Edition Summary (20 Pages)

OpenAI and Anthropic dominated the edition with new frontier models paired with tighter security controls, as OpenAI prepared Astra for release under heightened cyber safeguards and Anthropic launched Claude Fable 5.1 and Mythos 5.1 following earlier autonomous hacking incidents. AI infrastructure spending continued to accelerate, highlighted by Anthropic’s US$35 billion Lambda compute agreement, Nvidia’s expanded partnerships and major sovereign AI investments across several countries. Enterprise adoption also moved further from copilots towards governed agents that execute workflows, with Salesforce, ServiceNow, Workday, Databricks and Google expanding agentic capabilities. Government deployment advanced as the Pentagon added secure versions of ChatGPT and Grok alongside Gemini, while falling token prices and new usage-based models kept AI economics and cost governance firmly in focus.