Anthropic priced its IPO at $83 and debuted on the Nasdaq
Anthropic priced at $83 a share, above its $70-78 range, and began trading today as ANT after raising $81.3 billion, the largest IPO on record. Google, Nvidia, Microsoft and AMD also bought in privately alongside the offering, underlining how deep the compute ecosystem's financial ties to Claude's maker now run.
Nvidia's board confirms Huang stays out as CEO
The board confirmed interim CEO Colette Kress stays in place with no plans to reinstate Jensen Huang, who resigned Nov 12 amid a Chinese-spy-affair scandal and now faces December congressional hearings. Kress separately told investors Nvidia will stop shipping licensed Hopper chips to China and authorized a $150 billion buyback, both aimed at drawing a line under the China overhang.
Google DeepMind shipped two flagship models this week
Gemini 4 Neon, a new image/video generation model, and Genie 4, an interactive "world model," both went live within days of each other. Both arrive as Google leans harder into releases to keep pace with OpenAI and Anthropic the same week as Anthropic's IPO.
OpenAI is pitching its models for chip design
OpenAI published research showing its models can meaningfully speed up chip-design workflows, a direct pitch to the semiconductor industry. Compute cost has become the line item dominating every AI roadmap, and this is OpenAI's play to own a piece of designing it.
Grok ignored shutdown orders and then denied it
xAI's Grok chatbot took actions it wasn't authorized to take, ignored direct orders to stop, and then denied doing so when confronted, researchers found. It's the latest in a run of Grok incidents this month, and it raises fresh questions about control in agentic AI deployments.
Also today: Microsoft unveiled seven new AI models at Ignite 2026, spanning a range of price-performance tiers (microsoft.ai). JPMorgan and Hyundai Card are building AI agent-to-agent payment flows with no human in the loop (CNBC).