As advances in graphics processing units accelerate enterprise artificial intelligence adoption, GPU virtualization is emerging as the foundation for where critical workloads ultimately run. While public cloud remains useful for early pilots, production deployments are shifting toward controlled environments that prioritize performance, security and efficiency.

This transition is sparking closer collaboration between Broadcom Inc. and Advanced Micro Devices Inc., as the two companies engineer infrastructure purpose-built for private AI with GPU virtualization at its core. Running on AMD’s latest processors and GPUs, the joint platform is designed to deliver an open, streamlined stack that meets enterprise demands for both scale and trust, according to Kumaran Siva (pictured, left), corporate vice president of strategic business development at AMD.

“I’ve heard so many customers say POC in cloud, deploy on-prem,” Siva said. “Private cloud is going to be a real thing, absolutely. Regulated industries need to keep data secure. This is going to be a challenge that … the partnership here really goes a long way to addressing.”

Siva and Paul Turner (right), vice president of products, VMware Cloud Foundation Division, at Broadcom, spoke with theCUBE’s John Furrier and Paul Nashawaty at VMware Explore, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed how their companies are supporting private AI deployment at scale, the role of GPU virtualization in the enterprise and what it takes to bring agentic platforms into production. (* Disclosure below.)

GPU virtualization powers enterprise-scale AI

One focus of the Broadcom and AMD partnership is enabling advanced virtualization across AI workloads. As enterprise use of agents and large models grows, organizations need ways to support more concurrency and greater resource efficiency. GPU virtualization allows multiple application servers to access trusted models and inference engines simultaneously, improving scalability across the AI private cloud, according to Turner.

“We are building an ability for our inferencing engine,” he said. “You can actually do shared inferencing at the same model. We already have model sharing that we can actually manage as trusted models that you share across your enterprise. Now you’re going to be able to deploy them onto your GPUs and actually have multiple application servers actually share those.”

The partnership is also developing a streamlined software stack to simplify private AI deployment. With components such as a vector database and model manager, the system supports GPU virtualization for rapid deployment and improved workload performance in the AI private cloud, according to Siva.

“We really view VMware and the VCF package as being a critical component to making our GPUs easy to use and easy to adopt,” he said. “We have our GPU operator, we have essential components that we are working very closely together with Broadcom to make sure that we enable the full ease of use.”

Optimizing performance in the AI private cloud

A key goal of the collaboration between Broadcom and AMD is to help enterprises extract more value from their hardware. With AI workloads running at scale, virtualization is essential to meeting SLAs, reducing latency and ensuring dynamic workload placement. These capabilities are foundational to a responsive private cloud environment, according to Turner.

“We have to actually virtualize GPUs properly,” Turner said. “We have to be able to take advantage of them. We can go and build a pool, [a] GPU as a service pool of GPUs, with AMD. We can actually use our [Distributed Resource Scheduler] capability to be able to do automatic … identification of best and optimal placement of that resource against that pool of GPUs.”

The companies are also focused on optimizing the hardware-software stack for enterprise AI. Deep integration between AMD hardware and VMware Cloud Foundation enables seamless workload mobility and high concurrency across AI models. That level of infrastructure harmony is only possible with GPU virtualization at the core, according to Siva.

“I think its great utilization is super important to enterprises,” Siva said. “We’ve seen … that in the CPU world where … we have tens of thousands of applications that enterprise is running today on VCF with Epic CPUs underneath, seamlessly running live, migrating all of those features. We will bring that to GPUs.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of VMware Explore:

(* Disclosure: Broadcom Inc. sponsored this segment of theCUBE. Neither Broadcom nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE