February 14, 2025 • 53 mins
Join David Hanson and Eric Bleeker as they discuss Europe's $100-Billion jump in AI and earnings from several of the companies Eric has backed in his $500,000 AI portfolio. Additionally, Eric shares the details behind an exciting new company he's buying in the portfolio -- a $11 billion market cap company that he calls the "best pure-play investment into the growth of coding assistant and the AI revolution in software development."
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
On today's episode of the AI Investor Podcast from
24-7 Wall Street, we're checkingin on earnings from tons of
companies inside of the $500,000AI portfolio and adding a new
one to the mix.
All of that and more is nextWelcome back to the show.
I am David Hanson, joined byEric Bleeker.
Eric, it's been two weeks sinceour last episode, which, in AI
(00:24):
dog years, it feels like it'sbeen about 13 years.
There's so much we've missed.
I'm happy to be back.
We did a couple of weeks wherewe had it weekly.
Then we went to buy it weekly.
We were traveling last week sowe didn't have a chance to get
the show up.
So apologies if anyone wasexpecting a show last week, but
we're here now, excited to jumpin.
We've had some time to digestthe big DeepSeek story which we
(00:47):
talked about last episode.
We've had a few weeks to digestit.
Hear from some other companies,so we're going to get into some
earnings.
But I wanted to start with justa couple other news stories.
Deepseek obviously dominatedthe headlines a couple weeks ago
and everyone was wondering well, does that mean all this
capital spending is going tolighten up?
We haven't necessarily seenthat a lot, and one of the
stories that caught my eye wasyou know, we had the big
Stargate announcement, thenDeepSeek and now Europe is
(01:07):
getting.
Remember Europe, that continentwhich at one point dominated
the world, which had a rough run.
So Europe is getting in the mix.
So walk us through.
What is the AI announcementcoming out of Europe this week?
Speaker 2 (01:16):
Yeah, you know you talk about how long it's been.
I feel like that Titanic memewhere it's like it's been 84
years.
Uh, yeah, and, as you mentioned, we were traveling.
I got this thing called thevegas flu.
That's, that's not a flu.
It doesn't show up on tests orcovid tests.
But you know, everyone outthere be be be careful, there is
(01:38):
something weird going around.
It has laid laid me up.
I'm glad I can even talk.
So, yeah, you're up just aboutout of every single conversation
in technology for about thepast decade, aside from seems
like passing some sort of newregulation with relatively
sweeping consequences.
But, like you said, the USbegins kind of the narrative in
(02:00):
2024 with the Stargateannouncement and before that the
CHIPS Act.
In 2024 with the Stargateannouncement and before that the
CHIPS Act, we've got DeepSeekcoming out of China and, as we
noted in the last podcast,deepseek was paired with a very
large funding bill.
I believe it's like $137billion worth of funding for AI
from China.
And now we have Europeannouncing it's France, if we
(02:24):
want to be specific a $113billion plan.
It was kind of funny.
The prime minister, thepresident of France, macron, he
had I'm sorry I'm saying thatwrong as well.
He had done his version ofdrill baby drill with plug baby
plug, promoting their nuclearpower.
(02:46):
So you know, david, I do thinkthat this shows, you know, as we
head into 2025 and we're goingto be continuing to follow this
kind of storyline and givingeveryone this kind of front row
seat to this storyline of therise of AI, geopolitics go front
and center.
(03:07):
We've seen Europe, which hasbeen hesitant, now a major
country on the continent,announcing a very sweeping and
ambitious plan, and I'm surethis isn't going to be the last
area.
We see Areas like the MiddleEast, even if they're not
getting the headlines, they'rein the background, providing
funding to areas like Stargate.
We have, you know, japan doinga lot of funding for their
(03:31):
semiconductor industry andtrying to, you know, bring
funding back there, and alsowe'll talk more about this later
, I'm sure but we also now haveIntel rising this week on kind
of speculation that the nextmajor push from the Trump
administration is going to begiving them a lifeline.
So this, this stuff reallymatters, and I think it shows
(04:00):
that AI is going to move from abig investing story to the pre.
You know, companies in the AIspace.
You're going to need to payattention to developments
because, as AI's impacts ramp,how to make sure that its
benefits accrue across societybecomes the number one issue
facing the world, and it's goingto be a mad scramble to figure
(04:21):
out how to do that.
Speaker 1 (04:23):
Now, this is a recently announced plan and,
just like Stargate, we don'thave all of the details of what
exactly they're doing.
But is it in the same realm ofa Stargate, of building out
larger data centers, the typesof companies that would benefit
from a Stargate type project?
Is this the same type ofinvestment that France is
talking about here?
Speaker 2 (04:38):
Yeah, I think people have called it Le Stargate, so
it's smaller and absolutefigures.
But yeah, it's similar.
It's trying to get companies tobuild large data centers,
potentially using France'savailable nuclear power.
You do look at how geopoliticshas gone.
A big export market of France'snuclear power is Germany, which
(05:01):
turned off its nuclear power,and now France is offering this
for AI.
So you see some of the dominoesfalling, but smart move for
France, in my opinion, becausethey have a great asset in their
amount of nuclear power andthey're going to make it kind of
the European hub for AI.
Speaker 1 (05:18):
Well, as an investor who has unfortunately had maybe
too much of his portfolio tiedto Europe over the last couple
of years I'm excited to at leasthave maybe some I don't know if
it's good news, but just someexcitement or growth prospects
for some of our European friendsover there.
So that's exciting.
Again, we talked about DeepSeeka lot last episode.
(05:38):
We've had a few weeks to digestit and I wanted to get your
take on some of the companiesbuilding models here in the US.
So that's OpenAI, anthropicGrok what have we been hearing
from those companies?
