes the norm, or whether major absentees like Meta and possibly Apple will set the tone. One thing is clear: the CoP has reshuffled the cards in the European AI market.
Practical tip: Organizations that implement the Code now position themselves for competitive advantage in a market that increasingly values responsible AI development.
Written on July 28, 2025
Want to know more about implementing the Code of Practice in your organization? Contact us for a personal consultation.