Nvidia on Tuesday announced the general availability of Nvidia AI Enterprise, a software suite that lets companies virtualize AI workloads on mainstream servers running VMware vSphere. As part of the GA release, Nvidia also said that it’s partnering with Domino Data Lab to integrate its MLOps platform on top of Nvidia AI Enterprise.
Nvidia AI Enterprise brings together three components: The Nvidia RAPIDS suite of software libraries; frameworks like TensorFlow and PyTorch; and Nvidia Tensor, the open-source inference serving software. The software suite packages these elements together with full enterprise-level support. It’s designed to run on mainstream servers from OEMs like Dell, HP, Lenovo and others.
“The idea here is the same servers that have been racked and stacked into private clouds and enterprise data centers today, can now be utilized for AI, with a small amount of GPU added to the server [at] affordable, accessible, incremental costs,” Manuvir Das, head of Enterprise Computing at Nvidia, said to reporters.
Nvidia decided to partner with VMware to bring this software suite to market, he said, since vSphere is “the de facto operating system of the enterprise data center.”
Das laid out the opportunity that Nvidia sees for accelerating AI adoption: While many businesses are already adopting AI for core strategic use cases, “there’s a variety of functions that every enterprise customer does every day, regardless of what business they’re in,” he said. “Whether it’s human resources, or managing their sales team, or operations, supply chain, etc. And this is where we expect to see a big wave of AI to follow, where all of these line of business applications that are used by enterprise customers will be infused with AI.”
Nvidia says early adopters of Nvidia AI Enterprise include dozens of companies in a range of industries, including automotive, education, finance, healthcare, manufacturing and technology. That includes Cerence, which is using it to develop intelligent in-car assistants, as well as the Italian public research university the University of Pisa.
OEMs offering Nvidia-certified systems that support Nvidia AI Enterprise include Atos, Dell Technologies, GIGABYTE, Hewlett Packard Enterprise, Inspur, Lenovo and Supermicro. Additionally, Dell separately announced Dell EMC VxRail as the first hyperconverged platform to be qualified as an Nvidia-Certified System for Nvidia AI Enterprise. Nvidia AI Enterprise is available via channel partners. Subscription licenses start at $2,000 per CPU socket for one year and include Business Standard Support. Perpetual licenses are $3,595 and require additional support purchase.