Dell, Nvidia Taking Users to School with New Generative AI Blueprints
Dell Technologies and one of its most illustrious partners, Nvidia — sensing “urgency” among enterprises for more and easier-to-understand AI coaching and implementation instruction — today launched a new suite of reference-model blueprints aimed at teaching enterprises how to adopt and manage generative AI (GenAI) for immediate use.
Dell Generative AI Solutions, developed side-by-side with Nvidia, were built to help enterprises with the relatively new aspects of deploying AI, namely managing and designing inferencing, strategy, implementation and scaling. The new reference models, which manage entire enterprise stacks, also will support workloads locally on devices and provide training and other critical tools for developers.
With a well-planned and verified inferencing approach, enterprises can rapidly deploy GenAI projects and scale applications to upgrade processes in key areas, such as customer operations, content creation and management, software development and sales. Ultimately, enterprises will use the Dell/Invidia GenAI suite for managing both legacy and emerging needs in edge management, securing generative AI development, installing and maintaining zero trust security and managing multicloud workload management, the company said.
So why is there such “urgency” for companies to begin using this in the marketplace?
“Well, there are data sovereignty and privacy regulations, because your data is stored all over, and allowing customers to have an on-prem or an enterprise generative AI effort allows for that maximization of those constraints,” Dell VP of Cross-Portfolio Software and Solutions Carol Wilder told The New Stack. “That’s where the urgency is. I think that generative AI also presents a watershed moment for customers to re-envision their own data and what their own data can do for them. They may not want to do this in the public cloud.
“I look at Dell’s role as Switzerland. We’re here to help people have options.”
New Blueprints Will Work with NativeEdge, Project Helix
Dell and Nvidia’s news was a followup on a set of announcements last May at Dell Technology World, which included the introduction of its NativeEdge platform; the partnership between itself and Nvidia to build Project Helix, a secure on-premises-based generative AI management package; and several other announcements. All of these products will be utilized in the reference models that comprise the Generative AI Solutions package.
Edge operations, which make use of automated processes powered by AI and large language models, are generally deployed in locations ranging from manufacturing floors and retail stores to remote wind turbines and long-distance hospital operating rooms. The new Dell-Invidia GenAI services also span IT infrastructure, PCs and professional services in order to simplify the adoption of full-stack GenAI with large language models (LLM).
Adding AI to Legacy Systems Can Be Painstaking
Users can start with a pre-built foundation rather than testing and experimenting from scratch. Wilder said that Nvidia has a number of pre-tested, pre-trained and pre-validated models built for many use cases.
Adding AI or machine learning-based components to legacy systems that were not designed with that technology in mind can be a painstaking — and often frustrating — process.
“When we (Dell) did it (implementing AI into its own various applications), we documented what options we used, so that it simplifies the choices where they (users) can say, ‘Look, I know Dell did this, these were the results,'” Wilder told The New Stack. “And then if there’s a SKU or changes that customers want to make, then they can delta off that versus the wild wild west of ‘Wait a minute: What do I choose?’
“When we look at users, there’s a set of very mature customers that are just going to do it by themselves. Then there are very immature customers that are starting right now in the consultation role. And then there’s this middle part, which has some traction, and they’re like, ‘Okay, what do I do next?’ This provides a really good blueprint for those customers.”
Naturally, Dell data center hardware to run the Nvidia software is specified in the blueprints. The solution calls for Dell PowerEdge XE9680 or PowerEdge R760xa servers with Nvidia Tensor Core GPUs, Nvidia AI Enterprise software, and Nvidia NeMo. Customers can also combine this with unstructured data storage. Components ostensibly can be swapped out to the caution of users.
GenAI Market Is Billowing up, and Fast
The worldwide GenAI market is projected to reach a whopping $52 billion by 2028, representing a CAGR of more than 35% from 2023. The technology is anticipated to create at least $450 billion in the enterprise across a large number of verticals over the next seven years.
A recent survey by Deloitte found that more than one-third (37%) of CEOs are implementing some form of GenAI, while more than half (55%) are evaluating or experimenting with the technology. In addition, 79% of CEOs believe GenAI will increase efficiency and 52% believe it will increase growth opportunities.