Microsoft announced Azure Maia 300, an inference-first accelerator system intended for long-context reasoning and sustained multi-step agent tasks. Microsoft claims 40% more tokens per megawatt than leading GPU systems and up to 7× Maia 200 rack throughput; Maia 300 deployments are planned for 2027, while Maia 200 is already serving Copilot production workloads. These are vendor projections, not independent benchmarks. Source: https://blogs.microsoft.com/blog/2026/10/12/microsoft-to-expand-azure-maia-system-to-multi-gigawatt-deployments/
Samsara launched Agent Studio, a no-code agent builder for physical operations, where agents can use live operational context and act through voice/frontline and dashboard workflows under a shared permission model. Samsara reports platform agent adoption has grown 7× since June; that traction figure is company-reported. Source: https://www.samsara.com/company/news/press-releases/agent-studio
DigitalOcean introduced Agent Data Services (private preview), a managed multimodal context store that ingests/indexes changing business data and captures agent execution traces, linking live context to its observability/evaluation and Agent Learning services. It targets the fragmented data-plus-feedback-loop layer; availability is limited preview, and terms exclude standard compliance certifications/SLA. Source: https://www.digitalocean.com/blog/introducing-agent-data-services