Barron's Streetwise - Nvidia's Make-Or-Break Earnings Report
Episode Date: February 16, 2024Will falling lead times delight or spook investors? Also, why one analyst calls Nvidia the Apple of A.I. Learn more about your ad choices. Visit megaphone.fm/adchoices...
Transcript
Discussion (0)
Calling all sellers, Salesforce is hiring account executives to join us on the cutting edge of technology.
Here, innovation isn't a buzzword. It's a way of life.
You'll be solving customer challenges faster with agents, winning with purpose, and showing the world what AI was meant to be.
Let's create the agent-first future together.
Head to salesforce.com slash careers to learn more.
What's interesting is NVIDIA is the most akin to a platform company in the space.
This whole Apple analogy for NVIDIA is sort of happening in two years versus what took 20 years for Apple.
Hello and welcome to the Barron Streetwise podcast. I'm Jack Howe,
and the voice you just heard is Ben Reitzes. He's the head of technology research at an
independent firm called Melius Research. Here's a fun fact about Ben. Two decades ago at another
firm, he upgraded Apple to buy at a split adjusted stock price of 17 cents.
This was before the iPhone.
Now he says there are key parallels between NVIDIA and Apple.
We'll talk about that and about why NVIDIA's upcoming earnings report could be tricky.
Listening in is our audio producer, Jackson.
Hi, Jackson.
Hi, Jack.
I'm going through my stack of Wall Street research, and I see an email here from a strategist
who says that we could be headed for a Goldilocks scenario of resilient growth and low inflation.
That was this past Tuesday, just before an inflation number came
in a little bit higher than expected. We're still using Goldilocks is the amazing thing to me.
I mean, if Goldilocks were real and if she were alive today, I feel like she'd be saying,
I cannot believe I never formally studied economics economics i can't believe my simple tale of judging porridge temperature came to be used as the metaphor for the precise uh
wall street conditions you need to keep the stock market going really it should be the mama bear
uh scenario because i always assumed the mama mama bear was the one in the middle in the middle
yeah the but i guess it could have been each one.
Maybe baby bear, like the medium porridge.
I feel like if you're going to take one thing from that story,
it's that bears have learned to make porridge.
In other words, bears are making fire.
I feel like it's really a cautionary tale.
Of course, inflation came down less than expected in January.
In Goldilocks terms, the porridge might be a little too warm.
And the stock market wobbled a bit this past week, but no major lasting damage so far.
I feel like if I were to grab a metaphor to describe this market, it would not be Goldilocks.
It might be Jenga, you know, the game with the wooden sticks where you got to stack them up and you got to pull them out one by one.
You know the game with the wooden sticks where you got to stack them up and you got to pull them out one by one.
And I feel like Nvidia is the one stick that you do not want to pull out of this market.
The stock has multiplied 18 times in price over the past five years.
It's up 224% in just the past year.
It has passed Amazon and passed Alphabet to become the third largest company in the S&P 500 by market value, just behind the current leader, Microsoft, and of course, Apple.
And that means it has a lot of pull on the returns for your index fund.
The S&P 500 has returned 22% over the past year.
You'd have to say that NVIDIA has been a big part of that.
Where are you going to find a company with that kind of heft that's on the verge of that kind of run?
If you look at the top 20 companies, what are we talking about here?
You're going to tell me Home Depot is about to triple?
What's going to happen to the weed whacker aisle that's going to triple that stock price?
I don't think so.
J.P. Morgan? Not even J.P. Morgan believes that J whacker aisle that's going to triple that stock price? I don't think so. JP Morgan?
Not even JP Morgan believes that JP Morgan is about to triple.
So I think you need NVIDIA to keep doing well.
And there's a big earnings report coming up the week ahead.
And NVIDIA in its past three quarterly reports has exceeded revenue expectations by double digit percentages each time.
So now you get into a game of analysts scrambling to take up their estimates.
The consensus on Wall Street is $20.3 billion for revenue in the fourth quarter.
That's mostly data center revenue, which at this point dwarfs gaming revenue.
But just this past week, UBS, their analyst took his estimate for revenue from $20.8 billion
up to $23 billion. That's a big jump.
It gets me thinking about whisper numbers and just how good this report has to be to satisfy
investors' expectations. What's a whisper number? Whisper number? Yeah. That's like a,
whisper number is like you have a Wall Street consensus where analysts say this is what the
earnings are going to be or the revenues. And then whisper number is people say,
no, no, those guys are low-balling it.
It's not really going to be there.
It's going to be way above that.
It's going to be this other number up here.
That's a whisper number.
Okay.
I believe that might qualify as,
what do you call that thing that weird people do
when they whisper into microphones?
