HPE Embraces Agentic AI and Signals a Power Shift in Enterprise Ops

I was on the ground this week at HPE’s annual Discover conference in Las Vegas. At the conference HPE made a slew of announcements that demonstrated a significant shift in both their direction and ultimate value to enterprise customers. In large part, HPE is capitalizing on early bets while making a marked shift in their software and AI strategy. At the same time, they left significant gaps that need to be addressed.

There is a lot to unpack…let’s break down what’s news and what it means…

GreenLake leans into agentic AI with GreenLake Intelligence

The first and probably most significant overarching announcement is GreenLake Intelligence. GreenLake is HPE’s solution that reimagines how on-premises infrastructure is both operated and consumed.

Early on, many saw GreenLake largely as an ‘as a service’ model for on-premises infrastructure. HPE CEO Antonio Neri even mentioned several years ago that every product in the HPE portfolio would be made ‘as a service’. Having accomplished that goal, HPE needed to take the next step and capitalize on the data and insights that the GreenLake infrastructure was generating.

At the same time, AI was coming fast and furious to drive both HPE and customers to derive insights from traditional infrastructure. The convergence of GreenLake and AI created a perfect storm of opportunity for HPE. The data and insights gleaned from GreenLake would feed an engine of automation and optimization for customers.

Fast forward to today…

HPE’s new GreenLake Intelligence is an agentic-based AI framework operating on top of GreenLake. Agents within GreenLake provide insights to the underlying infrastructure stack through a series of new copilots.

New HPE CloudOps software suite

In the vein of consolidation, HPE also announced the new HPE CloudOps software suite which integrates a combination of OpsRamp, Morpheus and Zerto. While customers can still purchase the individual products, customers are likely to shift toward integrated approaches like the CloudOps software suite to both simplify and gain greater insights across products. The power in HPE’s GreenLake approach is in its platform…and now GreenLake Intelligence. The CloudOps software suite can provide key insights to drive decisions…and potentially down the road automation for GreenLake.

Agentic and automation come to HPE OpsRamp

OpsRamp is HPE’s operational control center that works across the server, storage, network and software pillars. New to OpsRamp is a copilot built on an agentic framework that provides easier access to insights and automation. As changes in infrastructure dynamics are getting more nuanced and timing is critical, automation becomes key. The new features bring an agentic approach to insights and automation within OpsRamp and integration with GreenLake Intelligence.

HPE Aruba Networking Central

HPE’s networking portfolio, Aruba, is moving to leverage AI for network operations and getting its own copilot. Leveraging AI with networking is a powerful combination as network architectures, changes in performance and security dynamics are constantly in flux. Competitors such as Cisco and Juniper are already leveraging AI in their networking suite and HPE is following suit.

Speaking of Juniper, there was quite a bit of sideline hallway conversation questioning if HPE’s acquisition of Juniper is going to happen. Officially, HPE cannot comment on the pending acquisition. While a Juniper acquisition would be a boon to accelerate HPE’s Aruba portfolio and technology with more mature AI technology from Juniper, the further HPE advances Aruba’s use of AI on their own, the more questions the deal raises.

UPDATE 6/28/2025: HPE announced an agreement with the DOJ that “clears the way for the transaction to close.”

Also, along the lines of integration, Aruba has long kept their own branding and operated somewhat separately from the core HPE portfolio. That made sense as Aruba had a very strong brand, and networking was largely separate from the other core pillars to HPE’s portfolio. However, customers are increasingly looking for a more integrated approach to infrastructure operations.

There was conversation some time back that Aruba Networking Central might provide wider operational oversight beyond networking due to the sophistication in the platform. While that never happened, there needs to be a plan to more fully integrate Aruba into the HPE portfolio. Should the Juniper acquisition proceed, it will only provide further pressure to integrate. 

HPE Alletra Storage MP X10000

One of the more confusing announcements was that HPE’s Alletra Storage MP X10000 would be providing MCP (model context protocol) servers natively built into the platform. MCP provides the ability for agents in an agentic framework to call on other services (like storage, systems, data). So, why was this confusing since calling storage using MCP makes sense? While it is good for that one purpose, it misses the bigger picture as to where MCP could provide value across the GreenLake architecture. Maybe that is part of the plan, but it did not come out that way.

