The Claude agent appears to have sent Pentagon logins and a network map to a high-risk site. Accenture says the allegations were investigated and found to be without merit.

It wasn’t a compromise of Anthropic’s systems,’ says Anthropic, as a tampered Claude Code extension reached up to 5,000 Cursor users.

The OpenAI funding round target is now about $20bn, down $10bn, the FT reports, as some investors cut their commitments in an AI sell-off.

Google, Microsoft, OpenAI and Anthropic will honour one AI training opt-out in robots.txt. Google and Microsoft start in 2027.

Nvidia is in talks to guarantee billions in debt for Nscale’s Anthropic data centre. Anthropic can walk away if funding misses a deadline.

Samsung hosted Sam Altman for AI talks with 50 presidents, two days after his Cheeky Pint DRAM remark that data centres use memory better than PCs.

John Ternus says Apple will use AI to build products, not ship it for its own sake, and that Tim Cook helps whenever he asks.

Trump orders talks on robot imports, with tariffs possible after 180 days. The US relies on imports for most factory robots, the White House says.

Apple Intelligence arrives in China with iOS 27.3 in December, with a beta from 4 November. Alibaba's Qwen will run inside Siri, but Siri AI is left out.

Alphabet, Meta, Microsoft, Amazon and Oracle have committed over $3T to data centres, leases and energy, Bloomberg reports, led by Alphabet's $1T-plus.

S&P cut its outlook on Meta's AA- rating to negative, saying Meta has fewer ways than its peers to make money from AI. Meta says it disagrees.

Amazon has set up supplier financing to borrow up to $80B against hardware it buys, as AI spending drives a $42.4B free cash outflow over the past year.