REPORTED DEVELOPMENT
HPE announced Private Cloud AI capabilities developed with NVIDIA: governed APIs, versioned MCP servers, role-, tenant- and project-based permissions, machine identities and audit records.
Availability has three stages. Heterogeneous H200/RTX PRO GPU clusters and KAI Scheduler are available now; new HPE AI Essentials capabilities are due at the end of 2026; self-service agent registration, API/MCP discovery and credential management are planned for 2027. HPE also cites planned industrial edge agents with Schneider Electric, rather than an independently verified implementation.
WHY IT MATTERS TO THE MARKET
Announced architecture and available functionality must be distinguished, alongside permissions, environment isolation, operational records and infrastructure costs. This is HPE's implementation roadmap, distinct from NVIDIA's reference-architecture announcement.