Transcription
Set you up to cut you off. GT, go ahead. I set you up to cut you off. GT numbers are coming in right now. Hold that thought.
30.8 billion for data center revenue beats the estimate of 29.14 billion. That's the data center revenue, not the total. Let me get you the total: 35.1. That is almost a $2 billion beat versus expectations of 33.2. Earnings, to John's point, 7% beat their $0.81 versus 74.
Seems like the market was a little bit not negative, but perhaps too cautious on the profit side. I mean, that's a pretty good beat: $0.81 versus 74. And then almost $2 billion, George, on the top line. That's pretty good. Stocks coming off a little bit touched 138 and change.
What do you think, GT?
Yeah, I mean, I'm not surprised by the reaction. The reason I say this is at least right now it's too early to tell what will happen. Of course, the open tomorrow and then the follow-through tomorrow is going to be extremely important on this name. I think the market may be adjusting accordingly considering what you have coming out with the Blackwell infrastructure. It's new; it's effectively something that they have to generate scale on.
And, you know, as you mentioned, there's some issues associated based on what they've told us in terms of the thermal displacement of that new technology. But nonetheless, I mean, these numbers are still impressive. 28.9 billion was the estimate on data centers. They came in ahead of that at 30, beating the top-line sales in total by $2 billion.
And that is, again, just the $2 billion beat is more than 12% of last year's sales at 18 billion. So effectively, you're seeing the top-line sales more than double. You saw bottom-line EPS more than double over 101%. And of course, we know that they have about 90% market share in the data center market. So they are still running extremely robust in terms of their sales, their earnings, and their profitability profile.
It's amazing if you just look at it from a hardware company standpoint. Net income margins, I think, are going to be important here at 55% over the last year, 85% gross margins. And these are typical of software companies. So I think this is still an impressive company. Maybe the market agrees, considering it's a $3.6 trillion name.
But like I said, when you bring out new technologies, the cost curve, you know, your average cost curve starts to adjust because you don't necessarily have the scale—the production scale—and new technology. So what they're doing is this smoothing out of discontinuing. Eventually, the hopper chips and then ramping up Blackwell, and that's a smooth and steady process.
But I think initially, there may be a little bit of a hang-up on adjusted earnings. More detail on that. And let me give you some more detail on that, George.
The timeline for Hopper and Blackwell: So Hopper and Blackwell are both going to be shipping in the fourth quarter of fiscal year 25. Okay? And then Blackwell production is going to ramp into fiscal 2026. Blackwell production shipments are going to begin in the fourth quarter of 2025. So they're going to overlap for a little bit, and then it's going to go fully over to Blackwell.
Basically, to your point about the valuation, that's kind of why I was asking John about any potential for any recurring element because it's like a software valuation. But to his point, I mean, these customers are so happy with what they've got and what they're doing that they're going to be in line for whatever comes next.
Oh yeah, definitely. And look, over time, and again, you just don't know what timetable we're looking at, they're going to face some margin challenges from technology hardware. They have to make up for it. And I think you bring up a great point, and it's an excellent, you know, attribute to Nvidia. It's: where's the recurring revenue? And that's going to come from software and services eventually.
So they're building out their Omniverse software to then bolt on vertically their hardware. And of course, this is the long-run strategy for Nvidia. But right now, it's just to keep running, you know, on all cylinders relative to any competition. And that's what's remarkable about this. It happens so fast, and there's so much demand. The pricing power is impeccable.
And right now, the hardware margins are, you know, almost unrealistic or unfathomable. But they're pretty incredible considering even the size of the business. But again, it's going to come down to the switchover from old hardware infrastructure to the new infrastructure.
And that's where there may be a little bit of a hang-up in terms of some earnings estimates because, again, the scalability will be lost on the hopper chip as they ramp up Blackwell production.
Well, that's the thing too: the faster they can ramp it, the better. I don't know, I mean, maybe for the stock it's nice to have demand outstripping supply. You get to kind of trade at a premium for expectations. But it is true that they are not meeting the demand right now. They have, you know, production is a big deal, and that means any headline about any, you know, hiccups or anything production-wise will be noted because they have a demand exceeding supply for several quarters, they say still.
So I don't know what if several—like at least three, maybe 4 or 5.
Yeah. And so when we talk about scale, scale occurs not from price. It occurs from unit quantity. And so again, prices have been firm. That's where their margins are really coming from because they can't keep up with demand. And so the scale will economics will come eventually from being able to mass-produce these units once they get their production capabilities up to where the market demand or equilibrium is.
But you know, overall, again, it's tough to digest this just with the numbers coming out. But I do think tomorrow's open and reaction on an intraday basis will be important because, you know, there was one quarter going back—I believe it was perhaps even last year—where we saw a huge pickup in price early in the day and it sold off. That was a short-term top in the stock.
Okay, George, the trade right now is pretty mixed. We dropped; we bounced. Right now trading at one—let's see—we're at 143, okay? So still, you know, a couple bucks off.
Pretty calm. You know, I mean, like when you've got a bar this high with the valuation, as you point out, very uncharacteristic of its sector and its group. This is a pretty nice response. You know, to ramp up directly on earnings would probably require some degree of shock beat. But this certainly seems like the type of number that continues the theme, that continues the trend, and probably the bid.
I would think.
Yeah, absolutely. Look, I mean, like I said, it's almost like you can't just compare this company to the industry group. I mean, even Carol and I talked—Caroline and I talked about this earlier on the panel this morning. You know, this is a stock that is narrowing in terms of performance. It's outperformed the rest of the semiconductor space because of the fact that it's its own industry group at the moment.
I mean, data center sales, if you just go back as a percent of total sales, were about 60% about six quarters ago. And now they're running around 88 to 89%. So that gives you an idea that there's massive concentration. They're still selling more gaming GPUs, but as a percent of total revenue, because revenues just ramped up so quickly, the data center market is where it's at for Nvidia.
Understood, George. Good analysis, good context. Like the sector comparison as well. Shares just kind of middling around a few bucks.