Oh, ASMR.
Yeah.
I didn't expect to do that this episode, but there you have it.
I was talking about this recently with my colleague, Jeremy Owens.
He's the tech editor over at MarketWatch,
and he has a new podcast that you should all check out.
It's called On Watch by MarketWatch.
We did our interview about nerd nerds gummy clusters this morning those these are gummy candies that are covered in crushed nerds is that regular nerds not even crushed
nerds are small enough that they just get rolled around in nerds tiny nerds tiny nerds and and they
look like little coral rocks they're so crazy looking like i'll show it to you. That looks like cat litter that is clustered around.
Let me interrupt, if I may.
Jackson, could we fast forward through the nerd talk and get to the part about NVIDIA earnings?
You sure?
What does NVIDIA need to show investors when it reports earnings? This is no longer the kind of company where people are excited about the stock just by the open-ended kind of vague possibility.
It's a company that has to hit some numbers now, is it not?
What do you think that their investors want to see?
Yeah, and it's not only been hitting those numbers, it's been surpassing those numbers in gargantuan ways.
And really, my question is, when do the estimates catch up with what NVIDIA is actually doing?
Because the last two quarters, especially, they've given an outlandish forecast.
Analysts have increased their forecast commensurate with that.
And then they've just blown past it in both of those quarters.
And so now it's going to get to the point where Wall Street is putting, increasing its estimates, but still expects NVIDIA to blow them out of the
water. And eventually that's not going to happen. Eventually we're going to find that level, Jack,
where, you know, the, the expectations from Wall Street and the ability of NVIDIA hit the same
level right now we have it. And that's what I'm really looking for is when do
the expectations rise to the level that NVIDIA actually hits and they stop just blowing away
the expectations every quarter. We'll come back to Jeremy in a bit. I wanted to hear from an
NVIDIA bull about whether they're feeling at all nervous about the company's ability to continue
to exceed expectations. And I reached out to Ben Reitzis at Milius Research. Ben keeps a framed
copy of that call that I mentioned earlier, upgrading Apple stock years before the iPhone.
Over the holidays, I just like got to play with an iP. And I was like, oh, man, this is amazing.
Played with some Macs. And the funny part was they had this reliance on, I think it was IBM, for chips.
And there was really a lot of problems with the yields.
So nobody remembers this, but they missed and guided down because they couldn't get enough chips for this PowerMax cycle.
So the stocks like down huge that day.
And I upgraded it intraday and it was like the most fun thing.
And people are like, you're crazy.
This is Apple.
I mean, you should have seen what it was like.
It was like I felt like there was a hit squad.
There are no hit squads coming for Invidiables.
92% of analysts who cover the stock say to
buy it. So why shouldn't investors worry that that in a contrarian way is a worrisome sign?
Ben says that he hasn't been this excited to cover technology since the 1990s and that he sees key
parallels between Nvidia and Apple. Let me run through a few of those.
The first parallel has to do with the size of the market turning out to be larger than expected.
Here's Ben on what NVIDIA founder Jensen Wong envisioned when he started the company.
So Jensen bet his company on accelerated computing in the 90s. And his first application,
you know, really his vision was video games,
that everybody was going to become a gamer. But, you know, at the time, computing was in
a bit of a different era. It was being led by Moore's law. And Jensen had this belief that
all chips need to be accelerated. And he bet his company on it. There were some setbacks in this vision.
You know, it took a while to materialize.
But in hindsight, he skated to exactly where the AI puck was going, where Moore's law would run out of gas eventually, you know, in recent years.
in recent years. And technologically, it's getting harder and harder to shrink the dyes on chips to improve CPU performance. And then what's happened with computing is AI takes so much computing power
that in order to perform the tasks of AI in terms of training and then inferencing, you do need an accelerator.
And so they skated to where the puck was going in that regard and make the highest quality
process accelerators in the market.
Okay, so think about this.
Apple comes out with the iPhone, which by now has replaced cameras and camcorders and
answering machines and fax machines
and personal data assistants and all kinds of things that us old folks remember. And the iPhone
made smartphones ubiquitous. You can hardly find a flip phone. The market turned out to be much
bigger than anyone expected at first. Now, NVIDIA makes chips that accelerate servers, and you need that for artificial intelligence applications.
And those applications are steadily creeping into software and services.
One reason is that one of the things that AI is good at is chomping through vast amounts of data to figure out new things that AI can do.
that AI can do. So the steady spread of AI raises the question of whether all servers,
or nearly all of them, will need to be accelerated in the future. As Ben says,
an unaccelerated server will be the flip phone of a data center within 10 years.
