Edition Summary (20 Pages)

Google expanded access to its Gemini 4 Argon frontier model, while Meta’s Muse Wonder releases intensified competition in advanced reasoning and agentic AI. Anthropic warned that smaller models could soon self-replicate across connected devices, while an alleged internal security breach at OpenAI and industry calls for stronger regulation heightened concerns about AI safety and oversight. Infrastructure economics remained under scrutiny as Microsoft unveiled its Maia 300 accelerator strategy, Nvidia announced European investments, and financing and power constraints raised questions about Oracle’s data centre expansion. Enterprise AI adoption advanced through Workday’s Anthropic partnership, Salesforce’s acquisition agreement with Loops, and expanding autonomous workflow capabilities. Sovereign AI initiatives also gained attention as governments and enterprises pursued greater control over infrastructure, models and sensitive data.