# regulation

Escalating compute ambitions are colliding with real-world infrastructure, with a Fortune investigation projecting a 17GW energy appetite for advanced models while AMD’s GAIA broadens access through new Linux support. Governance tensions are sharpening, as a preliminarily approved $1.5B Bartz v. Anthropic settlement and looming privacy policy changes put data rights and accountability at the forefront. Emerging capabilities from synthetic companions to AI-designed bacteriophages underscore the urgency of safety and incentive design.

Policymakers are moving from pilots to enforcement, with OECD cataloging 200 public-sector AI uses and Italy advancing rules criminalizing deepfakes and mandating human oversight. At the same time, Nvidia’s $5B stake in Intel signals a hardware reset to make AI features native, even as interfaces mature unevenly and creators wrestle with credit, consent, and hallucinations.