VMware Inc. and Nvidia Corp. today announced an expanded partnership with a new solution that will enable enterprise businesses to build, train and deploy generative artificial intelligence models for applications customized with their own data that preserves security and privacy.

The new offering, called VMware Private AI Foundation with Nvidia, is a full-stack platform that features generative AI software along with everything an enterprise customer needs to build, train, customize and deploy an AI model out of the box. It’s built on VMware Cloud Foundation and optimized for AI, featuring generative AI software and accelerated computing from Nvidia.

“Generative AI and multicloud are the perfect match,” said VMware Chief Executive Raghu Raghuram. “Customer data is everywhere — in their data centers, at the edge and in their clouds.”

As more businesses look to use generative AI large language models in their applications, they have begun to face risks when it comes to using proprietary business data, including the potential that it might be exposed. With the Private AI Foundation solution, enterprise customers will be able to run AI services adjacent to their data and preserve data privacy to enable increased security while using VMware.

The number of models to which the enterprise has access is also extremely broad and choosing the right one can make or break an application given how different models handle data. The Private AI Foundation will allow customers to train and deploy any given model including Falcon LLM, Meta Platform Inc.’s LLaMA 2 or MosaicML Inc.’s MPT. Users can easily work with their own models or with open-source models from the community available on Hugging Face.

Under the hood, customers will get access to deep learning virtual machines, vector databases, graphical processing unit scaling and prebuilt images that are prepared specifically for use cases common to various industries. Developers will also have instant access to Nvidia’s AI Workbench, a workspace for developers that’s designed to allow the creation, testing and customization of pretrained generative AI models.

The platform also features NeMo, Nvidia’s cloud-native end-to-end framework that allows enterprises to build, customize and deploy their own generative AI models almost anywhere. NeMo acts as the underlying “operating system” for Nvidia’s AI platform and delivers customization operations, guardrail tools, data curation and models for customers to roll their own AI.

“Enterprises everywhere are racing to integrate generative AI into their businesses,” said Nvidia founder and CEO Jensen Huang. “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.”

The two companies said the Private AI Foundation platform will be supported by Dell Technologies Inc., Hewlett Packard Enterprise Co. and Lenovo Ltd. These companies will be among the first to offer systems that will be capable of providing LLM customization and inference workloads using Nvidia L40S graphical processing units, Nvidia BlueField-3 data processing units and Nvidia ConnectX-7 SmartNICs.

The VMware Private AI Foundation with Nvidia solution is expected to be released in early 2024.

Image: Pixabay