In case you have not heard, GenAI is outdated information. Now, it is all about agentic AI. A minimum of, that actually appears to be the theme primarily based on the newest bulletins from the main tech trade distributors. All of them are centered on driving the story of extra autonomous actions enabled by AI.
That stated, there’s nonetheless an incredible quantity of exercise and development occurring within the “conventional” period of generative AI – significantly round integrating the know-how into companies and their inside IT operations.
To its credit score, HPE was among the many first to actively talk about and exhibit the potential for operating GenAI-powered purposes inside the confines of a company knowledge middle or non-public cloud. At final yr’s Uncover occasion, the corporate unveiled Non-public Cloud AI, providing organizations the power to construct options on their very own GPU-equipped servers.
At this yr’s present, HPE confirmed real-world progress in these purposes, and took the following step towards enabling AI-powered brokers as a part of these options. Constructed on a framework the corporate calls GreenLake Intelligence, HPE launched a collection of multi-part brokers that may automate a variety of IT operations, together with storage, networking, configuration, observability, and extra.
In actual fact, the corporate delivered one of the vital compelling and complete tales round agent-based IT improvement I’ve seen from any vendor. HPE emphasised that these techniques are nonetheless designed to maintain people within the loop, a key consideration given how impactful autonomous brokers may grow to be.
Constructing on the corporate’s long-established GreenLake non-public cloud platform, GreenLake Intelligence leverages each LLMs and conventional machine studying fashions skilled particularly to deal with IT-related points utilizing many years of information HPE has collected.
A brand new chatbot-style GreenLake Copilot interface sits on high of those fashions, permitting IT professionals to ask questions, troubleshoot points, and discover options in plain English. From FinOps and workload optimization to sustainability metrics and community troubleshooting, the purpose is to simplify the complexity IT groups face every single day.
All GreenLake Intelligence choices start with the idea of a multi-agent orchestrator, which manages a wide range of extra task- or content-specific sub-agents able to figuring out points or strolling by way of processes. These instruments can present step-by-step steering on resolving points, performing duties, and even do them on their very own in an autonomous trend (once more, with human affirmation if desired).
A powerful instance is HPE’s Aruba Central community administration utility, which makes use of its personal model of the GreenLake Copilot interface. It may well shortly detect and assist resolve advanced issues utilizing AI and ML mannequin knowledge accessed by way of what HPE calls an agentic mesh. Like different GreenLake Intelligence companies, it may well additionally generate visible dashboards from incoming log knowledge, making it simpler to identify potential points in actual time.
One other key a part of HPE’s up to date portfolio is its new CloudOps software program suite, which mixes an enhanced model of the OpsRamp observability platform, Morpheus virtualization and cloud administration instruments, and Zerto knowledge administration and safety software program.
OpsRamp, specifically, now options expanded GenAI-powered capabilities designed to watch, handle, and repair numerous parts of a company’s IT infrastructure, together with compute, storage, and networking.
After all, determining precisely what instruments and infrastructure to undertake might be difficult for a lot of organizations. To assist, HPE additionally launched an upgraded CloudPhysics Plus evaluation software to assist corporations in planning their migration to extra trendy hybrid and AI-powered operations.
Past software program and companies, HPE additionally debuted a sequence of latest {hardware} choices at Uncover, together with additions to its ProLiant Gen12 line utilizing AMD’s newest Epyc CPUs. That is nice information for corporations in search of better alternative in CPU suppliers.
Constructing on final yr’s launch of Nvidia GPU-equipped servers for Non-public Cloud AI, HPE additionally launched the second technology of its AI factories, together with techniques that includes Nvidia’s new Blackwell RTX Professional 6000 GPUs – designed particularly for enterprise AI use circumstances. Notably, the corporate confirmed that these new techniques are backward- and forward-compatible, that means they will work alongside each first-generation fashions and future releases. HPE additionally introduced that its OpsRamp software program is now licensed to be used with any Nvidia AI Manufacturing facility-based compute system, together with these made by rivals.
On the storage facet, HPE unveiled the Alletra Storage MP X10000 system, designed particularly for AI workloads. The all-flash, software-defined system integrates with GreenLake platform instruments to handle the massive knowledge units required for AI, and might be optimized by way of GreenLake Intelligence.
One particularly attention-grabbing characteristic of the MP X10000 is its built-in help for MCP (Media Context Protocol) servers, which permit it to speak throughout the complete IT stack – together with servers, networks, observability instruments, storage parts, and extra. MCP is quickly gaining adoption as a typical for enabling connections between LLMs and digital agent platforms, so it is encouraging to see HPE proactively embracing it.
Taken as an entire, the Uncover bulletins present HPE is concentrated squarely on the long run. On the similar time, the corporate balanced that imaginative and prescient with a practical message, highlighting the effectivity and efficiency features organizations can notice by upgrading ageing infrastructure. That is probably not as flashy, nevertheless it displays the fact of at the moment’s blended IT environments.
Trying forward, the agentic imaginative and prescient for IT that HPE offered at Uncover is undeniably compelling, nevertheless it’s necessary to do not forget that we’re nonetheless very a lot within the early phases of AgenticOps. For many organizations it is doubtless going to be a multi-year course of to achieve that imaginative and prescient, nevertheless it’s nice for corporations to have a roadmap in direction of the place they should go.
Bob O’Donnell is the founder and chief analyst of TECHnalysis Analysis, LLC a know-how consulting agency that gives strategic consulting and market analysis companies to the know-how trade {and professional} monetary group. You’ll be able to observe him on X @bobodtech
Picture credit score: Philip Oroni