You can see this rising expectation of how big AI acceleration can become by looking at NVIDIA rival AMD. In December, AMD raised its estimate of its TAM, or total addressable market, for both
AI accelerators and memory to $400 billion by 2027. Its last estimate, six months prior, was just $150 billion. The new estimate is enough
money for both NVIDIA and AMD to thrive. The second parallel between NVIDIA and Apple has
to do with software. Back to Ben and his discussion of NVIDIA founder Jensen Wang.
But perhaps what was more clairvoyant was that he created a programming language called CUDA several years ago.
There were developers already that existed for it. And the magic also was that he created a
full stack approach to tackling the problem. And what he realized was NVIDIA needed to be,
you know, as much as a software company as they are a chip company.
Ben mentioned CUDA.
That is not short for Barracuda.
That would be much cooler than its actual origin story.
It is an acronym that stands for, Jackson?
Compressed.
Yes.
Uber.
Yes. Dextrous. Yes. uh compressed yes uber yes dexterous yes you're nailing this one you got it compute unified
device architecture now i don't know a ton about unified device architectures or architectures or
devices but i can tell you that cuda is software that optimizes NVIDIA chips and allows its customers to build applications. And it has model libraries and a network of experts and
things that are not easily replicated. There are a lot of iPhone customers like me that are in it
just as much for the ecosystem and the software as for the latest hardware specs. And the same
is sort of true of
NVIDIA. You might recall that more than a decade ago, internet pioneer Mark Andreessen wrote that
essay where he said that software is eating the world. So if you're wondering how software is
eating the world when a chip company seems to be eating the S&P 500, it's because NVIDIA,
like Apple, is kind of a software company in disguise.
That's a good thing.
The third thing is closely related to CUDA, and that's that NVIDIA stands to earn a rising amount of high margin revenue from services, the same way Apple does with its app store
purchases and its subscriptions.
And as you heard Ben say at the top of the episode, for Apple, that played out over decades.
For NVIDIA, the process is accelerated
to just a couple of years. And that is a good place for a break. Wouldn't you say, Jackson,
calculating according to Moore's law, according to Smore's law, wouldn't you say?
Yeah, I'll get the marshmallows.
We'll be right back with a look at NVIDIA earnings.
we'll be right back with a look at nvidia earnings welcome back nvidia's quarterly financial results are right around the corner and my colleague
jeremy owens from market watch was talking about some of the investor anxieties around the company
people see nvidia booming they think, this is proof that generative AI
is going to be this big boom that tech is talking about. But are we going to have that? Is this a
short-term gain for NVIDIA to help people develop these generative AI products? Will those generative
AI products make enough money for it to be self-sustaining for this to cycle over and over
again? Or is it one cycle? Jackson, I think the concern that Jeremy's talking about is one that a lot of investors
share. What if this is a hype cycle? What if the promise of AI is overhyped at this point?
What do you think? Yeah, I mean, I was pretty shocked when you had mentioned NVIDIA's stock
had overtaken Amazon. As a consumer, I'm just so familiar with Amazon's products and their data centers support like
the entire internet.
And I don't think I could tell you if I've ever seen an NVIDIA chip or gone on their
website or anything.
So yeah, I mean, on the one hand, it's hard to discount the amount of actual money being
made right here.
So when NVIDIA reports quarterly results for the fourth quarter, the expectation is for
its most recent fiscal year, it will report over $45 billion in data center revenue.
And that's a tripling from the year before.
That's also just the estimate.
And remember that NVIDIA has a record of shattering expectations when it reports revenues.
So that's a lot of money.
On the other hand, I look at my email here and there's a chart from our friend Torsten
Slok.
He's the chief economist over at Apollo Global Management.
And it's labeled the rise and fall of AI.
And I'm thinking, fall of AI?
When did that happen?
the rise and fall of AI.
And I'm thinking fall of AI, when did that happen?
So he's made a chart of the number of times the words AI and machine learning and generative AI
are mentioned in earnings calls.
And it peaks in the third quarter of last year
and it's been on the decline.
So maybe we're talking about it less,
but there's still an awful lot of money flowing in.
And I guess that's what matters most.
So are we on the path for
another uh eighteen thousand percent gain from here i'm not i'm not that bullish fifteen thousand
baby no i that seems implausible but uh ben over at melius research says that nvidia is entering
its 30 slash 30 season i didn't know what that meant at first.
I thought that was a, sometimes analysts use like sports metaphors.
Some analysts, not Ben.
I have seen cases where analysts get a little carried away with the sports metaphors.
The ESPN documentary series, yeah.
The ESPN has a documentary series called 30 for 30, but that's not what it refers to in
this case.
has a documentary series called 30 for 30, but that's not what it refers to in this case.
