Korean
At HPE Discover 2026, HPE unveiled AI infrastructure upgrades to empower enterprise agentic AI with enhanced governance, security and scalability. Expanding its NVIDIA AI Computing co-developed AI Factory portfolio, the updates accelerate enterprise AI transformation from pilot trials to stable production.
Focused on HPE Private Cloud AI and large-scale AI Factory environments, the upgrades optimize agent governance, data preparation, inference performance and confidential computing capabilities.
“Autonomous AI demands secure, governable and cost-scalable modern architectures,” said HPE CEO Antonio Neri. “HPE and NVIDIA deliver full-stack AI solutions covering networking, servers, storage and software, building a solid agentic enterprise foundation for reliable AI production deployment.”
“The entire computing stack is being rebuilt for AI agents,” said NVIDIA CEO Jensen Huang. “HPE and NVIDIA jointly build next-gen AI factories with NVIDIA Vera CPUs, accelerated hardware and secure AI software, helping enterprises convert data into intelligent operational actions.”
HPE Private Cloud AI: New Agentic AI Capabilities
Co-developed with NVIDIA, HPE’s turnkey Private Cloud AI platform receives major upgrades to support enterprise-grade AI agent deployment.It fully supports NVIDIA Agent Toolkit, including Nemotron open models, NemoClaw and OpenShell runtime, delivering complete frameworks for agent reasoning, policy enforcement, behavior monitoring and operational governance.
HPE also launched ProLiant Compute DL394 Gen12, a Vera CPU-powered server optimized for agentic AI workloads and high-speed data processing.
Extended Zerto capabilities monitor AI agent behaviors to identify anomalies and unauthorized operations. Its continuous data protection enables rapid restoration to secure baseline states to mitigate risks.Built-in local agent registration allows enterprises to centrally govern and approve internal AI models, tools and functional capabilities via unified security policies.
Streamlined Data Preparation & Higher Inference Efficiency
Targeting enterprise AI’s prevalent data readiness bottlenecks, HPE rolled out targeted optimizations for data processing and inference workloads.
HPE Alletra Storage MP X10000 features intelligent automation for unstructured data metadata tagging and governance policy enforcement, speeding up AI dataset preparation. It delivers up to 20x faster token response and 20% higher prompt-processing efficiency and token throughput.
HPE Data Fabric Software now supports MCP for Apache Airflow and enterprise AI inventory management, enriching distributed dataset metadata for better discoverability and governance. A standalone ProLiant-based Data Fabric appliance is also available for simplified deployment.
AI Infrastructure Scaling & Cost Optimization
HPE introduced new features to boost resource utilization and cut operational costs for large-scale AI deployments.A unified model gateway enables centralized, governed multi-model access with workload prioritization and 256-GPU scalable multi-node inference. Integrated NVIDIA NeMo tools support enterprise fine-tuning of Nemotron pre-trained models.These improvements elevate GPU utilization, reduce token consumption costs and sustain long-term AI infrastructure expansion.
Confidential Computing for Sovereign AI Factory
HPE integrates NVIDIA Confidential Computing across its AI Factory portfolio to serve large-scale and sovereign AI deployments.Hardware-rooted encryption and cryptographic attestation safeguard runtime AI models and sensitive data, helping enterprises meet regulatory and data sovereignty requirements with no performance loss.
NVIDIA BlueField DPUs and DOCA software enable zero-trust enforcement, runtime threat detection and full-stack encrypted networking, applicable to both scaled and sovereign HPE AI Factory deployments.
Upgraded NVIDIA Hardware Ecosystem Compatibility
HPE expands support for NVIDIA’s latest-generation AI hardware across its AI Factory lineup.The portfolio now supports NVIDIA RTX PRO 6000 Blackwell GPUs, Spectrum-X Ethernet, BlueField-3 DPUs and ConnectX-8 SuperNICs. Based on NVIDIA reference architectures, the solutions cover end-to-end AI lifecycles, integrated with NVIDIA AI Enterprise software and HPE’s Unleash AI partner ecosystem.
Availability Timeline
The unified model gateway and core Private Cloud AI enhancements will launch in July 2026, with Data Fabric Software updates scheduled for October 2026.Q4 2026 will bring additional upgrades including agentic observability, Alletra storage integration, and full NVIDIA Agent Toolkit and NemoClaw support.
Zerto agent monitoring/recovery workflows and NVIDIA Confidential Computing support for AI Factory will also arrive in Q4 2026.
Compatibility with the latest NVIDIA Blackwell GPUs, Spectrum-X networking and modern DPUs/SuperNICs is immediately available.
Beijing Qianxing Jietong Technology Co., Ltd.
Sandy Yang/Global Strategy Director
WhatsApp / WeChat: +86 13426366826
Email: yangyd@qianxingdata.com
Website: www.qianxingdata.com/www.storagesserver.com
Business Focus:
ICT Product Distribution/System Integration & Services/Infrastructure Solutions
With 20+ years of IT distribution experience, we partner with leading global brands to deliver reliable products and professional services.
“Using Technology to Build an Intelligent World”Your Trusted ICT Product Service Provider!
Sandy Yang/Global Strategy Director
WhatsApp / WeChat: +86 13426366826
Email: yangyd@qianxingdata.com
Website: www.qianxingdata.com/www.storagesserver.com
Business Focus:
ICT Product Distribution/System Integration & Services/Infrastructure Solutions
With 20+ years of IT distribution experience, we partner with leading global brands to deliver reliable products and professional services.
“Using Technology to Build an Intelligent World”Your Trusted ICT Product Service Provider!