VMware and NVIDIA unlock Generative AI for enterprises

VMware Private AI Foundation with NVIDIA will enable enterprises to customize models and run generative AI applications, including intelligent chatbots, assistants, search and summarization. The platform will be a fully integrated solution featuring generative AI software and accelerated computing from NVIDIA, built on VMware Cloud Foundation and optimized for AI.

“Generative AI and multi-cloud are the perfect match,” said Raghu Raghuram, CEO, VMware. “Customer data is everywhere — in their data centers, at the edge, and in their clouds. Together with NVIDIA, we’ll empower enterprises to run their generative AI workloads adjacent to their data with confidence while addressing their corporate data privacy, security and control concerns.”

“Enterprises everywhere are racing to integrate generative AI into their businesses,” said Jensen Huang, founder and CEO, NVIDIA. “Our expanded collaboration with VMware will offer hundreds of thousands of customers — across financial services, healthcare, manufacturing and more — the full-stack software and computing they need to unlock the potential of generative AI using custom applications built with their own data.”

Full-Stack Computing to Supercharge Generative AI

To achieve business benefits faster, enterprises are seeking to streamline development, testing and deployment of generative AI applications. McKinsey estimates that generative AI could add up to $4.4 trillion annually to the global economy.(1)

VMware Private AI Foundation with NVIDIA will enable enterprises to harness this capability, customizing large language models; producing more secure and private models for their internal usage; and offering generative AI as a service to their users; and, more securely running inference workloads at scale.

The platform is expected to include integrated AI tools to empower enterprises to run proven models trained on their private data in a cost-efficient manner. To be built on VMware Cloud Foundation and NVIDIA AI Enterprise software, the platform’s expected benefits will include:

Privacy — Will enable customers to easily run AI services adjacent to wherever they have data with an architecture that preserves data privacy and enable secure access.

Choice — Enterprises will have a wide choice in where to build and run their models — from NVIDIA NeMo™ to Llama 2 and beyond — including leading OEM hardware configurations and, in the future, on public cloud and service provider offerings.

Performance — Running on NVIDIA accelerated infrastructure will deliver performance equal to and even exceeding bare metal in some use cases, as proven in recent industry benchmarks.

Data-Center Scale — GPU scaling optimizations in virtualized environments will enable AI workloads to scale across up to 16 vGPUs/GPUs in a single virtual machine and across multiple nodes to speed generative AI model fine-tuning and deployment.

Lower Cost — Will maximize usage of all compute resources across, GPUs, DPUs and CPUs to lower overall costs, and create a pooled resource environment that can be shared efficiently across teams.

Accelerated Storage — VMware vSAN Express Storage Architecture will provide performance-optimized NVMe storage and supports GPUDirect® storage over RDMA, allowing for direct I/O transfer from storage to GPUs without CPU involvement.

Accelerated Networking — Deep integration between vSphere and NVIDIA NVSwitch™ technology will further enable multi-GPU models to execute without inter-GPU bottlenecks.

Rapid Deployment and Time to Value — vSphere Deep Learning VM images and image repository will enable fast prototyping capabilities by offering a stable turnkey solution image that includes frameworks and performance-optimized libraries pre-installed.

The platform will feature NVIDIA NeMo, an end-to-end, cloud-native framework included in NVIDIA AI Enterprise — the operating system of the NVIDIA AI platform — that allows enterprises to build, customize and deploy generative AI models virtually anywhere. NeMo combines customization frameworks, guardrail toolkits, data curation tools and pretrained models to offer enterprises an easy, cost-effective and fast way to adopt generative AI.

For deploying generative AI in production, NeMo uses TensorRT for Large Language Models (TRT-LLM), which accelerates and optimizes inference performance on the latest LLMs on NVIDIA GPUs. With NeMo, VMware Private AI Foundation with NVIDIA will enable enterprises to pull in their own data to build and run custom generative AI models on VMware’s hybrid cloud infrastructure.

At VMware Explore 2023, NVIDIA and VMware will highlight how developers within enterprises can use the new NVIDIA AI Workbench to pull community models, like Llama 2, available on Hugging Face, customize them remotely and deploy production-grade generative AI in VMware environments.

Broad Ecosystem Support for VMware Private AI Foundation With NVIDIA

VMware Private AI Foundation with NVIDIA will be supported by Dell Technologies, Hewlett Packard Enterprise and Lenovo — which will be among the first to offer systems that supercharge enterprise LLM customization and inference workloads with NVIDIA L40S GPUs, NVIDIA BlueField®-3 DPUs and NVIDIA ConnectX®-7 SmartNICs.

The NVIDIA L40S GPU enables up to 1.2x more generative AI inference performance and up to 1.7x more training performance compared with the NVIDIA A100 Tensor Core GPU.

NVIDIA BlueField-3 DPUs accelerate, offload and isolate the tremendous compute load of virtualization, networking, storage, security and other cloud-native AI services from the GPU or CPU.

NVIDIA ConnectX-7 SmartNICs deliver smart, accelerated networking for data center infrastructure to boost some of the world’s most demanding AI workloads.

VMware Private AI Foundation with NVIDIA builds on the companies’ decade-long partnership. Their co-engineering work optimized VMware’s cloud infrastructure to run NVIDIA AI Enterprise with performance comparable to bare metal. Mutual customers further benefit from the resource and infrastructure management and flexibility enabled by VMware Cloud Foundation.