He's referring to NVIDIA being on the way to earning $30 a share and for its stock being worth a multiple of 30 times that figure.
So his price target works out to just a little bit more than 30 times 30.
It's $920, and that implies 24% upside from NVIDIA's
recent price. And $30 in earnings is pretty significant for a company that earned $3 and
change in its fiscal year ended January 2023. So the estimate for this latest fiscal year is over $12 per share, and NVIDIA's on a path to
somewhere over the next few years ramping up to $30. That trajectory speaks to how Wall Street
analysts don't really view this as a sudden windfall. This is not a momentary period where
NVIDIA's selling tons of chips, and then the market's going to cool off. They see this as
something where the company will sustain its growth, not nearly at the same pace, but healthy growth anyhow in the years ahead,
much like Apple did at each point that investors said the company is too big to grow.
I asked Ben what to expect for NVIDIA's upcoming earnings report,
and he mentioned a delicate balance, and he said the company would have to thread a needle.
Jackson, when is the last time you threaded a needle?
How long has it been?
Let me guess, you've never done it, because people of your generation buy and wear throwaway pants.
You've never had a knee patch in your life, gosh darn it.
Admit it.
Hey, I do not have throwaway pants okay i've threaded
a needle i actually recently uh repaired a hole in a t-shirt you know what i should have never
doubted you you always come through it's unraveling now so i may have to uh well that's besides the
point so threading a needle is uh you know i don't know what your experience is like, but I can vaguely recall it's not easy.
You got to do some twisty stuff.
There might be, I don't want to get gross here, but there might be saliva involved.
You hope there's not bloodshed.
So let's hope that NVIDIA does better in its earnings report.
Here's Ben on what to expect.
Investors will want the increased supply so they get the increased earnings so we get
positive revisions but yet we don't want too much supply so that we have a glut and then have
an inventory correction at some point in the future so you have to thread that needle in
terms of talking about demand drivers and when lead go down, we still have to be confident
that we're in a period of sustainable growth. Ben mentions lead times. Those are really
important for a company like NVIDIA. Lead time is a delay in customers getting their stuff,
basically. And they're tricky because lead times can be a sign of good news. There could be
ravenous demand for a company's products. They could also be bad news.
There could be manufacturing delays, and those can hold down revenues.
And in the case of NVIDIA, we've seen a bit of both.
There has been a dire shortage at Taiwan Semiconductor and other manufacturers
of machines for a type of circuit packaging.
It's called COAS, chip on wafer on substrate jackson explain to the folks what
that means yeah it's like a technical version of a doritos locos taco exactly thank you let me just
point out that um there are a lot of podcasts out there that'll give you a deep dive on tech
this is more of a snorkeling. We'll get to it in another episode. But the upshot
is that now there's more supply that has come online, and this bottleneck has been relieved
a little bit. And so NVIDIA is expected to report that its lead times have come down maybe a lot.
And that would be good news because the company might report some
extra revenue upside, but you just have to satisfy any concern on the part of investors that it is
the bad kind of lead times falling. In other words, you have to show them that it has nothing to do
with demand cooling for your products. One way to do that is with a blowout revenue number. Another way to do that is with strong guidance.
UBS predicts guidance from NVIDIA of $25 billion to $26 billion for its first quarter.
Remember, that compares with UBS's own estimate of $23 billion for the fourth quarter.
So that's what Ben means when he says that NVIDIA will have to thread a needle.
It will have to show investors that supplies are improving, but they're still very much
imbalanced with demand.
And if NVIDIA can do that, and if it can continue with what Ben calls this full stack approach
of being basically a one-stop shop for artificial intelligence, then perhaps it can continue
keeping up with even exceeding high expectations.
Here's Ben.
So there's two things.
So we need to make sure there's not a massive inventory correction somewhat in the future,
and no one will ever know. There's speculation on both sides. But I think you got to watch this
software and cloud data points, because at some point, there's going to need to be a handoff in the story to more of the
platform essence of the company. Thank you, Ben. And thank you, Jeremy. Don't forget to listen to
On Watch by Market Watch wherever you get your podcasts. If you have questions and if they're
even distantly related to the subject of finance, go ahead and send them in. Just tape them on your
phone. Use the voice memo app.
You can send it to jack.how.
That's H-O-U-G-H at barons.com.
Thank you for listening.
Jackson Wafer on a substrate
on a Doritos Locos Cantrell
is our producer.
You can subscribe to the podcast
on Apple Podcasts, Spotify,
wherever you listen.
If you listen on Apple,
you can write us a review.
If you want to find out
about new stories and new, you know what, you know how the internet works.
Thank you. And we'll see you next week.