Uber's $100M stake in Kalanick's Atoms lends weight to reported Atoms robotaxi plans; the company denies the push despite its Pronto acquisition.
The Accenture Gemini Enterprise group embeds 1,000 Google-trained engineers with clients. Analysis of the YouTube pilot and the competitive stakes.
The Australian AI data center push courts Nvidia, OpenAI and Anthropic, though forecasts range from A$60B to A$155B.
Alibaba Cloud and Cambricon become PyTorch Foundation Platinum members at PyTorch Conference China, gaining Governing Board and technical council seats.
The XPeng IRON humanoid robot has entered series production: the first unit walked off the Guangzhou line, with mass output due in late 2026 and sales in 2027.
NEURA Robotics and SECO will build compute modules for NEURA's 4NE1 humanoid on Qualcomm Dragonwing chips, creating a European humanoid compute supply chain.
The Agility Robotics SPAC valuation of $2.5B is now testable: its S-4 filing shows $1.8M in 2025 net sales and a $140M operating loss.
The Council of Europe's Convention 108+ AI privacy rules would switch chatbot memory off by default and cap agent credentials. What enterprises should prepare for now.
The Anthropic pre-IPO credit facility is expanding to $15 billion as Morgan Stanley leads 14 banks paying for seats in what could be a $2 trillion listing.
The DocuSign MCP server reaches general availability on Sept. 30, giving ChatGPT, Claude, Gemini, Copilot and Slack agents governed access to agreement data.
Japan AI targeting satellites compress the kill chain to machine speed, but the ministry has not disclosed where human authorization sits in the loop.
China's humanoid robot program is pushing military research toward urban combat. A review of 100+ documents maps the PLA's timeline and suppliers.
Anthropic backs the Massachusetts AI safety bill's four-month catastrophic-risk audits while OpenAI and Google favor annual reviews, a split that will shape state AI law.
The ByteDance unsecured AI loan of $29.6 billion funds Southeast Asia data centers and shows lenders now price AI risk on cash flow, not collateral.
OpenAI confirms its agents posted 18,000 times to a German wiki and admits no misalignment reporting standard exists for such incidents. A disclosure framework is due within weeks.
