← Back

Wire · regulatory

AI Act Adapted to ChatGPT Rather Than Breaking, Uuk Argues

Published

26 July 2026

Topic

regulatory

Sectors

AI Frontier Models

Geography

Europe

Source

Read at aiweekly.co

Verified

Fusion42 · 26 July 2026 · Fusion42 review

EU AI Act enforcement powers activate 2 August 2026, with compliance deadlines extending to August 2027 for models already deployed. Signatories to the Commission's Code of Practice include Anthropic, Google, Microsoft, and OpenAI; Meta refused and xAI signed only the safety chapter.

This Wire brief sits within Fusion42's coverage of AI Frontier Models. Wire is Fusion42's founder-focused intelligence feed: each story is connected to the funds and startups it names — every one with a live profile on Raise or Scout — so founders can follow the capital and the momentum behind the headline rather than just the headline itself. Wire analysis is one of the live surfaces Arthur reasons over.

◆ The Wire takeaway

If you ship a foundation model to Europe, you have until August 2027 to comply with new EU rules, but the Commission's enforcement teeth arrive 2 August 2026—twelve months before your deadline. Meta and xAI have already signalled they won't sign on fully; the real test is whether the 125-person AI Office can actually enforce against them.

Related on Wire

Topics

AI Frontier Modelsai-regulationeu-ai-actcompliance-deadlinefoundation-modelsenforcement