A Lot Can Go Wrong for Nvidia, Says Analyst Goldberg

Bloomberg

Watch A Lot Can Go Wrong for Nvidia, Says Analyst Goldberg - Bloomberg on Dailymotion

Transcript

00:00NVIDIA on track to become the first five trillion dollar company as President Donald Trump fuels

00:05hopes of Blackwell chip sales in China. Jay Goldberg of Seaport Global holds the sole

00:11sale rating on NVIDIA with a $100 price target saying there's a lot more that can go wrong with

00:17NVIDIA than can go right. Jay joins us now for more. Jay, welcome back to the program, sir.

00:22NVIDIA 73 buys, six holds, your sale. Jay, what can go wrong? Well, I think there's a lot that can go

00:31wrong. There's just not as much that can go right as well, right? NVIDIA is the most scrutinized

00:37company on the planet. Investors track it down to every detail. Everyone looks at all the supply

00:42chain data coming out of Asia. We know what their capacity is. They say they're sold out this year.

00:47They say they're probably going to be sold out next year. Once you're sold out, how do you go up from

00:51there, right? There's some talk that maybe they got some more capacity, but there's not a lot of

00:56upside, right? And that's part of the reason China is such a focus topic is because that's sort of the

01:02one swimming factor quarter to quarter, as opposed to you start to look through the downside risks here

01:09is AI is, I maintain it's a bubble. No offense to Tony, who was just on, but I think there's a lot of,

01:18there's only really six companies driving all this spend. And you sort of look through the chain

01:23of what's going on in the data center construction, electricity, supply chain, the increasingly

01:28labored financial engineering that's going into putting these deals together. And I just think

01:34there's a lot of things that could sort of throw a wrench in the works. Jay, what if, though, the

01:40president of the United States does strike a deal on Blackwell chips? Would that change your mind

01:44about NVIDIA and the price target? No, I think that's, that's the, that's the big swing factor

01:49quarter to quarter. They, they wrote those chips off a few, what, two quarters ago. What's going to

01:56happen with their China, their China parts? I think we all anticipate there's going to be some resumption

02:01of sales into China. So I don't think that's the, yes, it's, it is some upside, but I don't think it's,

02:06it's a sort of, I don't think it changes the overall arc here. China is 20% of sales,

02:11something like that. So. You have a sell on NVIDIA and you think we're in a bubble.

02:15If you think back to the dot-com era, NVIDIA equals who?

02:21NVIDIA equals Amazon, right? Amazon went up a whole bunch and then it fell a whole bunch and

02:27then it came back. And I think that's what we're going to see here too, right? I agree. AI has the

02:31potential to be immensely powerful and NVIDIA is really well positioned for that. But I think that

02:36they're going to be growing pains. No technology this big has ever been adopted in a straight line.

02:41They're going to be hiccups. They're going to be pullbacks. I think it's just going to take us

02:44a few more years to figure out what AI is actually useful for and what it's good at and what are the

02:48underlying business models. And there are going to be some, some pretty, pretty serious swings in

02:54share prices between now and then. If it's reminiscent of Amazon, then at what level would you want to

03:00get in? Because I imagine everyone listening to this thinks, oh, Amazon, I wish I had bought that

03:04sometime in the past. Well, buying Amazon in 1999 and then holding it, it probably took another 10

03:11years for it to get back to that price. So I think there is going to be disruption in the whole AI

03:18chain. Right. And so I know that any sort of quarter to quarter dips, people will just buy

03:24into weakness. But I do think there are sort of more fundamental issues here around the adoption

03:29of AI, how long it's going to take and how some of these deals, how some of these big data center

03:34things that are baked into numbers, how do those actually get built? Right. What happens if one of

03:38these major AI labs can't build their data center in time, they can't get the electricity or they

03:47can't get the funding right? I think that slows things down. I think that, you know, and then I

03:51think there's risks in consumer adoption. We still don't have a really good consumer AI app. And I think

03:57at some point, all of this spend has to be justified. And the return on that investment is not entirely

04:06clear in a lot of these cases. Jay, could you just describe, spend some time on describing what

04:09that moment of realization would look like when the other 70 plus names get what you're looking

04:15for? What does that actually look like? Because your price target right now at 100. I just wonder

04:19what the rest of the market would look like if we came down from 200 to 100. So I probably have to

04:24review my price target. I'll give you that. But I do, like I said, I think, I think what happens is

04:29over the next year or so, gravity just reasserts itself, right? I'm increasingly concerned by some

04:37of the financing that's taking place in some of these neocloud deals, some of these data centers.

04:42I think that's sort of almost a house of cards in some cases. There are good companies there too,

04:48but there are just so many of these companies chasing this build. But when you get down to root

04:54demand, it's all coming from the same six companies, right? And I think if you push those

05:00companies on what their business case for all this massive spending on NVIDIA parts is,

05:06it's challenging for them to come up with a clear, coherent business case for it, right? And in that

05:12kind of situation where they're sort of investing heavily, spending huge amounts of money based on

05:18some form of fear of somebody else getting to some unspecified destination first, that's a very

05:23heavily psychologically driven spending wave. And that can, just as easily that came on,

05:29it can also turn off quickly too. And so I think what we'll see, probably not in this earning season,

05:34but somewhere in next year, we'll start to see some easing back, some lowering of CapEx.

05:40And that sort of causes everything to reset.

Category

Transcript

Be the first to comment

Recommended