Jensen Huang, CEO of Nvidia: AI language models as a service “potentially one of the greatest software opportunities ever created”

" image-credit="Nvidia" image-alt-text="jensen-huang-gtc-2022-crop-for-twitter" image-filename="jensen-huang-gtc-2022-crop-for-twitter.jpg" image-date-created="2022/09/20" image-width="3445" image-height="2302" image-do-not-crop="false" image-do-not-resize="false" image-watermark="false" lightbox="false" preload="true" class="c-shortcodeImage u-clearfix c-shortcodeImage-large c-shortcodeImage-hasCaption">

Nvidia co-founder and CEO Jensen Huang opened the company’s fall GTC conference by announcing the general availability next month of the company’s new “Hopper” GPU in systems from Dell and others. The keynote also featured computers for healthcare, robotics, industrial automation, and automotive, as well as several cloud services, including an Nvidia-hosted cloud service for deep learning language models such than GPT-3.


As reported yesterday, Nvidia co-founder and CEO Jensen Huang opened his company’s fall GTC conference with numerous product and service announcements, including the introduction of two cloud computing services that the company will operate.

At a press conference on Wednesday, Huang told ZDNET that the two services will be “very long-term SaaS platforms for our business.”

One service, Large Language Model Cloud Services, allows a developer to take a deep learning artificial intelligence program such as GPT-3 or Nvidia’s Megatron-Turing 530B, and tailor it to particular applications, to make it task specific while reducing the effort the customer has to put in.

The second service, Omniverse Cloud Services, is an infrastructure-as-a-service offering from Nvidia that will allow multiple parties to collaborate on 3D models and behaviors.

Also: Nvidia CEO Jensen Huang announces availability of “Hopper” GPU, a cloud service for large AI language models

ZDNET asked Huang, How big is SaaS [software-as-a-service] business for Nvidia for many years?

Huang said it’s hard to know, but the big language model service has such broad applicability that it will be one of the biggest opportunities in any software.

Here is Huang’s response in full:

Well, it’s hard to say. It really is, in a way, the answer. It depends on the software we offer as a service. Maybe another way to take it is just a couple at a time. This GTC, we announced new chips, new SDKs and new cloud services. And that’s what you’re asking. I highlighted two [cloud services]. One of them is the large language models. And if you haven’t had a chance to look at the effectiveness of large language models and its implication on AI, really do. It is really something important. Large language models are difficult to train and the applications of large language models are quite diverse. He was trained on a large amount of human knowledge. And so it has the ability to recognize patterns, but it also contains a coded amount, a large amount of coded human knowledge, so that, if you will, it has a kind of human memory, if you will. In a way, it’s encoded much of our knowledge and skills. And so if you wanted to adapt him to something that he was never trained for – for example, he was never trained to answer questions or he was never trained to summarize a story or publish breaking news, paraphrase, he was never trained to do these things – with a few extra strokes of learning, you can learn these skills. This basic idea of ​​fine-tuning, adapting to new skills, or learning zero or few shots, has big implications in a lot of areas, that’s why you see so much funding in computational biology. . Because the great language models have learned to structure the language of proteins and the language of chemistry. And so, we put this model in place. And how big can that opportunity be? My guess is that every company in every country speaking every language probably has dozens of different skills that their company could adapt to our grand language model moving forward. I’m not sure exactly how big this opportunity is, but it’s potentially one of the greatest software opportunities of all time. And the reason is that intelligence automation is one of the greatest opportunities ever created.

The other opportunity we talked about was Omniverse Cloud. And remember what the omniverse is. Omniverse has several features. The first characteristic is that it ingests, it can store, it can compose physical information, 3D information, on several layers or what are called diagrams. And it could describe geometries, textures and materials, properties like mass and weight, etc., connectivity. Who is the supplier? What is the cost ? What is it related to? What is the supply chain? I’d be surprised if – the behaviors, the cinematic behaviors. It could be artificial intelligence behaviors. And so, the first thing Omniverse does is it stores data. The second thing it does is it connects multiple agents. And agents can be people, robots, autonomous systems. And the third thing it does is it gives you a window into this new world, another way of saying, a simulation engine. And so, Omniverse is basically three things. It is a new type of storage platform, it is a new type of connection platform. and it is a new kind of computing platform. You can write an application on Omniverse. You can connect other apps through Omniverse. Like, for example, we showed many examples of Adobe connected to Autodesk apps connected to, you know, various apps. And so, we connect things, and you could connect people. You could connect worlds, you could connect robots, you could connect agents. And so the best way to think about what we did with Nucleus [Nucleus Cloud, a component of Omniverse Cloud, is a facility for developers to work on 3-D models using the Universal Scene Description specification], consider it the easiest way to monetize this, is probably like a database. And so, it’s a modern database in the cloud. Except this database is 3-D, this database connects multiple people.

And so, these are two SaaS applications that we have implemented. One is called a large language model. The other is basically Omniverse or a database engine, if you will, that we’re going to install in the cloud. So I think those two announcements – I’m really glad you asked – I’ll have plenty of opportunities to talk about them again and again and again, I’m going to talk about them again and again, but those two platforms SaaS are going to be very long-term SaaS platforms for our business, and we’ll run them across multiple clouds and so on.

Comments are closed.