'Stock is trading at the very low end of its historical range': Luria on Nvidia

Added: 2026-05-22

[00:00:00]Well, Nvidia surpassed first quarter Wall Street expectations on massive demand for its AI chips, but reaction from investors rather muted. Let's find out why investors might be so skeptical.

[00:00:11]Joining me now is Gil Luria, head of technology research at DA Davidson.

[00:00:16]Thanks as always for joining us, sir.

[00:00:18]Uh why the muted response? Let's just get that out of the way now.

[00:00:23]Yeah, that's been the pattern the last few quarters. It doesn't matter how good Nvidia results are, the stock doesn't necessarily react positively to earnings. A big reason for that is that we actually knew that Nvidia would report this type of result a month ago when Microsoft, Amazon, Google, and Meta reported. Those companies drive more than half of Nvidia's revenue, really.

[00:00:47]And so we knew that as those companies continue to spend on CapEx, continue to increase CapEx, Nvidia would do well.

[00:00:54]The stock is up since then, which tells you that that was reflected in the stock. And today they just delivered the formality, or last night they delivered the formality, but mind you, very good results, beat, raised, staggeringly high numbers.

[00:01:09]They they talked about selling 20 billion of standalone CPUs this year.

[00:01:15]That that was at zero last year, so their opportunity continues to grow.

[00:01:20]There was nothing bad about last night's result. It was all good. It's just that investors already knew that was coming.

[00:01:26]Okay, so 90 billion exceeding 90 billion for their revenue for the next quarter. Is that achievable? I mean, you just mentioned about that 20 billion, so is that achievable then?

[00:01:38]Yeah, in fact, over the last few months the company's given us much more visibility than in the past. They've talked about a trillion-dollar cumulative business for Blackwell and and Rubin, which means that they can continue to grow at these rates well into next year. And that's something that as investors get more comfortable with, they'll appreciate Nvidia more.

[00:02:00]Nvidia is already trading in the low 20s on forward earnings. It's unlikely that they'll trade at any less than that next year, and their earnings will be at least 50% higher. So, so that's why this is such a good setup for Nvidia. They have more visibility than ever, and the stock is trading at the very low end of its historical multiple range. All right. And and you mentioned Vera Rubin.

[00:02:22]Uh what makes it so unique? And is there any anybody even threatening it right now to compete with it?

[00:02:31]It's just much more performant in the sense that you can get a lot more tokens per dollar, which ultimately is what is important for for AI both mostly on the inference side is what does it cost you to to create a token to drive AI. And and their products continue to stay ahead of the competition. All All AMD and and Broadcom and the proprietary chips from the hyperscalers can do is try to keep up. And and at least for some applications have a slightly lower price point cuz Nvidia is just much much more performant with every generation of of of new semis that they introduced, new GPUs, and again new CPUs as well. So, they're staying ahead of the competition, and that's really the story of Vera Rubin is that by the time the other the other chips catch up to Blackwell, Nvidia shows up with Vera Rubin.

[00:03:25]And and with that, can AI replace these chips? Could Could it you know, could somehow Nvidia be replaced by AI?

[00:03:34]Well, I just wait. I went on a tangent there for a second.

[00:03:39]I went on a tangent for a second there.

[00:03:42]>> [laughter] >> Um Go.

[00:03:45]Yeah, Nvidia doesn't actually own its manufacturing plants.

[00:03:50]If you distill what Nvidia does, they make a very complicated design file that they send to TSMC in Taiwan to manufacture. So, could at some point AI be so good that it can design a better file and do that instead of Nvidia or for Nvidia? That's possible. But again, at that point none of us have a job anyway, so we [laughter] won't even have to talk about it.

[00:04:13]>> Well, you and I are sitting on a beach then and talking about this, hopefully.

[00:04:17]Okay, other competition, China.

[00:04:20]Do they need to crack that market or are they going to be good without it?

[00:04:24]Well, it's a quarter of the global market, so they very much would like it to be part of their market long term. In the short term, they're just caught in this back and forth between the US government and the Chinese government about whether they can sell even degraded last generation chips. And at some point the US government agrees and the Chinese don't, and then the Chinese don't agree and then the US does. So, it's been just back and forth for several months. Right now, none of us have Chinese sales in any of our Nvidia expectations, but longer term they do want to participate in the market. Again, it's quarter of the global market.

[00:05:03]And are the Chinese chips close or getting closer?

[00:05:08]They're getting closer, but it's the same thing. Nvidia runs so fast that as the Chinese catch up to the last generation, Nvidia comes up with the next generation. But they will address the domestic need in China because Nvidia has been shut out of that market and that helps them grow to scale and get better, and that's what Nvidia doesn't want. It It doesn't want this to be making Chinese semi companies better and bigger, and then using that leverage to sell outside of China. That's why Nvidia wants to continue to sell even degraded chips into China in order to prevent the Chinese industry from developing to a point where they are competitors on a global scale.

[00:05:51]All right. And and Nvidia they're trying to diversify and how important is that?

[00:05:57]Yes, Nvidia really controls the whole ecosystem. They're really the orchestrator of the great data center build out. They have had this foresight about how quickly it would come to be.

[00:06:10]So they've coordinated with TSMC in making the chips, with the memory companies, Hynix, Samsung, Micron for develop delivering enough memory and then they quietly decided that they're going to sell CPUs on a standalone basis cuz they anticipated that using AI in an agentic manner would require a lot more CPUs.

[00:06:33]They have networking equipment. They're working with the optical companies.

[00:06:38]They're the ones orchestrating the whole build out because they have all this foresight and that allows them to participate in the most significant way and be in the middle of the build out, which means in the middle of the AI trade. All right. And you still have it as a buy at $300 US.

[00:06:56]That's right.

[00:06:57]And no changes.

[00:07:00]Uh not until they get over 300 and then we'll have to reassess. All right. We'll leave it on that note there, Gil. Thanks as always for joining us.

[00:07:07]Thank you. Gil Luria is head of technology research at DA Davidson.