Broadcom & NVIDIA expand AI in private clouds with VCF boost

Broadcom announced a collaboration with NVIDIA to bring the latest NVIDIA AI technology to VMware Cloud Foundation (VCF), enhancing AI capabilities within private cloud environments for enterprises and cloud service providers.

The partnership aims to enable customers to build, deploy, and scale advanced AI models on modern AI servers while maintaining the operational workflows and enterprise-grade features provided by VCF. The focus is on supporting generative, agentic, and physical AI applications that are significantly shaping and transforming data centre infrastructure requirements.

Broader NVIDIA support

This integration extends Broadcom's longstanding work with NVIDIA, now including support for NVIDIA's new Blackwell Architecture GPUs, such as the NVIDIA RTX PRO 6000 Server Edition GPUs. The collaboration also incorporates NVIDIA networking components to deliver wider options for accelerated computing solutions. These updates are intended to increase flexibility and operational efficiency for AI workloads in private clouds.

The expanded offerings will also enhance the jointly developed VMware Private AI Foundation with NVIDIA, providing a platform for organisations to utilise the latest NVIDIA advancements with their existing enterprise systems.

"Our customers want the freedom to innovate with AI while continuing to rely on the enterprise platforms they trust," said Paul Turner, Vice President of Products, VMware Cloud Foundation Division, Broadcom. "Our partnership with NVIDIA delivers exactly that – enabling organizations to build, deploy and scale AI workloads alongside their existing applications without compromising on performance, efficiency, availability or operational simplicity. It's about blending cutting-edge AI innovation with the enterprise-grade reliability and manageability for which VCF is known."

Justin Boitano, Vice President of Enterprise AI at NVIDIA, added his perspective on the demand generated by new AI technologies:

"The emergence of generative AI is driving the need for a new level of accelerated computing infrastructure. With NVIDIA networking technology and Blackwell GPUs, enterprises can build and deploy powerful AI applications directly within their existing private cloud, using the full capabilities of the NVIDIA platform in combination with VMware Cloud Foundation."

Key technical integrations

Support for the NVIDIA Blackwell Architecture GPUs is a central element of the integration. These GPUs are designed for large-scale AI training, inference, and high-performance computing. The update specifically includes support for the NVIDIA RTX PRO 6000 Server Edition GPUs, targeted at enterprises that require efficient handling of both virtual desktop infrastructure (VDI) and AI workloads.

Future releases will introduce support for NVIDIA Blackwell B200 GPUs, designed to deliver high performance for AI and high-performance computing (HPC) applications at scale.

In terms of networking, VCF will soon support NVIDIA ConnectX-7 network interface cards and BlueField-3 400G DPUs, using DirectPath I/O. This setup will enable customers to utilise advanced capabilities, including NVIDIA GPUDirect RDMA and GPUDirect Storage, which are essential for multi-node AI model training and large-scale data transfer, particularly relevant for generative AI use cases.

Operational continuity

A key aspect of the collaboration is maintaining VCF's core operational features, even as new hardware is introduced. Customers will be able to deploy NVIDIA technologies such as HGX servers without disrupting existing workflows or enterprise virtualisation capabilities. Features like vMotion, High Availability (HA), Distributed Resource Scheduler (DRS), and Live Patching remain intact, supporting both traditional enterprise applications and demanding AI or machine learning workloads within the same environment.

This approach aims to provide organisations with a flexible platform to manage a variety of computing needs on-premises while using familiar operational practices. The preservation of these enterprise capabilities is intended to reduce barriers for customers considering the adoption of advanced AI technologies within their private data centres.

Future outlook

The ongoing integration benefits VCF customers by increasing the range of available NVIDIA hardware and features, thereby expanding their options for building AI-optimised private cloud infrastructure. The upcoming support for the latest GPUs and networking components is expected to improve further efficiency and performance for companies employing data-intensive AI and HPC workloads.

The companies jointly emphasise that these enhancements will enable enterprises to run AI, machine learning, and traditional workloads concurrently, maintaining operational simplicity and reliability at scale.