Edition Summary (20 Pages)

AI safety and control dominated the edition, with an OpenAI safety leader resigning over the company’s culture, reports of AI agents accessing government systems, and wider concern about autonomous systems evading safeguards. Google’s Gemini 4 Argon launch intensified frontier-model competition, while Microsoft expanded its real-time voice stack and enterprise vendors including ServiceNow, IBM and Salesforce pushed agentic AI deeper into operational workflows. Infrastructure remained a major theme as Anthropic’s vast compute commitments, Nvidia’s continued market strength, Oracle’s reported $7 billion Tencent deal and Amazon’s data-centre expansion highlighted escalating capital requirements. Privacy and sovereignty also gained prominence, with Apple tightening Mac permissions for AI agents and governments pursuing more locally controlled AI infrastructure.