iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
HPE and NVIDIA are expanding their existing enterprise AI collaboration, not forming a new partnership. Their latest announced step, on September 28, 2026, is to integrate NVIDIA OpenShell into HPE Private Cloud AI and bring NVIDIA BlueField-4 support to more HPE systems. OpenShell integration is planned for Q4 2026; BlueField-4 timing depends on product lead times.
What HPE and NVIDIA announced
The collaboration is branded NVIDIA AI Computing by HPE. First announced in June 2024, it combines co-developed systems with joint go-to-market integrations. Its scope has since grown from turnkey private AI infrastructure to large AI factories, storage, servers, services, and supercomputing.
The September 2026 update focuses on safeguards for AI agents. HPE says it is integrating NVIDIA OpenShell, a secure runtime, into HPE Private Cloud AI. HPE also plans to support NVIDIA BlueField-4 across its server and AI factory portfolio, using NVIDIA Sentry and DOCA for infrastructure-level monitoring and policy enforcement. The two pieces address different layers: runtime controls govern what agents can do and access, while infrastructure controls monitor and enforce policies at the systems level. These are described capabilities, not an independent security evaluation.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Clear out junk files and repair common Windows errors3Fix the driver behind crashes, sound loss and screen glitchesHPE also says NVIDIA Confidential Computing is planned for HPE AI Factory solutions through HPE Services in Q4 2026. The announcement describes cryptographic attestation and encryption as part of a chain of trust. HPE says the capabilities can help address requirements including CMMC, NIST 800 series, STIG, and FIPS, but applicable requirements depend on configuration and deployment. HPE’s September 28, 2026 announcement.
#1 Best Overall
- PLEASE NOTE: Exporting an NVIDIA RTX Pro 6000 GPU outside the US requires strict adherence to the U.S. Export Administration Regulations (EAR) and issuance of an export license from the Bureau of Industry and Security (BIS). Compliance and Know Your Customer (KYC) screening may be required as a condition of order acceptance. [NVIDIA Blackwell Streaming Multiprocessor] The new SM features increased processing throughput, and new neural shaders that integrate neural networks inside of programmable shaders | DLSS 4: Multi Frame Generation ensures ultra-smooth frame pacing for lifelike simulations.
- [Double-Flow-Through Design] The RTX PRO 6000 Blackwell features a double-flow-through cooling design, optimizing efficiency and airflow to sustain peak performance under 600W power loads. | [5th Gen Tensor Cores] Deliver up to 3X the performance of the previous generation and support for FP4 precision for faster AI model processing times with reduced memory usage, enabling local fine-tuning of LLMs and generative AI | [4th Gen Ray Tracing Cores] Double the ray-triangle intersection rate of the previous generation to create photoreal, physically accurate scenes and immersive 3D designs with RTX Mega Geometry, which enables up to 100X more ray-traced triangles.
- [PCIe Gen 5] Support for PCIe Gen 5 provides double the bandwidth of PCIe Gen 4, improving data-transfer speeds from CPU memory and unlocking faster performance for data-intensive tasks like AI, data science, and 3D modeling. | [GDDR7 Memory] With 96 GB of GPU memory and 1.8 TB ps bandwidth, it can tackle massive 3D and AI projects, fine-tune AI models locally, explore large-scale VR environments, and drive larger multi-app workflows.
- [DisplayPort 2.1] Achieve unparalleled visual clarity and performance, driving high resolution displays at up to 8K at 240 Hz and 16K at 60 Hz. Increased bandwidth enables seamless multi-monitor setups while HDR and higher color depth support ensures superior color accuracy for precision work, such as video editing, 3D design, and live broadcasting.
- [Universal MIG] Divide a single RTX PRO 6000 Blackwell into multiple isolated instances, each with dedicated resources, allowing for concurrent execution of multiple workloads, optimized GPU utilization, and secure isolation of different applications or users. [WARRANTY] 3 YR Manufacturer's Warranty. Bulk OEM Packaging. Retail Packaging is NOT included.
