Broadcom and NVIDIA Team Up to Make VMware Cloud Foundation AI-Native

Broadcom has announced a new partnership with NVIDIA aimed at transforming VMware Cloud Foundation into an AI-native platform for private clouds. The collaboration, unveiled at VMware Explore 2025 in Las Vegas, focuses on weaving advanced GPU acceleration and networking into enterprise infrastructure so that companies can more easily build, train, and deploy AI models in secure environments.
Central to the upgrade is support for NVIDIA’s latest Blackwell GPU architecture, paired with high-performance networking technologies such as ConnectX-7 adapters and BlueField-3 DPUs. Together, these components are designed to move massive volumes of data quickly across clusters, enabling enterprises to train complex models without the typical bottlenecks that slow multi-node systems.
Broadcom also introduced VMware Cloud Foundation 9.0, now positioned as an AI-native release. It includes built-in Private AI Services at no additional cost, giving organizations access to tools like GPU monitoring, vector databases, model storage, agent builders, and data indexing. The idea is to give IT teams a unified foundation for conventional workloads while also streamlining AI adoption under the same operational playbook.
Executives stressed that the new AI features are fully integrated into VMware’s established workflows—like vMotion, High Availability, and Distributed Resource Scheduler—so enterprises can scale their AI projects without disrupting core operations. By tying NVIDIA’s hardware muscle with Broadcom’s cloud management expertise, the move signals a push to keep advanced AI development inside corporate walls, where data privacy and control remain top priorities.