00:00NVIDIA says that mainstream AI will create new customers that could become a bigger source of
00:06revenue and its hyperscaler clients' robotics become more of a thing. Let's bring in T.D.
00:13Cowan, senior equity research analyst, Joshua Bookalter. He has a buy rating on the stock.
00:19Joshua, can I just get your initial reaction? Because the numbers are big, 85% revenue growth,
00:24huge amounts of buybacks, a 91 billion quarter upcoming. What were your overall thoughts?
00:30Yeah, thank you for having me on. Yeah, no notes, really. I mean, it was what's become a typical
00:36NVIDIA earnings call with a meaningful beat and raise. You get by $3 billion in the print,
00:41$4 billion in the guide. They're making the extraordinary look ordinary here. And the
00:47other angle that's typical about an NVIDIA earnings call recently is the stock didn't move on it.
00:52But I think you alluded to a very important point. So NVIDIA, for the first time,
00:57separated its data center segment into two buckets, hyperscale and then one that includes the AI
01:03startups, neoclouds, enterprise, and others. And I think there was an important point there in that
01:09only half of it is stemming from hyperscalers. And so I think they're leaning into the diversity of
01:15both their customer base and the applications, which should lead itself to more confidence in the
01:22durability of the spending as well. You know, we're going to talk to Joe Kaiser a little bit later,
01:27formerly of Mercado and Blackstone. He says, you know, AI is moving to the edge, meaning away from
01:34these, you know, big data centers in like Danny's hometown and back here towards the city, maybe even
01:41in robots or in cars that we drive. And that's why they broke apart that revenue reporting.
01:48How soon are we going to see AI that we interact with in real time, you know, in those kind
01:55of
01:55products? Yeah, I think it's a fair question. I think, you know, the robotics and more edge
02:01inference and edge inferencing and AI, I do think is not going to be material from NVIDIA standpoint for
02:06a bit, simply because, I mean, let's be real. NVIDIA's revenue is tracking towards $400 billion. And it's
02:12hard for what are inherently more singles and doubles wins to meaningfully move the needle. I think the
02:18bigger portion of that bucket is going to be how much is spending is done by the AI labs like
02:23OpenAI and
02:24Anthropic, who, by the way, might be IPO-ing in the near to medium future. And then also larger enterprise
02:32applications and customers that are increasingly going to be utilizing AI, more so than some of the
02:36robotics and edge inferences, even if, you know, the revenue numbers that are within that bucket are, you
02:42know, running errors through NVIDIA, but are very meaningful for everybody else.
02:45Joshua, at the same time, there was some time spent on this call about supply chain risks that are quite
02:51complicated. You also have what's happening in the Middle East, adding to concerns that the macro
02:56environment and demand for these chips obviously has been outweighing any of those. But at what point,
03:01if at all, do they become a real issue for NVIDIA and for the wider chip universe?
03:07Yeah, I'm less worried about the impacts of the Middle East for NVIDIA specifically. I do very
03:13much think that demand remains a much more important piece of the story here. But on the supply side,
03:18one thing that was interesting is NVIDIA's supply chain commitments went up from $95 billion last
03:24quarter to $119 billion this quarter. Now, it's really hard to make an explicit and quantified read
03:29from that number. But I do think what we should be taking away is we are in an environment where
03:35everyone is clamoring for capacity. There are concerns about supply. And NVIDIA has supply
03:41commitments locked up that that are bigger than most companies revenue. And that naturally lends
03:48itself to better visibility than others. And I do think there's an element where visibility into capacity
03:53and supply is important, not just from a pricing standpoint, but also for demonstrating to your
03:59customers that you have more visibility and ability to get supply through the network than your peers.
04:06Durability is like the key word, I guess, in a lot of the analyst reports. Is this kind of growth
04:14sustainable, maybe is a better word, Josh, in the future? I mean, will we continue to see
04:20these massive numbers or does it, you know, scale down as we get as we get further into the AI,
04:27you know,
04:27reality? Yeah, I mean, listen, like NVIDIA is going to track. Are they going to do 75, 80 percent year
04:32over year
04:33growth forever on the massive revenue base that they have? No, that's you know, there is law of large numbers
04:38that will certainly apply here. But I think from that standpoint, what's most compelling to me is that, again,
04:44we're four years into the AI, into the AI boom, and NVIDIA is the clear leader here from a market
04:51share
04:51and revenue standpoint, and their revenue growth is going to accelerate this year. Again, that can't
04:56continue forever. But the other point here is, again, there were concerns of hyperscale, the hyperscale
05:03portion of contribution to NVIDIA needing to naturally hit a wall as, you know, NVIDIA's revenue was
05:09becoming so high within the CapEx figure. We're actually learned by splitting out the segments that
05:15NVIDIA's revenue is actually closer to 20 percent of hyperscale CapEx run rate versus the 40 to 50 percent
05:21if you just took the whole data center segment revenue. And the other interesting point was their
05:25hyperscale segment grew 115 percent year over year. It's going to decelerate, but it's not really doing
05:31it right now is the reality.
Comments