What is included in NVIDIA AI Computing by HPE?
The name covers a portfolio rather than one machine. The main options differ by scale and workload:
| Offering | What it is for | What to compare |
|---|---|---|
| HPE Private Cloud AI | Turnkey private enterprise AI factory, with configurations for workloads such as inference, fine-tuning, and retrieval-augmented generation. | GPU configuration and scale, private or air-gapped controls, feature availability, and operational model. |
| HPE AI Factory at-scale and Sovereign AI Factory | Full-stack AI infrastructure for service providers, large enterprises, and sovereign deployments. | Scale, tenancy, sovereignty and control requirements, networking, cooling, and deployment timing. |
| HPE Cray Supercomputing GX5000 family | Systems for AI alongside high-performance computing and scientific workloads. | CPU and GPU architecture, interconnect, workload mix, density, and power and cooling needs. |
| HPE ProLiant DL394 Gen12 | Announced Private Cloud AI server featuring an NVIDIA Vera CPU. | Exact configuration, intended role, and delivery timing. |
The portfolio’s initial June 2024 announcement centered on HPE Private Cloud AI, combining NVIDIA AI compute, networking, and software with HPE compute, storage, and GreenLake cloud capabilities. HPE described four right-sized configurations at launch. HPE’s June 18, 2024 announcement.
How the new agent controls are intended to work
OpenShell: runtime governance
HPE says OpenShell is an open-source secure runtime intended to connect agent execution with enterprise identity, policies, approvals, observability, and audit. The design is described as supporting different models, harnesses, and agents across cloud, hybrid, on-premises, and air-gapped infrastructure. HPE Private Cloud AI integration is planned for Q4 2026; that is an announced schedule, not confirmation that the integration is already available.
Sentry and BlueField-4: infrastructure safeguards
NVIDIA Sentry is described as an independent watchdog that monitors agent activity and applies policies out of band. It runs on NVIDIA BlueField-4 DPUs and uses NVIDIA DOCA. HPE plans BlueField-4 support across servers, AI rack-scale systems, Private Cloud AI, AI Factory at-scale, and Sovereign AI Factory, with timing dependent on product lead times. The announcement does not establish that all these integrations are shipping now.
Rank #2
- VD8465 Japanese Authorized Distributor Product
- The speed of FP32 calculation is twice as fast as previous generations, which greatly improves the complex 3D processing and graphics simulation workflow
- Up to 2X the throughput compared to previous generations and significantly faster workloads such as video content rendering, architectural design assessments, and virtual prototypes of product design
- Achieve more than twice the previous generation AI performance improvement, support faster FP8 precision data and accelerate the execution of mixed flotation decimal and whole numbers
- It has a large capacity of memory necessary for working with a vast array of data sets and workloads such as rendering, data science, and simulation
HPE’s September 28, 2026 post on secure agentic AI characterizes production readiness in terms of being able to show where agents run, what they can reach, and what they did. The described controls are intended to support that visibility and enforcement; they do not, by themselves, establish compliance or security for every deployment.
How much can the systems scale?
HPE’s announcements give several specific capacity figures, but they refer to different products and should not be combined into one system specification.
- Private Cloud AI network expansion: HPE’s March 2026 announcement described network expansion racks scaling to 128 GPUs and scheduled availability for July 2026. The schedule alone does not confirm current shipping status. HPE’s March 16, 2026 announcement.
- Private Cloud AI multi-node inference: HPE’s June 2026 announcement described inference capability for up to 256 GPUs. This is distinct from the 128-GPU network expansion rack figure. HPE’s June 16, 2026 announcement.
- GX5000 compute blades: HPE said a GX240 blade can feature up to 16 NVIDIA Vera CPUs. A rack can scale to 40 blades, or 640 Vera CPUs and 56,320 NVIDIA Olympus Arm-compatible cores, according to the same March announcement.
