HPE & Nvidia Launch Blackwell AI Infrastructure Portfolio
Working across servers, private cloud platforms and compute systems, the company is eager to address growing demand for on-premises AI infrastructure among organisations requiring data sovereignty or air-gapped operations.
The company’s announcement at HPE Discover Las Vegas 2025 introduces three distinct solutions targeting service providers, governments and enterprises seeking alternatives to hyperscale cloud deployments.
HPE positions the offering as a direct challenge to cloud providers’ dominance in AI deployment, providing turnkey solutions that reduce setup complexity for data centre operators.
“Generative, agentic and physical AI have the potential to transform global productivity and create lasting societal change, but AI is only as good as the infrastructure and data behind it,” says Antonio Neri, president and CEO at HPE.
HPE Private Cloud AI Integrates Blackwell Architecture
The centrepiece deployment combines Nvidia’s Blackwell accelerated computing with HPE ProLiant Compute Gen12 servers. These servers have achieved top rankings in over 23 AI performance tests according to HPE's internal benchmarking data.
The Private Cloud AI platform includes air-gapped management capabilities for organisations with strict data privacy requirements.
Multi-tenancy features enable resource partitioning across teams within single deployments. HPE has introduced a try-and-buy programme allowing customers to test Private Cloud AI across Equinix’s global network of data centres before purchase commitments. This approach reduces deployment risk for data centre executives evaluating AI infrastructure investments.
The company targets three market segments with distinct requirements. Large-scale solutions combine HPE ProLiant Compute XD systems with Nvidia AI Enterprise software and comprehensive cooling technologies for service providers and model builders.
Likewise, sovereign AI solutions address government and public sector requirements through specialised air-gapped management and data sovereignty capabilities.
“We are entering a new industrial era – one defined by the ability to generate intelligence at scale,” says Jensen Huang, CEO of Nvidia.
Targeting data-intensive AI workloads
The HPE Compute XD690 represents the company’s latest AI-focused hardware addition, supporting eight Nvidia Blackwell Ultra GPUs per system. HPE Performance Cluster Manager provides integrated systems management across large AI environments scaling to thousands of nodes.
This management layer addresses operational complexity concerns among data centre operators deploying multi-node AI clusters.
Storage infrastructure receives attention through the HPE Alletra Storage MP X10000, which will support Model Context Protocol servers to accelerate data pipelines for AI applications. The storage system targets data-intensive AI workloads requiring high-throughput access to training datasets and model repositories.
HPE’s Unleash AI partner ecosystem has grown to include 26 new partners, enabling support for over 75 AI use cases across various industries. These span agentic AI, sovereign AI, smart cities, industrial applications, data governance and cybersecurity solutions.
The partner network approach allows HPE to address vertical-specific requirements without developing industry expertise internally.
Assessing the impact
HPE has also announced a collaboration with Accenture to develop agentic AI solutions for financial services organisations.
The partnership utilises the Accenture AI Refinery platform deployed on HPE Private Cloud AI infrastructure, addressing regulatory compliance requirements specific to financial institutions.
Additionally, HPE Financial Services has introduced a programme offering reduced payments for Private Cloud AI during the first six months of deployment. The programme allows customers to leverage existing technology assets as capital for funding additional AI projects. This financing approach addresses budget constraints among enterprises evaluating AI infrastructure investments.
The payment structure recognises that AI deployments often require proof-of-concept phases before full production deployment.
Data centre executives could use the reduced payment period to demonstrate value to stakeholders before committing to standard pricing terms.
The expanded portfolio positions HPE as a competitor to cloud providers in enterprise AI infrastructure deployment. Rather than migrating workloads to public cloud platforms, organisations can deploy AI capabilities within existing data centre facilities using HPE’s integrated solutions.
Jensen explains: “Together, HPE and NVIDIA are delivering full-stack AI factory infrastructure to drive this transformation, empowering enterprises to harness their data and accelerate innovation with unprecedented speed and precision.”
Explore the latest edition of Data Centre Magazine and be part of the conversation at our global conference series, Tech & AI LIVE and Data Centre LIVE.
Discover all our upcoming events and secure your tickets today.
Data Centre Magazine is a BizClik brand
