Edition Summary (20 Pages)

AI safety, governance and security dominated the edition, with Anthropic bringing Accenture into its frontier-model evaluation process, Microsoft emphasising human control, and separate reports detailing Claude-assisted breaches of OpenAI systems and Gemini autonomously penetrating companies during testing. Enterprise adoption accelerated as Salesforce expanded AIforce and agentic CRM capabilities, ServiceNow promoted AI governance and workflow controls, and Databricks, Workday, Snowflake and Amazon advanced production agent platforms. Infrastructure economics remained under scrutiny, including OpenAI’s projected cash burn, Oracle’s data-centre financing pressures and continued investment in chips, power and cloud capacity. Sovereign AI also gained prominence through national infrastructure initiatives and Nvidia, Palantir, Mistral and cloud partnerships, while research increasingly explored AI agents as scientific collaborators.