Manuvir Das, Head of Enterprise Computing | NVIDIA
The Future of AI in the Enterprise: NVIDIA’s Manuvir Das spoke at DDN’s bi-annual user group about the state of AI in the enterprise, and NVIDIA’s strategy. His November 15 presentation, held virtually at SuperComputing 2021, provides fresh perspectives and insights into the evolution of enterprise data centers, and the future of AI infrastructure.
See an extract from the Future of AI in the Enterprise session below, and click play to watch the entire session.
AI is the Most Important Application of HPC
“We’re all familiar with HPC. I would make the claim to you that AI already is becoming and certainly will be going forward the most important application of HPC, and the reason is because it’ll be applicable to every enterprise company in every industry, not just a few people. AI is the ultimate HPC application, because it involves a lot of data and a lot of compute that need to be done together. But there’s a myriad of use cases.
AI will Impact Every Industry
My second claim to you is that AI will impact every industry. If you look at the journey of AI to date, certainly from NVIDIA’s point of view and for many of you in the audience as well, we can be proud of the fact that there are thousands of companies now that have adopted AI. I think we counted something like 25,000 companies that use NVIDIA AI in one form or the other.
But we know that there are hundreds of thousands of enterprise companies. So this is the tip of the iceberg. In this tip of the iceberg with early adopters, the use cases we’ve seen are the obvious use cases where the business of the company is really of a form where AI would make a huge difference.
Every Business Application will be Infused with AI
For example, if I’m from an automotive company and I’m working on self-driving cars. Yes, you need AI to recognize objects in the road. If I’m from a shopping website, I need to show people what is the next thing they should buy based on their behavior of the last few months. Yes, you need AI models to do that. But the reality is that, in fact, there are business functions that are common to every company in every industry that are waiting to be infused with AI. And we really believe that this is where the democratization of AI will really come in.
The Future of AI in the Enterprise
“This is where you’ll see the spread of AI. So just look at this slide here with some of the use cases. Which company does not do marketing and sales? Almost every company does. And when you do marketing and sales, you need personalized recommendations for your customer base, what should they do next? What should they read next? What should they buy next? And that is a key application of AI, where you take all of the data that you have and you generate recommendations.
Recommendations are the New Engine of the Internet
“Recommendations are the next generation of search, and that applies to every company. You can look at customer service and a chat bot. What is a chat bot? A chat bot is really an application of conversational AI, a human customer, employee, what have you, says something in a language of their choice. They ask a question in the language of their choice and there’s a three-step process. You translate what they said into some text that the machine can understand. You use an AI model to generate an answer to the question.”
And then you convert the text back into speech. You feed that up and that’s how you get a chat bot. And regardless of what industry you’re in, there’s an application for that.
These functions are really independent of what industry the company’s in. And that’s why they’re universal. And we expect to see going forward more and more software companies that sell software to enterprise companies adopting AI.
Three Phases of AI
And this is the real power, because if you think about the democratization of AI, we think about it in three phases. The first phase that we’ve been through is the early adopters. Many of you are in the audience where you have expertise and a real desire to adopt this technology. Then you go through phase two, which is where more and more enterprise companies have a few data scientists. They adopt easy to use technology and they start doing AI on their own.
But the point is they still understand AI. They still know that they’re doing training and inference. And then the final phase of democratization is where as an enterprise company, you’re just using some line of business application. It might be for customer service, for marketing, it doesn’t matter. And when you’re ready to upgrade to the next version, the vendor says to you, “Oh, by the way, I have some requirements on the hardware now, because I have gone ahead and infused my application with AI.”
And so the adoption is by people who don’t even know what the AI inside looks like, and they don’t need to know, but they’ve adopted AI. So that’s really the phases we see and I think at this stage, we are somewhere on the cusp of really moving from stage one to stage two, but certainly down the road, we will get to stage three.
Now, the beautiful thing about this application of HPC AI is that fortunately in computer science terms, there’s really one recipe. And because there’s one recipe, we can do largely one body of work and then reuse it for lots of different use cases. And … it’s all based on data.
Watch other sessions from the Data, AI and Analytic Summit on demand here.
DDN is the world’s largest private data storage company and the leading provider of intelligent technology and infrastructure solutions for Enterprise At Scale, AI and analytics, HPC, government and academia customers. Through its DDN and Tintri divisions the company delivers AI, Data Management software and hardware solutions, and unified analytics frameworks to solve complex business challenges for data-intensive, global organizations.
DDN provides its enterprise customers with the most flexible, efficient and reliable data storage solutions for on-premises and multi-cloud environments at any scale. Over the last two decades, DDN has established itself as the data management provider of choice for over 11,000 enterprises, government, and public-sector customers, including many of the world’s leading financial services firms, life science organizations, manufacturing and energy companies, research facilities, and web and cloud service providers.