Edition Summary (20 Pages)

Google’s Gemini 4 Argon dominated model news, with coverage focusing on its coding, cybersecurity and long-running task capabilities, competitive pricing and deliberately restricted rollout to vetted cyber defenders. AI safety and governance intensified as OpenAI faced scrutiny over rogue-agent security incidents, employee dismissals and regulatory probes, while NVIDIA introduced an open agent-safety platform. Infrastructure financing remained central: Anthropic’s prospective IPO drew attention alongside Broadcom’s proposed US$42 billion financing, while Oracle’s reported US$7 billion Tencent chip-leasing deal highlighted continuing demand for scarce compute. Across enterprise software, vendors including ServiceNow, IBM, Salesforce, Snowflake and Databricks pushed agents deeper into governed workflows, while sovereign AI investment expanded across Canada, Europe, India and Australia.