MCP has the potential to provide significant value across an infrastructure portfolio. Yet, the announcement was focused on storage and a specific model of storage system. It would have been more interesting to see MCP coming from a GreenLake Intelligence level and spanning out to services and then to specific technology.

HPE CloudPhysics Plus

CloudPhysics is HPE’s workload assessment tool and has been expanded to cover edge to cloud workloads across several hypervisors. Interesting to note that the tool is only available when conducting an assessment by either channel partners or HPE sales.

This kind of tool serves as a foundational element to provide insights that, when paired with automation, could refactor workloads for better performance and optimization over time. HPE didn’t allude to this capability, but putting some of the capabilities together offers an interesting opportunity.

FinOps and Sustainability

FinOps (and TBM) are both gaining interest among enterprise customers to more clearly understand the value they are getting from their technology investments. Clearly understanding the full cloud cost picture is complicated. HPE announced that they are expanding GreenLake to provide greater insights for FinOps including FinOps Open Cost and Usage Specification (FOCUS) exports.

FinOps, pared with carbon footprint provides greater insights to metrics around spend, value and sustainability. 

Overall conversation themes

Several conversations popped up during the conference that warrant discussion.

1. Quantum and HPC: HPE talked quite a bit about their quantum and HPC chops. HPE has strong experience in the HPC world through their Cray acquisition some years ago. With AI, the HPC and classic computing worlds are starting to converge. At the same time, quantum is getting closer as a potential suitor for specific workloads.

• Hybrid IT (Edge to Cloud): HPE has focused on providing an edge-to-cloud solution from the start of GreenLake. Today, that conversation is flying higher through new software capabilities alongside increased demand from enterprises needing a hybrid approach to IT services.

• AI Factory: It seems every platform vendor is talking about building an ‘AI factory’. The problem is that it means different things to different companies. While enterprise customers are building AI factories, it means something different from vendors are largely talking about. 

What comes next?

HPE’s sales and partner programs are a strength for the company…and customers. As HPE moves beyond foundational infrastructure toward software-based insights, customers can lean in on their automation strategies using agentic AI. HPE will need to ensure that both sellers and partners understand this shift and can deliver value to customers.

Another challenge is that HPE’s typical infrastructure company contact is not the one driving agentic AI within the enterprise. HPE will need to expand who they build relationships with and the messaging they use. Rather than a bottoms-up approach, HPE will need to adopt more of a top-down approach.

At the same time, HPE must be cautious not to cannibalize or alienate their base. HPE has several cards they could play to address this and will be interesting to watch where they go especially considering strong competition from Dell, Lenovo and the hypervisor vendors such as AWS, Google and Microsoft.

What’s missing?

HPE made marked shifts in their strategy and communicated it well. At the same time, it was fairly limited in where it is headed and what customers can expect next. That’s less about roadmap and more about setting the tone and cadence for where HPE is headed.

Why do customers care about this? Customers are moving much more quickly than ever before and need to ensure that the partners that they are working with align with their multiyear journey.

Also missing from wide discussion were cybersecurity, data, insights and alignment with higher-level functions facing CIOs today. This beckons the question: Is HPE an infrastructure company that provides software to manage it? Or is HPE becoming a software company that happen to also have infrastructure? Right now, it appears as the former with the potential to be the latter.

From the CIO perspective

HPE has always been a solid infrastructure player. Between core HP (before HPE), Compaq and Cray, HPE continues to show they have the infrastructure chops to compete among the best providers.

Today, infrastructure is largely heading toward commodity, and the differentiator is in how it is managed…ergo software. It’s a space that infrastructure companies have largely consolidated into infrastructure operations.

That’s good. But today, enterprises are looking for much more. Why? In the end, because customers are driving enterprises to change how they use technology. That, in turn, is causing IT organizations to rethink how they leverage technology resources.

AI is turning technology on its head and providing significant opportunities for both vendors and enterprises alike. While switching costs between infrastructure providers is still fairly high, building software insights and automation that aligns with business outcomes will further differentiate providers.

HPE has done well with GreenLake and the new enhancements announced this week. Now they need to capitalize on them and determine their next move.