Google released Gemini 4 Hydrogen for fast everyday work, and Anthropic started the Claude Meetings beta, letting Claude join Google Meet meetings directly. In the US, four civic groups sued five major AI companies seeking a freeze on model development, and reports of an Nvidia CEO change have heightened tension among Korean semiconductor and AI companies.
Google released Gemini 4 Hydrogen on November 12, rolling it out globally in stages across Search, the Gemini app, and developer tools. The model is focused on handling everyday work faster and supporting agent workflows, extending the Gemini 4 lineup after the security-focused Gemini 4 Carbon — first unveiled on October 28 — was opened widely to developers, enterprises, and scientists on November 10.
Anthropic has started the Claude Meetings beta for paid Team and Enterprise plans. Claude joins Google Meet video meetings as a participant, answering questions and drafting project proposals and schedules live with attendees. External meetings are excluded by default, and safeguards require prior approval for looking up internal data.
On November 12, four nonprofits sued Anthropic, Google, Meta, OpenAI, and SpaceX in San Francisco Superior Court. They ask the court to block the development and deployment of models more powerful than each company's current frontier models, citing risks of cyberattacks, bioterrorism, and loss of control over autonomous systems. Filing the suit alone does not halt development; whether the court grants the injunction is the key question.
Google DeepMind's AI agent framework AlphaProof Code, unveiled on November 10, mathematically verified the core security properties of the code behind Signal Messenger's encryption protocol. It shows AI assisting security research reaching beyond model performance competition to formally proving the safety of a real protocol.
Reports on November 12 that Nvidia CEO Jensen Huang faced dismissal put Samsung Electronics, SK hynix, Hyundai Motor Group, LG, Naver, and other Korean companies allied with Nvidia on AI on edge. A management change could slow next-generation HBM4 and AI chip development and investment and force revisions to long-term supply contracts. Nvidia's board said the appointment of a permanent CEO is still at the search stage.
OpenAI has reorganized its robotics research team into a core research division and an applications and hardware division. The reorganization aligns with a strategy of controlling robots directly with a general-purpose frontier model, separating AI research from real robot hardware development so each can move at its own pace.
Nonprofit research institute Unemployment Labs, founded by a former OpenAI researcher, said it found the weight-initialization fingerprint of Nvidia's export-controlled Blackwell GPUs in six open-weight models released since April 2026 by four major Chinese AI labs. The analysis suggests the models may have been trained on Blackwell chips, and the debate over the effectiveness of AI chip export controls is set to grow.