OpenAI Agents API enters public beta, hosting the Codex harness at no extra fee and shifting the enterprise build-versus-buy call on agent orchestration.
OpenAI backs mandatory national AI safety rules, urging Congress to act before December while 29 states enforce their own AI laws.
China's AI infrastructure plan commits $532 billion to reach 9,800 EFLOPS by 2030, with 100,000-card clusters built on domestic chips.
Skild AI S1 robot foundation model learns ten-minute tasks from one video with no fine-tuning. What its 66% success rate means for industrial buyers.
Amazon Quick desktop app reaches general availability on macOS and Windows, taking AWS's agentic assistant directly up against Microsoft 365 Copilot.
Nvidia is pitching AI compute as an investable asset class, pairing $279B in supply commitments with a $500B Wall Street financing push.
The Katzenberg AI video studio pairs Jeffrey Katzenberg with ex-Sora lead Bill Peebles and ex-Dropbox CFO Sujay Jaswa to build video models for filmmakers.
EU Cyber Resilience Act reporting starts September 11, 2026, with a 24-hour vulnerability clock, an unproven ENISA portal and no rules for AI agents.
Newsom signed California AI audit laws SB 813 and AB 1405, creating a state auditor registry backed by OpenAI and Anthropic, and pressed for federal rules.
The CDAO awarded Red Cell Partners a $100M OTA to test Pentagon shared savings AI contracts, paying vendors only from costs they eliminate.
Meta acquired Stockholm startup Stilla.ai to accelerate its Meta Business Agent, aiming to turn WhatsApp, Messenger and Instagram chats into completed sales.
Anthropic's Claude Fable 5.1 ships day-one across four cloud platforms at flat $10/$50 pricing, with a 75% cache cut and a gated Mythos 5.1 tier.
OpenAI ends its $1 federal pilot, moving OneGov ChatGPT pricing to metered usage at 50% off through 2028 with no platform fee or spend commitment.
Samsung runs Mistral Large on-premises in its own fabs as sovereign AI shifts from slogan to budget line, with the EU-backed Scaleup Europe Fund co-leading.
Salesforce's Trusted Enterprise AI Harness adds six governed capabilities and an AI Control Plane to register, secure and meter enterprise AI agents.
