putilov_denis - stock.adobe.com

HPE offerings dive into AI features and new bundle packages

The new HPE Cloud Ops Software suite bundles pieces of the vendor's private cloud data centering offerings, and HPE expands AI capabilities in software and hardware.

AI agents are now swimming throughout Hewlett Packard Enterprises' GreenLake cloud and its associated software, bringing copilot interactions and agentic AI capabilities to the vendor's entire catalog of capabilities.

GreenLake Intelligence, first shown by HPE at its Discover conference Tuesday, is a new collection of agentic AI assistants for a variety of common IT needs, including storage, networking, observability and security. These agents enable customers to interact and optimize their own HPE environments by using conversational interfaces now common for enterprise AI.

The vendor expects to launch a beta of the service by the third quarter of calendar year 2025.

These agents show how enterprise use of AI is continuing to mature and find specific use cases, said Matt Kimball, vice president and principal analyst at Moore Insights & Strategy.

The agentic AI push by many vendors may be coming too fast for enterprise adoption and production workloads, he said. Those customers could soon have a clearer vision of what exactly they'll need for the future, following AI pilots and experiments today.

"I don't think enterprise IT is there yet, for the most part," Kimball said. "[But] enterprise IT is starting to form a solid and real vision for what activating AI across the organization looks like, and it's agentic."

HPE also plans to release HPE CloudOps Software this year, a bundle package of its data center software built from recent acquisitions, according to the vendor.

Agentic offerings

GreenLake Intelligence is described by HPE as a collection of AI agents designed to operate across HPE's stack and multi-cloud hybrid environments. The automation technology of Morpheus enables these agents to view a customer's stack for optimizations, issues and general upkeep, said Varma Kunaparaju, senior vice president and general manager of OpsRamp Software and cloud platforms at HPE.

"My view of this [offering] is not about giving more dashboards and more reports," he said during a media briefing. "It's about embedding intelligence directly into the operational stack so that IT teams can act with speed and confidence."

Kunaparaju highlighted specific use cases, including infrastructure provisioning through a chatbot interface and highlighting issues through a cloud command console.

"This is not a black box," he said. "This is accountable, expandable and explainable reasoning that teams can trust."

Although agentic AI continues to expand in capabilities and operations, GreenLake customers will likely demand even further automation and offloading of management, said Mike Matchett, founder and principal analyst at Small World Big Data.

GreenLake's original pitch as a private cloud with HPE managed services attracted customers who wanted to offload those management and operations concerns onto a third party, he said. Agentic AI brings a new layer of management concerns that HPE customers might already be paying to avoid.

If you could automate that [process] already, why haven't you and why do you need an agent to see that?
Mike MatchettFounder and Principal Analyst, Small World Big Data

"What people would be buying [GreenLake for] is to not do system management anymore," Matchett said. "If you could automate that [process] already, why haven't you, and why do you need an agent to see that?"

Keeping a human in the loop will remain vital even with a managed service, until agentic AI accountability and responsibility increase, Kimball said.

"I don't think enterprise IT orgs are ready for full autonomy yet," he said. "Human in the loop has become a very common phrase, and critical in the agentic journey."

New AI capabilities aren't limited to HPE's GreenLake software.

The HPE Alletra MP Storage X10000, the vendor's latest storage hardware, will support the Model Context Protocol (MCP) to connect storage into a variety of data services for AI. The feature is expected for the second half of calendar year 2025.

MCP, released by AI vendor Anthropic last year, is a framework for orchestrating and integrating AI agents. The new support should help speed up AI implementation and adoption, according to Kimball.

"Embedding MCP into a universal storage platform should deliver considerable performance gains," he said. "And in a disaggregated architecture, you get this considerably better performance with virtually unlimited scale."

Data center bundle

HPE will combine its three recent acquisitions into a single data center offering.

The new HPE CloudOps Software suite combines the OpsRamp visibility platform, virtualization capabilities with HPE Morpheus Enterprise Software and data backup with migration capabilities through HPE Zerto.

Customers can still purchase these offerings separately or with a new pay-as-you-go model combining all three.

Even though virtualization is considered a stable, if aging, data center technology, the pricing changes to VMware have continued to rattle the industry and sent customers looking for alternatives, Matchett said.

"People are feeling locked into VMware by Broadcom," he said. "The services may be getting sleeker and narrower, but it's legacy. I don't think [enterprises] want to retrain VMware admins into Morpheus admins."

The HPE CloudOps Software will be available in the fourth quarter of calendar year 2025, according to HPE.

Tim McCarthy is a news writer for Informa TechTarget covering cloud and data storage.

Dig Deeper on Data center hardware and strategy