Every year, HPE Discover offers a glimpse into where enterprise infrastructure is heading. This year's event in Las Vegas was no different, with AI dominating the agenda.
While generative AI has been the focus for the past couple of years, the conversation has now evolved towards agentic AI - systems capable of planning, reasoning and taking actions autonomously. Supporting these workloads requires far more than powerful GPUs. Organisations need secure infrastructure, well-managed data, resilient storage, automated operations and trusted technology partnerships.
Here's our roundup of the biggest announcements from HPE Discover 2026.
Hardware: Building infrastructure for AI at scale
HPE's hardware announcements centred around creating infrastructure that can support increasingly demanding AI workloads while simplifying hybrid cloud operations.
New ProLiant Compute DL394 Gen12
One of the headline announcements was the ProLiant Compute DL394 Gen12, powered by NVIDIA's new Vera CPU.
Designed for AI training, inference, financial modelling and high-performance data processing, the platform delivers significantly higher memory bandwidth through LPDDR5X memory, enabling AI applications to process large datasets more efficiently.
While many AI announcements focus exclusively on GPUs, memory performance is increasingly becoming the limiting factor for large AI models, making this an important development.
HPE Private Cloud Gen4
HPE also introduced the fourth generation of its Private Cloud platform, built on ProLiant Gen12 servers.
Perhaps the most significant enhancement is the ability to manage both traditional virtual machines and Kubernetes workloads through a single platform. As organisations continue modernising applications at different speeds, unified management helps reduce operational complexity while supporting hybrid environments.
Alletra Storage continues to evolve
Storage remains one of the biggest challenges for enterprise AI.
The Alletra Storage MP X10000 now expands beyond object storage to include native file storage, giving organisations a single platform capable of supporting AI pipelines, analytics, backup and cyber resilience. Increased scalability and RDMA support also improve performance for demanding AI workloads.
Meanwhile, the Alletra Storage MP B10000 gains autonomous storage management capabilities, allowing it to detect, analyse and resolve issues automatically. Combined with increased controller scalability and enhanced data reduction guarantees, the platform continues HPE's push towards self-managing infrastructure.
Expanding the HPE AI Factory
HPE also expanded its AI Factory portfolio with additional NVIDIA Blackwell-based infrastructure, including RTX PRO 6000 Server Edition GPUs, Spectrum-X networking, BlueField-3 DPUs and ConnectX-8 SuperNICs.
The focus here is less about individual hardware components and more about delivering validated, sovereign AI infrastructure that organisations can deploy with confidence.
Software: Making AI operational
Powerful infrastructure alone doesn't make AI successful. Organisations also need the software to manage data, monitor workloads and maintain operational resilience.
That was the focus of HPE's software announcements.
Private Cloud AI becomes more production ready
HPE Private Cloud AI received several updates aimed at moving AI projects from proof of concept into production.
New capabilities include secure local AI agent registration, governed access to frontier models, workload prioritisation and support for larger multi-node inference clusters. These additions reflect the growing importance of governance, security and operational management as organisations deploy AI more widely.
Better data management with HPE Data Fabric software
Data remains the foundation of every successful AI initiative.
HPE announced several enhancements to Data Fabric, including policy-based data movement across hybrid environments, improved metadata management and conversational AI capabilities for interacting with enterprise data.
Support for Model Context Protocol (MCP) and enterprise AI inventory management should also make it easier to understand where AI-ready data exists across the organisation.
GreenLake Intelligence expands observability
One of the more interesting announcements was the continued development of GreenLake Intelligence.
Rather than simply monitoring infrastructure health, GreenLake is evolving towards full AI operations management, providing visibility into AI utilisation, token consumption, operational costs and even agentic root-cause analysis across multi-vendor environments.
As AI becomes business critical, this level of observability will become increasingly important for managing both performance and cost.
Data protection keeps pace with AI
Resilience remains a key consideration as AI workloads become more valuable. HPE announced several updates across its data protection portfolio, including:
• Live VMware workload migration using HPE Zerto
• AI-powered protection enhancements
• Detection of potentially rogue AI agent activity
• Deeper integration between StoreOnce, SimpliVity and HPE Private Cloud
These developments reinforce an important message: organisations can't separate AI strategy from cyber resilience and disaster recovery.
Partnerships: AI is still a team sport
Perhaps the biggest takeaway from Discover 2026 was that no single vendor can deliver enterprise AI alone.
HPE continues to strengthen partnerships across the ecosystem to provide integrated solutions rather than isolated products.
NVIDIA
The partnership with NVIDIA continues to deepen, spanning hardware, AI software and infrastructure.
Support now extends across NVIDIA Vera CPUs, Blackwell GPUs, NeMo, Nemotron, Agent Toolkit, OpenShell, Confidential Computing and high-performance networking technologies.
Together, HPE and NVIDIA are clearly positioning themselves to deliver complete AI platforms rather than individual components.
Veeam
Backup and recovery also received attention through expanded integration with Veeam.
The new capabilities provide agentless, image-based backups for HPE Private Cloud, alongside improved recovery and data mobility across platforms.
Microsoft Defender
Security remains front and centre, with HPE Zerto integrating more closely with Microsoft Defender to improve threat visibility and accelerate recovery during cyber incidents.
ServiceNow
Finally, HPE announced closer integration with ServiceNow, connecting GreenLake Intelligence and OpsRamp Operations Copilot with ServiceNow's autonomous AI capabilities.
The goal is to create a single operational view spanning infrastructure monitoring through to automated service management.
Final thoughts
Many of the announcements at HPE Discover 2026 focused on AI, but the underlying theme was operational maturity.
As organisations move beyond experimentation, the challenges become less about building AI models and more about managing infrastructure, protecting data, controlling costs and maintaining resilience.
For organisations considering AI adoption, the most successful projects won't simply have the fastest hardware. They'll combine modern infrastructure with strong governance, resilient data platforms, effective monitoring and robust recovery capabilities.
Those are all areas where good operational practices matter just as much as technology.
As always, we'll be watching closely as these new capabilities become generally available over the coming months.
Learn more
• HPE introduces ProLiant Compute DL394 Gen12 with NVIDIA Vera CPU: https://www.hpe.com/us/en/newsroom/press-release/2026/05/hpe-introduces-cpu-server-with-nvidia-vera-cpu-purpose-built-for-agentic-ai.html
• HPE delivers unified private clouds and data platforms: https://www.hpe.com/us/en/newsroom/press-release/2026/05/hpe-delivers-unified-private-clouds-and-data-platforms-to-accelerate-enterprise-modernization-and-ai-data-readiness.html
• HPE brings agentic AI into production with NVIDIA: https://www.hpe.com/us/en/newsroom/press-release/2026/06/hpe-brings-agentic-ai-into-production-with-nvidia-delivering-security-governance-scale-and-sovereignty.html
• HPE delivers unified agentic IT operations with GreenLake Intelligence and Morpheus Software: https://www.hpe.com/us/en/newsroom/press-release/2026/06/hpe-delivers-unified-agentic-it-operations-with-greenlake-and-hpe-morpheus-software.html
With over 20 years of experience supplying HPE technology, COOLPSIRiT's expert team is here to help guide you through features, functions, and options available.
If you would like to learn more and understand how to ensure reliable recovery, then contact our team today by emailing hello@coolspirit.co.uk or call 01246 454222.
