Nvidia's Historic Earnings, Jensen is the King of AI
Revenue is expected to increase 52% quarter over quarter. And a writeup on other public AI companies.
Nvidia posted a historic quarter. There’s no point in focusing on this quarter but rather on the guide. The magnificent, glorious guide. The most important thing is they guided revenue to $11 billion, or 52% quarter-over-quarter growth. I don’t think anyone has ever grown revenue that fast on that business size.
Revenue $11.00B +/- 2% vs FactSet $7.17B
Moreover, they guided to record gross margins of 70%, which we have never seen in Nvidia’s history.
Non-GAAP gross margin 70.0%, +/- 50bp vs consensus 66.6%
I understand the hype, and I have been waiting for this quarter to strike it out one way or another decidedly, and now I have to say, Nvidia has outdone itself. I modeled out something that I thought would be appropriate for revenue estimates, and the shocking thing is that it seems that consensus estimates are now ~50-60% below.
The biggest “debate” in this calculation is likely my continued sequential revenue growth for the next few quarters, but I think it’s unlikely they cannot grow sequentially given the step up in demand. And remember, this is Nvidia, the company that has missed revenue estimates thrice since 2011.
Their secular growth is not to be trifled with, and I think the team at Nvidia is likely having a hard time forecasting revenue, so why not forecast a bit more conservatively? Let’s talk about some specifics about the call. Regarding specifics, this comment from Collette (CFO) was probably the best for sizing the rest of the year.
We believe that the supply that we will have for the second half of the year will be substantially larger than H1. So we are expecting not only the demand that we just saw in this last quarter, the demand that we have in Q2 for our forecast but also planning on seeing something in the second half of the year. We just have to be careful here, but we're not here to guide on the second half. But yes, we do plan a substantial increase in the second half compared to the first half.
In a perfect world, they would guide sequential revenue growth for the rest of the year. In a world where they grew 52% QoQ, and forecasting demand is almost impossible, the second half over first-half revenue is the best insight we will get.
Another part I found very compelling is Jensen’s defense of GPUs over accelerators, noting that pre and post-processing is a massive part of the inference burden, and the more generalized structure of GP-GPUs makes GPUs a great fit for inference today.
So the input and the output requires a lot of pre- and post-processing. The pre- and post-processing can't be ignored. And this is one of the things that most of the specialized chip arguments fall apart. And it's because the length -- the model itself is only, call it, 25% of the data of the overall processing of inference. The rest of it is preprocessing, post-processing, security, decoding, all kinds of things like that.
Also watch for more product announcements soon at Computex next week.
And then beyond that, if you want to connect the smart AI factory -- this AI factory into your computing fabric, we have a brand-new type of Ethernet that we'll be announcing at Computex. And so the -- this whole area of the computing fabric extending, connecting all of these GPUs and computing units together all the way through the networking, through the switches, the software stack is insanely complicated. And so we're -- I'm delighted you understand it. But we don't break it out particularly because we think of the whole thing as a computing platform as it should be.
Last, I’ll leave you on a turbo bull case that Nvidia will replace most of the annual $250 billion dollar in spending from CPUs to GPUs. Even say 30%
And what happened is when [indiscernible] today, I came along. It triggered a killer app for this computing platform that's been in preparation for some time. And so now we see ourselves in 2 simultaneous transitions: the world's $1 trillion data center is nearly populated entirely by CPUs today. And $1 trillion is $250 billion a year. It's growing, of course. But over the last 4 years, call it $1 trillion worth of infrastructure installed and it's all completely based on CPUs and dumb NICs. It's basically unaccelerated.
In the future, it's fairly clear now with this -- with generative AI becoming the primary workload of most of the world's data centers generating information, it is very clear now that -- and the fact that accelerated computing is so energy-efficient, that the budget of a data center will shift very dramatically towards accelerated computing, and you're seeing that now. We're going through that moment right now as we speak while the world's data center CapEx budget is limited. But at the same time, we're seeing incredible orders to retool the world's data centers.
And so I think you're starting -- you're seeing the beginning of, call it, a 10-year transition to basically recycle or reclaim the world's data centers and build it out as accelerated computing. You have a pretty dramatic shift in the spend of a data center from traditional computing and to accelerate computing with SmartNICs, smart switches, of course, GPUs and the workload is going to be predominantly generative AI.
This brings me to valuation. The most surprising thing about Nvidia’s earnings is that you can underwrite the numbers now. That sounds a bit nutty, but hear me out.
Assuming the operating leverage continues, they have ~55% operating profit for the rest of the year, a similar interest expense, and ~14% tax rate, which gets me to an eps number of about ~9.50 for fiscal 2024. It’s trading at ~390 after hours, or 41x CY 2024 earnings. I think it’s safe to say that they can grow EPS by 40% next year, that is ~30x earnings for FY 2025. That’s forward multiple that you can frankly feel safe about.
If we look at AMD, whose estimates are likely more correct, trading at ~38x earnings, Nvidia and AMD are trading at the same valuation multiple. I don’t think that means that AMD is a sell, but rather that Nvidia’s ceiling for valuation is higher.
I think that’s insanely compelling. ~401 dollars a share is a 1 trillion dollar valuation, which will almost certainly be broken in the coming year. This is an important psychological barrier.
Meanwhile, this is probably the most innovative and dominant company in the AI space, growing revenue at ~70% annually while trading at 30x next year’s earnings. AI is likely going to be a paradigm shifting event, and demand and revenue estimates still have upward bias. I can underwrite 30x earnings easily given the revenue growth, and I think that this thing is the easiest candidate for a bubble stock possible. Moreover, there really aren’t great publically traded alternatives.
In conclusion, despite the insane move, the math can be justified at Nvidia now. Their strong and meaningful revenue growth will convert to earnings, and despite the headline sales multiple, they are making 50%+ operating profit and converting that to earnings. 30x next year’s earnings seems palpable, and a price that I don’t think Nvidia will stay at for long.
There aren’t great alternatives (I’ll discuss behind the paywall) that benefit from things like Nvidia does. They are the purest play on generative AI and LLMs and trade at 30x forward earnings (likely conservative!). That doesn’t seem right in this environment. It’s hard not to be bullish.
I understand chasing the highest-flying semiconductor stock levered to AI might be tough, but luckily there are a few other “bets.” A few more esoteric, a few just second-level thinking. I discuss them behind the paywall.