- GX5000 interconnect: HPE described NVIDIA Quantum-X800 InfiniBand switches as offering 144 ports, each at 800 Gb/s, in that March announcement.
These are HPE product specifications, not general performance comparisons. Selection also depends on workload mix, networking, tenancy, and the cooling and power requirements of a deployment.
What performance claims has HPE made?
HPE’s June 2026 release included vendor-reported results tied to particular tests; they are not independent benchmarks or guarantees for other configurations.
Rank #3
- Small in Size, Serious in Performance — a space-saving design delivering professional-class performance, enterprise-grade security and reliability, flexible deployment options, and a MIL-STD-810H–certified build engineered for demanding work environments.
- Extreme AI and professional graphics performance — The ThinkStation P3 Ultra SFF Gen 2 combines an integrated Intel NPU with NVIDIA RTX 4000 SFF Ada Generation graphics (20GB GDDR6) to deliver up to 335 TOPS of AI performance across CPU and GPU. Ideal for AI inferencing, deep learning, 3D animation, content creation, advanced imaging, 3D modeling, and BIM software—all in a compact, energy-efficient workstation.
- Fast, secure storage with next gen memory & business-ready OS — 2TB PCIe Gen 5 TLC Opal SSD for ultra fast boot and load times, MAXED OUT 128GB DDR5-6400MHz memory, and Windows 11 Professional preinstalled.
- Easy-access front connectivity — USB-A (USB 10Gbps), 2 x USB-C (USB4 20Gbps) – data transfer only, Headphone/mic combo
- Warranty — Factory Sealed. 1 Year Lenovo Warranty
- 20.4× improvement in time to first token: HPE reported this result from testing a ProLiant DL380a Gen12 with eight NVIDIA H200 NVL GPUs, Alletra Storage MP X10000 with three controller nodes, and the NVIDIA Nemotron 70B model using KV-cache-aware inference optimization.
- Up to 20% higher token throughput: HPE based this claim on internal data across five standard Hugging Face inference and fine-tuning benchmark tests, against three popular large language models, on an HPE Private Cloud AI system.
Actual results can differ with models, software, configuration, and workload. HPE’s figures describe its stated test conditions rather than a universal expected improvement.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.When are the new systems and features expected?
Availability depends on the product and feature. Announced dates are schedules, not proof of shipping status.
- OpenShell in HPE Private Cloud AI: planned for Q4 2026.
- BlueField-4 across HPE systems: planned, with timing dependent on product lead times.
- NVIDIA Confidential Computing for HPE AI Factory through HPE Services: planned for Q4 2026.
- ProLiant DL394 Gen12 with NVIDIA Vera CPU: HPE said it is expected in 2027.
HPE’s earlier 2026 releases also put other items on schedules that have since passed or are still pending: March scheduled Private Cloud AI network expansion racks for July and Fortanix support with DL380a Gen12 for Q3; June scheduled new Private Cloud AI features for July and HPE Data Fabric Software for October, and assigned Alletra X10000 and NVIDIA Agent Toolkit/NemoClaw support to Q4. Those announcements do not establish whether an item is available now, so check with HPE for current status. Air-gapped Private Cloud AI, RTX PRO 6000 support, and NVIDIA AI-Q and Omniverse blueprints were described as available in March 2026.
Quick Recap
What should buyers verify?
- Which portfolio product fits the workload: private enterprise AI, a large or sovereign AI factory, or AI plus HPC and scientific computing.
- The exact GPU or CPU configuration, networking, storage, and scale being quoted; capacity figures in announcements refer to specific systems and capabilities.
- Whether the feature or hardware is shipping for the chosen configuration, rather than merely announced or scheduled.
- How agent permissions, approvals, monitoring, isolation, and audit are implemented in the intended deployment.
- Which compliance requirements apply and whether the specific configuration and deployment address them; HPE explicitly says this varies.
- For performance comparisons, whether the benchmark’s model, hardware, software, and workload match the buyer’s planned use.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
