VMware announced new collaborations focused on the Private AI they introduced earlier this year at VMware Explore Las Vegas, with long-time partners Intel and IBM being the big two, as well as a third with IBM spinoff Kyndryl.
Today at the European VMware Explore 2023 being held in Barcelona, VMware increased their efforts to empower customers to adopt generative AI in the enterprise. They built on the Private AI announcements they made this summer at VMware Explore 2023 Las Vegas, and highlighted the expansion of ecosystem partnerships around Generative AI, while announcing new fruits from several of these partnerships.
“We are starting to see this new infusion of AI into the application space as well, which we see as really that next wave of app innovation,” said Chris Wolf, VP of VMware AI Labs. “Through generative AI, we are seeing large language models have an impact on a large number of use cases, from software development to customer service and support, and content generation for marketing, among others.”
Wolf said that in reaction to the core challenge around this type of AI, which was centred around privacy, led VMware to Private AI.
“This is an architectural approach that balances the business gains from AI with the privacy and compliance needs of the organization,” he stated. “Some of the core tenants for private AI are not just around privacy, but also choice, such as services above the stack, whether that’s commercial open source, as well as choice of hardware and options below the stack as well. We believe Private AI will become the default architecture for enabling Generative AI in the enterprise.
“Our focus for Explore Barcelona has been around growing our ecosystem and our partnerships,” Wolf emphasized.
To this end, three new collaborations were announced. The first was the introduction of VMware Private AI with Intel, a new collaborative effort to help enterprises build and deploy AI models and boost AI performance by harnessing the power of VMware Cloud Foundation and Intel AI software kit, processors, and hardware accelerators. It will deliver reference architecture on building a validated and benchmarked AI stack for data preparation, model training, fine-tuning and inferencing to accelerate scientific discovery and enrich business and consumer services.
Specifically, this collaboration will enable the fine-tuning of task-specific models in minutes rather than hours and the inferencing of large language models at faster than human communication using the customer’s private corporate data. This makes it possible to fine-tune smaller, economical state of the art models which are easier to update and maintain on shared virtual systems, which can then be delivered back to the IT resource pool when the batch AI jobs are complete.
“In addition to the reference architecture, we will also be showcasing our partnership with Intel at Explore,” Wolf said. “We see good energy and good opportunity in the partnership.”
The second announcement was a collaboration with IBM, designed to bring IBM watsonx to on premises environments, through both VMware Private AI and Red Hat OpenShift.
“We have had good success with IBM’s watsonx in the cloud, but customers have told us that they also want to being it on-prem,” Wolf indicated. The collaboration will let enterprises access IBM watsonx in private, on-premises Infrastructure as a Service environments as well as hybrid cloud with watsonx SaaS offerings on IBM Cloud. Joint clients will be able to utilize IBM watsonx AI capabilities, data management, and in the near future, its governance capabilities across public clouds, virtual private clouds, data centers, and edge sites.
The third announced collaboration is between VMware Private AI and Kyndryl, which was IBM’s infrastructure services business until it was spun off in 2021.