Microsoft

VentureBeat Pulse Research finds 54% of enterprises have experienced AI agent security incidents or near misses, but only 32% assign unique identities per agent and 30% sandbox high-risk agents.

Microsoft is training sales staff to pitch its AI models as more efficient and cost-effective than OpenAI's and Anthropic's.

Microsoft CEO Nadella accuses OpenAI and Anthropic of a 'reverse information paradox': training on public data under fair use, but banning distillation of their models while learning from customer interactions.

OpenAI says GPT 5.6 will serve as the preferred model for Microsoft Copilot 365.

Microsoft uses AI to identify issues earlier, increasing fixes in each Patch Tuesday update.

Microsoft is replacing OpenAI and Anthropic models in products like Excel and Outlook with its own MAI models to reduce costs, with tens of thousands of weekly queries already using the new models.

To cut AI costs, Microsoft has started replacing OpenAI and Anthropic models with its own in Excel, Outlook, and other products.

Microsoft cut about 4,800 jobs (2.1% of global workforce) on Monday, impacting Xbox and commercial sales.

Microsoft published an official directory of MCP server implementations on GitHub for AI-driven data access and tool integration.

Microsoft open-sourced the 'MCP for Beginners' course on GitHub, teaching Model Context Protocol fundamentals and practical applications.

Microsoft creates a new unit with $2.5 billion, deploying 6,000 engineers on-site to integrate AI into core business processes and ensure measurable ROI.

Anthropic redeployed Claude Fable 5 on July 1 after US export controls were lifted, and added a safety classifier.