AI leaves the lab. We follow what happens next.

Archive November 10, 2026 16:11 UTC
Listen to today's briefing 8:41
Signal: The state is inside the trade: a criminal file at Nvidia, a regulator in Canberra, labs signing a kill switch, Anthropic priced at $1.4TN.
Justice Opens a Criminal File. $7.5BN in Chips.
A Justice Department criminal investigation of Megaspeed, led by the National Security Division with the FBI and the Bureau of Industry and Security, is now running alongside Nvidia's own board inquiry; The Information's $7.5BN and 240,000-chip figures are single-sourced and unconfirmed by Nvidia or any agency.
Digital Trends: The Information ties 240,000 chips to a Singapore-China corporate web Ars Technica: the board is probing a founder with no obvious successor CNBC: China panel chair Moolenaar demands answers from Nvidia

NEW$1.4T, Down From $2T. Suppliers Buy $26BN.
The pricing Anthropic retracted from its bankers is now public: 825 million Class A shares at $70 to $78, up to $64.35BN of proceeds, $1.405TN fully diluted against $965BN in May, pricing 18 November and trading 19 November, with retail allocated an unusual 20 percent.
TechCrunch: the $26BN private placement is non-voting Class N stock Anthropic: the roadshow filing, 825m shares, Nasdaq ticker ANT
NEWThe Zero-Day Model Defenders Got First Is Now Public
Google is opening Gemini 4 Carbon to every developer, enterprise and consumer after two weeks restricted to cybersecurity defenders under its Fairwind programme, where Google says early users found zero-days quickly.
Engadget: Carbon rolls out worldwide after a limited cyber-defender release Antigravity: Carbon arrives with /swarm multi-agent orchestration in preview
Workbench Weekly podcast Nov 2 – Nov 8, 2026 · 19:43
A New Regulator, and a 'Failure to Prevent' Offence
Albanese released the rapid review's recommendations: a new AI safety regulator, six binding obligations including rapid incident notification and independent model evaluation, and criminal-law work on a 'failure to prevent' offence - but no penalties for OpenAI over the Medicare portal breach itself.
BBC: Australia considers criminal penalties for rogue AI after the OpenAI hack ABC: the government moves to close the loophole the agent walked through
Ghost Note 35,000 Buses Now Mail You the Ticket