With AI use continuing to grow in adoption throughout enterprise IT, Deloitte is creating a new Deloitte Center for AI Computing to advise its customers, explain the technology and help them use it in their ongoing business and growth plans.
Designed to provide a cloud-accessible accelerated platform that Deloitte clients can use to test and explore various AI strategies and tools, the infrastructure features six Nvidia DGX A100 systems, Nvidia Mellanox networking, high-performance storage and Nvidia software. The platform will be physically hosted in Deloitte’s datacenter in Hermitage, Tenn., but will also be part of a virtual service.
Based on Nvidia’s DGX POD reference architecture and harnessing the power of 48 A100 GPUs, the installation will serve as an AI acceleration engine for Deloitte’s clients, the company said. The new facility was announced on Tuesday (March 2).
“The Nvidia systems are being deployed this month and the center will be available to clients starting later in March,” Nitin Mittal, the AI strategic growth offering leader for Deloitte Consulting, told HPCwire sister site EnterpriseAI.
Deloitte team members will serve clients through the center from several locations in the U.S. and India, said Mittal. The virtual services will start with 20-25 practitioners trained in skill sets including AI architecture, ML engineering, data science and front- and back-end development skills using Nvidia’s GPU-based architecture and accelerated computing infrastructure, he added.
“The center provides a platform for Deloitte to co-innovate and co-invest with its clients and speed the development of AI applications,” said Mittal. “As organizations accelerate the adoption of AI, the center provides them with a technology platform, deep AI expertise, and supercomputing resources to speed the development of a wide range of AI applications, from autonomous vehicles to digital contact centers to public sector innovation.”
Based on Nvidia’s specifications, six DGX A100 machines together provide ~500 teraflops of peak double-precision, 1.0 petaflops of peak single-precision, or 7.5 petaflops of peak single-precision AI tensor performance.
A price tag for the project was not publicly disclosed, however DGX A100 systems list for $199,000 each.
Jason Girzadas, managing principal of Deloitte Consulting, said in a statement that his company’s collaboration with Nvidia “can enable clients to quickly deliver on the full promise of AI solutions to transform both their businesses and the basis of professional services.”
The new facility will significantly expand Deloitte’s ability to develop innovative AI products and services that previously were being undertaken by Deloitte’s AI Exploration Lab in Austin, Texas, and the AI Factory in Canada, according to the company.
Read the rest of the story at EnterpriseAI.news.