AMD’s Helion rack blueprint separates AI agent execution from model inference—or combines both—giving platform teams a new capacity-planning model for production agents.

Meta has released Muse Wonder through its Model API and Muse Code, bringing million-token context and long-horizon agent capabilities while leaving key production questions around data modes, evaluation and safety.

Microsoft will make Maia its primary inference platform and begin multi-gigawatt Maia 300 deployments in 2027. Here is what the custom silicon shift means for developers and platform teams.

Anthropic’s free OSS Scanner will send critical open-source maintainers recurring AI-generated vulnerability reports, shifting the bottleneck from discovery to verification and safe patching.