THE DAILY BRIEF

AMD is investing up to $5 billion in Anthropic, with Anthropic committing to deploy up to 2 gigawatts of MI450 GPUs for training and running Claude models. This deal positions AMD as a serious challenger to Nvidia in the AI chip supplier space after similar major agreements with Meta and OpenAI.

► AMD's growing dominance in AI infrastructure deals signals Nvidia's CUDA moat is eroding, forcing a redistribution of semiconductor power and reshaping AI capex strategies across the industry.

AI lab Midjourney has acquired the popular astrology app Co-Star and is building its first standalone image-generation application. The moves represent Midjourney's push into new business areas following an ambitious expansion announcement last month.

► Midjourney's shift from a Discord-based service to acquiring consumer apps and building standalone products shows how AI image-gen startups are maturing into diversified platform companies.

Anthropic is rolling out Claude Opus 5, a new AI model designed to handle workplace tasks affordably while delivering near-parity performance with competing high-end models. The cost reduction addresses growing customer price sensitivity as competition intensifies globally.

► The race to commoditize frontier AI performance through cost optimization is accelerating, forcing all model providers to compete on unit economics rather than raw capability alone.

Microsoft has observed a sharp increase in attacks using ACR Stealer malware targeting enterprise customers to steal browser-stored passwords, authentication tokens, and sensitive documents. The threat represents a significant shift in attack vectors against large organizations.

► Enterprise security teams must now assume browser credential storage is a critical attack surface, forcing reconsideration of how authentication flows are architected in modern applications.