GPUs are often the most expensive part of an AI infrastructure investment. How you virtualize and share them determines how much value you get out of them.

At VMware Explore 2026, Shawn Kelly and I covered the practical side of running GPUs on VMware Cloud Foundation. We compared passthrough and vGPU, explained time-slicing and Multi-Instance GPU (MIG), and shared real-world benchmarks. We also walked through best practices for placement, DRS, and vMotion. Whether you’re starting your first AI project or running LLMs at scale, this session will help you get more from every GPU.

Watch the session on the VMware Explore replay page.

Wrapping up

Every GPU in your environment is a significant investment, and VMware Cloud Foundation gives you the tools to share, place, and monitor it wisely. This session walks through your options so you can match the right approach to each AI workload.

Also check out my session on Deploying and Governing AI Models and Agents with VCF Private AI Services.