Nvidia CEO Jensen Huang’s opinion on the hypergrowth of generative AI | interview

Connect with top gaming leaders in Los Angeles at GamesBeat Summit 2023 this May 22-23. Register here.

---

Jensen Huang will give a keynote address at the Computex 2023 event in Taiwan over the weekend, but he’s already riding high on a resurgence in AI demand thanks to the popularity of generative AI like OpenAI’s ChatGPT.

Huang reported earnings on Wednesday that beat Wall Street earnings expectations as well as expectations for the second half of the year. That, in turn, fueled a broad tech rally on Wall Street.

The graphics chip and artificial intelligence company’s share price rose 27% in the past two days, from $305 a share to $389.25 today. With a market capitalization of $963 billion, Nvidia’s valuation is approaching $1 trillion.

Building on the strength of data center chips, Nvidia reported revenue of $7.19 billion for the fiscal first quarter ended April 30, down 13% from a year ago but above expectations. Data center revenue in the fiscal first quarter was a record $4.28 billion, up 14% from a year ago and up 18% from the prior quarter.

I spoke briefly with Huang after the earnings announcement this week and he told me about his confidence in the rise of AI, as well as Nvidia’s ability to align manufacturing to meet this demand. But he didn’t think there would be a pick up in the overall economy in the second half. And he pointed out that gaming, while down from a year ago, is back to the quarterly growth of previous quarters.

Here is an edited transcript of our interview.

GamesBeat: Looks like you have a lighthearted report today.

Jensen Huang: We always have happy reports! It’s keynote time. I’ll say something funny.

GamesBeat: I was wondering how the transition would work as we go from hitting the gas, having shortages during the pandemic, hitting the brakes, and now hitting the gas again following the release of ChatGPT and the growth of generative AI. How do you think the pattern is going to be to meet these higher expectations now?

Huang: The good news is that we were already taking it down on Hopper. We went into production in August last year. Our timing was flawless. Ampere, of course, is still in high demand and already in volume. The first part is that our supply chain is very large and the flow of our supply chain is already very high. On top of that, we still have a huge increase in demand. We have purchased substantially more supply for the second half of this year. We responded very quickly and placed very large orders. We have acquired a lot of incoming supplies.

GamesBeat: Do you see something like an economic resurgence accompanying this? Does the second half of the year look good for that reason too? Or is this really just an AI phenomenon?

Huang: I think it’s an AI phenomenon. The reason for this is that the rest of the data center is still down. The data center, enterprise computing, as you know, is muted. But generative AI is doing amazingly. I think it’s very focused on generative AI. For the first time, people can see how they are going to make money with generative AI. All of these APIs can be connected to all of these services and applications. It is much easier to invest when you can see the return on investment. That is number one. Number two, people finally realized, when everything fell into place, that accelerated computing saves them money and it saves energy.

GamesBeat: Why is it so expensive to train large language models?

Huang: Well, they are not. They really aren’t.

GamesBeat: The impression I got from ChatGPT was that it was very expensive.

Huang: It doesn’t cost that much at all. The reason for this is that it just depends on where you started. If you’re a software engineer, you used to be able to write an app just by buying a Macbook. Now all of a sudden the fact that you need a supercomputer to help you develop the model seems like a lot. But let’s take it in perspective. If you were building a chip company and you were taping a chip, the tapeout for a chip is around $100 million, just the tapeout. Not to mention the tools, which probably cost another 100 million dollars, and not to mention all the engineers, all the systems you’re building, things like that. To build one of our chips, it’s a few billion dollars. And we’re just a chip company. There are a lot of chip companies. When they hit a chip, it costs no less than $25 million. Writing, building a big language model – burning a chip these days, what the software industry is learning is that building these big language models is like burning a chip.

GamesBeat: Is there a lot of concern that we could still see shortages of some kind due to this rapid change in demand? Or are you not so concerned about that?

Huang: I think the shortage of services is quite severe at the moment. But I think it’s going to improve tremendously in just a few months as all the systems are delivering in real time. This is going to improve in real time.

GamesBeat: The game seems smoother. Is it for any particular reason?

Huang: No, I thought playing was cool. We are seeing sequential growth.

GamesBeat: I guess it’s still down from a year ago.

Huang: Yeah, but that’s for all the stuff from a year ago. Yearly comparisons are difficult. But the quarterly results are good. The channel inventory correction is behind us. We are augmenting Ada now across the board. I’m pretty excited about it. There’s also a new app for creatives in town. It’s called generative AI. GeForce is for gamers and creatives. Now you have generative AI to help you create things. That’s the buzz. People are very excited about it.