Microsoft data center expansion targets 38 GW by 2032, adding 26 GW of AI compute under a $175B 2026 spending plan after capacity shortages.
Apple's iPhone 18 Pro and Pro Max start at $1,199 and $1,299, up $100, with the first 2nm A20 Pro chip and pre-orders opening September 12.
Zankore $3.1 billion GPU loan funds Nvidia-powered AI capacity in Indonesia, shifting capex and credit risk onto five banks.
Florida AI criminal liability plan from AG Uthmeier would add fines, monitorship and business suspension for chatbot makers tied to crimes.
Google's Dreambeans AI story feed is now free for all US accounts 18+ on Android and iOS, drawing on Gmail, Photos, Calendar, Search, YouTube and Gemini.
Amagi's four AI-native broadcast tools head to IBC 2026 as record Q1 FY27 revenue funds a cloud production pilot with a major U.S. network.
OpenAI's ChatGPT for Financial Services bundles GPT-6 Astra with LSEG and PitchBook data for junior bankers, but compliance and audit gaps remain.
Anthropic's AI misuse report details 154 pages of blocked Claude abuse, from bioweapon research to Yemeni missile programs and model distillation.
Microsoft's September 2026 Patch Tuesday fixes a record 974 CVEs, including two exploited zero-days in the Windows Update Stack and ALPC.
SpaceX is selling orbital compute to Anthropic and Google at over $3B a month, but the first satellites do not launch until Q4 2027.
This period's AI funding tracker covers twelve items, led by two $3B rounds for Crusoe and Moonshot AI plus Nvidia's $2.5B Thinking Machines stake.
Minneapolis's proposed robotaxi ordinance would require a paid human safety monitor in every driverless car, reshaping fleet economics and city permits.
Apple Watch Audio Intelligence brings Live Rewind, Siri Recap and sound alerts to Series 12 and Ultra 4, with raw audio locked inside Secure Exclave.
Apple's iPhone Duo foldable opens to 7.6 inches for $1,999 with the A20 Pro chip, October 16 pre-orders and the iPhone 18 Pro from $1,199.
DeepSeek V4.1-Flash ships MIT open weights at $0.15 per million input tokens, cuts KV cache to 890 bytes per token and retires V4 Pro on September 14.
