When first released, the joint platform enabled companies to deploy generative AI applications and included a vector database so that companies could use Retrieval Augmented Generation (RAG) to make their generative AI provide more accurate and timely answers.

“The piece of the puzzle we were missing was a model store manager,” says Paul Turner, vice president of products for VMware Cloud Foundation at Broadcom.

Companies can provide their developers with a curated selection of AI models and thus establish access controls for these models.

“And it makes sure that no one is just using general purpose large language models that you don’t want to support,” says Turner. “Because out there on the internet, you don’t know the provenance of that LLM or where it came from. This gives you a way to manage those LLMs across your entire user base, so you can really give them the opportunity to leverage their generative AI innovation.”

VMware customers can use Nvidia’s AI models, as well as models from Hugging Face and other partners, including Meta’s Llama 3 model and models from Google and Mistral. “Whatever Nvidia supports, we support,” Turner says.

Other new features include a model repository, tools for securing models with built-in access controls, an optimized deployment workflow, and reference AI workflows for specific use cases such as customer service, drug discovery, and PDF data extraction.