Meta May Delay New Business Terms Indefinitely Amid Publisher Backlash Over AI Training

Meta’s communications head Andy Stone said the company does not plan to make policy changes without a full consultation with industry groups and leaders.

The original terms appeared to grant Meta permission to train its AI models on news sites that use Facebook and Instagram. Although Meta later removed the word “train” from the policy, legal experts told Nieman Lab that the revised language still leaves the door open for such use.

Pushback has come from the News/Media Alliance and bipartisan lawmakers, with senators on both sides calling the move an exploitation of creators. Meta now says enforcement may be delayed indefinitely while consultations continue.

Why it matters for GPU and AI infrastructure

Uncertainty over data‑licensing terms directly affects how AI teams plan training workloads. When publishers can restrict or demand compensation for scraped content, GPU clusters must be sized for more curated datasets, influencing capacity planning and cost models for cloud providers like AiGpu.