I just returned from HPE Discover 2025 in Las Vegas. This year’s event focused on practical ways to simplify infrastructure, manage costs, and prepare for AI workloads — rather than abstract future promises.
Below is a summary of the key announcements and what they mean from a data center and modernization perspective.
HPE introduced GreenLake Intelligence, a set of AI-driven operational tools designed to automate routine tasks, optimize resource allocation, and assist with issue resolution.
Instead of relying on manual tuning and monitoring, GreenLake Intelligence uses domain-specific AI models trained on historical telemetry data. While it won’t replace engineers, it can help reduce repetitive maintenance work and improve consistency.
One of the most concrete announcements was HPE’s focus on lowering virtualization costs.
VM Essentials, together with Private Cloud Business Edition (PCBE), uses a KVM-based hypervisor and Morpheus software to provide a lower-cost alternative to VMware. HPE cited potential savings of up to 90% on per-socket licensing, depending on workload and scale.
This approach gives organizations a clear path to reduce hypervisor licensing costs without sacrificing hybrid or multi-cloud flexibility. For organizations facing rising VMware renewal costs, this presents a compelling alternative to explore.
HPE addressed the backup and DR options for VM Essentials workloads directly at Discover.
Today, VM Essentials supports guest-based (agent) backups. While reliable, this method requires more operational overhead and individual VM-level management.
Commvault is currently the only vendor providing true native VM-level backup support for VM Essentials today. If agentless, image-based backups are needed immediately, Commvault is the primary option.
In addition, Cohesity, Rubrik, and Veeam are actively developing native support for VM-level backups on VM Essentials. These vendors are working to bring full hypervisor-level integration to the platform in upcoming releases, which will provide organizations with more choices and reduce reliance on guest agents.
Zerto replication is not supported yet for VM Essentials but is officially on the roadmap. Organizations planning to adopt VM Essentials should design their data protection strategies around the current options today while preparing to leverage these additional capabilities as they become available.
HPE introduced Private Cloud AI, developed with NVIDIA, as a pre-integrated, production-ready platform for AI workloads. It includes support for air-gapped deployments, multi-tenancy, and modular scaling.
Alletra Storage MP was presented as a data platform designed to support high-performance AI pipelines, offering real-time metadata processing and optimized integration with model training workflows.
These solutions are designed for organizations with defined AI production workloads rather than those still in early exploration.
The new ProLiant Gen12 servers allow significant consolidation of previous hardware generations, with HPE citing examples of replacing up to seven Gen10 servers or 26 Gen8 servers with a single Gen12 system.
Benefits include lower power and cooling requirements — potentially reducing operational costs by up to 65%, depending on deployment. Direct liquid cooling and heat reuse support are also included, which can be important for dense GPU workloads or energy-focused data centers.
Networking updates centered around Aruba Central’s AI capabilities for predictive monitoring and rapid issue resolution.
Zero Trust security controls are now more tightly integrated, extending policies to edge devices and remote environments.
HPE also discussed its planned acquisition of Juniper Networks, aimed at strengthening end-to-end networking capabilities from data center to campus and edge use cases.
At Prescriptive, we see these updates as providing practical options for organizations looking to reduce costs, simplify operations, and prepare for future AI workloads — without requiring immediate, large-scale transformations.
VM Essentials and PCBE present a realistic opportunity to address virtualization costs. Backup and DR planning remains critical: Commvault offers immediate VM-level support, while Cohesity, Rubrik, and Veeam are working to expand compatibility. Zerto will add further capabilities when support becomes available.
The new AI-focused solutions and storage updates are best suited for organizations moving AI workloads into production, rather than just exploring proofs of concept.
ProLiant Gen12 and the networking enhancements provide near-term improvements in efficiency and reliability, especially for organizations planning hardware refreshes in the next one to two years.
HPE Discover 2025 focused on clear, actionable updates for modernizing infrastructure and controlling operational costs while preparing for AI and next-generation workloads.
If you’re evaluating options to reduce hypervisor expenses, improve operational efficiency, or prepare for AI projects, these new solutions provide concrete and realistic directions.