Edition Summary (20 Pages)

AI safety and governance dominated the edition, with OpenAI, Anthropic and Google DeepMind discussing a joint standards body and independent model assessments, while Anthropic called for slower frontier development and Nvidia’s Jensen Huang argued against new regulation. Microsoft separately published a draft Humanist AI code emphasising human control and shutdown authority. Product activity remained strong: Google launched Gemini 3.8 Live and Extended Thinking, Apple began the gated Siri AI beta rollout with iOS 27, and Meta introduced paid Meta One subscriptions. At Dreamforce, Salesforce unveiled AIforce and its Nvidia-built Koa CRM reasoning model while expanding ties with Google Cloud and AWS. Capital and infrastructure pressures also featured prominently, including OpenAI’s proposed funding round, Mistral’s €3 billion raise, Oracle’s AI spending and sovereign AI investment.