Covering Disruptive Technology Powering Business in The Digital Age

Written by: Martin Dale Bolima, Journalist, AOPG

NVIDIA announced on Tuesday, August 24, the general availability of its AI Enterprise Software Version 1.0, finally making it available after its limited early release. This development means that organisations looking to leverage the powerful combination of Artificial Intelligence (AI) and machine-learning (ML) can now use NVIDIA’s pioneering and much-anticipated software suite, which features NVIDIA’s software library suite RAPIDS, its open-source inferencing software Tensor and frameworks such as PyTorch.

“AI is real, and its value has been proven to many, many companies around the world,” said Manuvir Das, head of enterprise computing at NVIDIA, at the virtual media brief announcing the software’s general availability. “At the same time, we also realised both from our experience within NVIDIA… and our experience working with many companies that AI is really difficult to implement.”

The difficulty lies in AI being both a full-stack problem and an end-to-end effort. The former, in particular, requires AI implementation to have the “best of breed hardware,” the right software to accelerate different applications and “a wide ecosystem of application developers.” The latter, on the other hand, necessitates data acquisition and storage, along with data processing to make it useful for AI and inferencing later on.

NVIDIA, according to Das, covers all those bases.

“We believe that NVIDIA is the only company in the world that is capable of putting this entire solution together because of all the technology we’ve already built, as well as the experience that we have,” said Das. “It requires a full stack, and we believe NVIDIA is uniquely the company that can do this. This is our mission.”

The general availability of Version 1.0 means NVIDIA is now offering a complete software suite that goes hand-in-hand with its hardware products, like its powerful GPUs and DPUs, and its range of application frameworks, including Metropolis for smart cities, Isaac for robotics and Clara for healthcare.

Critically, Version 1.0, which Das describes as “the operating system of AI,” comes with full enterprise-level support and is specifically designed to run on mainstream servers, including HP, Lenovo and Dell.

“The idea here is the same servers that have been racked and stacked into private clouds and enterprise datacentres today can now be utilised for AI, with a small amount of GPU added to the server [at] affordable, accessible, incremental costs,” explained Das.

Critically, Version 1.0 can be easily installed by enterprises whose virtualisation platform is VMWare’s vSphere 7 Version 2. This is especially critical given how 70% of enterprise data centres are running VMWare as their virtualisation platform today. It is also the reason NVIDIA has partnered with VMware, with Das noting how vSphere is now “the de facto operating system of the enterprise data centre.”

With this collaboration, Das forecasts Version 1.0 further accelerating AI adoption beyond core strategic use cases. “There are a variety of functions that every enterprise customer does every day, regardless of what business they’re in—whether it’s human resources, or managing their sales team, or operations, supply chain, etc.,” said Das. “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.”

These developments bode well for AI’s early adopters, most of which are in the manufacturing, automotive, finance, technology, education and healthcare industries, as well as for soon-to-follow enterprises in a range of other sectors.

Another positive development is NVIDIA’s collaboration with Enterprise MLOps platform Domino Data Lab, which Das also announced in the same virtual briefing. Domino’s MLOps platform, according to Das, ensures “reproducibility” as well as improved efficiency to running AI applications supported by ML models.

“It makes the work of the researcher reproducible by the IT administration team that is going to deploy into production so that they can validate what they’re putting into production, and there’s governance and they can track versions and data sets and all of that,” Dar explained. “So, we are very pleased to announce that as part of the general availability of NVIDIA’s AI Enterprise, we now have a partnership with Domino Data Lab, and what they’re doing is integrating their platform with NVIDIA AI Enterprise.”

Organisations interested in the NVIDIA AI Enterprise Version 1.0 can get it through channel partners, with subscriptions starting at USD $2,000 annually for one CPU socket plus Business Standard Support. Perpetual licences are also available at USD $3,595, but support for it will have to be purchased separately. Version 1.0 is currently supported by NVIDIA-certified systems, including OEMs such as Dell Technologies, HP Enterprise, Lenovo and Atos.

(0)(0)