Have they changed their tune atall?
Again, the big discrepancy wasokay, deepseek does this for
less and OpenAI Sam Altman hasbeen out there for over a year
(05:59):
now saying I need trillions ofdollars Podcast bro, if we
recall of him going aroundtrying to drum up all this
enthusiasm for more spend.
Stargate have they changedtheir tune at all since the in
these last two weeks, or whathave we?
What have you heard from thosecompanies?
Speaker 2 (06:13):
Yeah, it's interesting.
You're seeing kind of abifurcation of kind of companies
talking their book, in a way,where out of China you've got
Alibaba saying the Americancompanies are doing all wrong
and that's a great narrative forthem because it works in their
favor and they're up 40% inrecent weeks as people kind of
look to stocks that have franklybeen stuck in the mud for years
(06:34):
and now have kind of catalystto get out of it.
But you look on the Americanside and it's kind of business
as usual.
In fact I would say it'sdoubling down on kind of the
storylines that we've had.
You know, david, the key ideawe've been following since
(06:58):
December has been thisbreakthrough in these new types
of AI models that use reasoningkind of a bull on AI and believe
that we're going to seecontinuing acceleration this
year, which I am A big reason iswe are still so early onto this
path of harnessing thesereasoning models and seeing
their capabilities skyrocket.
So, from OpenAI, sam Altman,who's the CEO of the company he
(07:22):
dropped a big blog post and heemphasized that intelligence of
a model it's still CEO of thecompany.
He dropped a big blog post andhe emphasized that intelligence
of a model it's stillproportional to the resources
used to train it.
If that's true, it's great newsfor all the data center
companies that we've talkedabout.
And he emphasized which wetalked about in our DeepSeek
episode that the cost of AIdropping by phenomenal levels,
(07:44):
which is what caused the DeepSesea panic is nothing new.
He had quoted their GPT modelsbetween 2023, mid 2024, 18
months, 150 fold drop in costperformance.
And he noted Moore's law, whichwas the fundamental centerpiece
of technology, being kind ofthe most transformative force on
(08:05):
earth.
That was 2x every 18 months.
So 150x versus 2x is striking,and he also emphasized that they
see no reason for investment tonot continue growing at an
exponential rate.
In interviews he was alreadytalking about you know, hey, we
just did Stargate with 500billion, the next one's 5
trillion.
Now again talking his book,right, talking his book.
(08:28):
But you know nothing.
Nothing has changed materiallyand in fact, with open AI, they
released a product named deepresearch.
David, have you played aroundwith deep research?
So it's on.
It's still under $200 tier.
They are going to move it to befree pretty soon.
So they've said I got a chanceto play around with it.
Really impressive Did someindustry reports in some areas
(08:54):
that we were looking at for ourbusiness.
It used novel ways of findingdata, it put it into tables,
making kind of unusualconnections that if I had
received this report and it wascoming from some management
consultant charging $50,000 forit, it wasn't far off from the
quality that you would expect.
(09:15):
So a lot of people have beencalling deep research kind of a
little bit of an iPhone momentfor the capabilities of AI, and
we're still near the front of it.
The other thing I would mentiontoo, the other companies.
We're about to see a lot ofleapfrogging because Anthropic.
They are presumably readyingtheir next model.
They came out with a statementkind of directed towards
(09:38):
politicians around the world,saying that by 2026 or 27, no
later than 2030, we're going tohave basically what they call a
nation state of intelligentbeings, which is AI agents.
And Elon Musk spoke at aconference I believe it was in
Dubai where he said Grok 3 willbe released before the end of
the month and that it, in theirtesting, outperforms all other
(10:01):
models.
So, david, we'll get to thehyperscalers the people actually
spending the money later inthis podcast, but from these
companies building the models,there's been almost for the size
of the deep seek reaction.
You have expected, maybe somekind of change their tune, but
in fact, if anything you knowthey are like I said earlier,
(10:23):
they are doubling down and Ithink what we're going to see in
the months ahead is kind of afurious pace of releases.
That, of course, will be funfor us to follow on this podcast
, right.
Speaker 1 (10:34):
Yeah, unfortunately, you know, openai, anthropic,
grok, those are.
They're not public companiesright now, so we can kind of
only assume how the market wouldmaybe be reacting.
We do have, obviously, meta,who's not a direct competitor to
those but is obviously buildingmodels as well.
We're going to get to Meta in alittle bit but, agreed, it does
seem like the vibes for lack ofa better term have got a little
(10:57):
more positive over the lastcouple of weeks, kind of what we
said in the DeepSeek episode,where there is something here,
but the overall story is notreally changing and in a way, as
long as it opens up more usecases, more use of AI, it could
be more bullish for thesecompanies, despite spending
billions and billions of dollarsBefore getting to hyperscalers
and companies that have madetons of money.
(11:18):
I wanted to touch briefly on acompany that loses money and
that's Intel, a company we'vetalked a little bit about.
You said I don't know what thefuture here is.
Is this company to break up?
Is it going to need a bailout?
What's going to happen here?
The stock is up around 20% thelast couple of weeks.
Are we getting some sort ofoutcome here or do you have any
speculation on what might begoing on with Intel.
(11:38):
This isn't a company in theportfolio.
I think you said you don'treally have an interest in
bringing a portfolio, but onethat we've just talked a little
bit about.
Speaker 2 (11:45):
Yeah, you know, the thing with Intel is it's it's
all speculation right now.
So you know, this is this is aat the moment, it's a gambler
stock Right and the thespeculation of what's happening
right now.
Donald Trump he has floated outtariffs and you know, it
(12:05):
depends how you want to look atTrump, but the kind of, I'll say
, industry consensus is thesetariff ideas for Taiwan.
It is largely the beginning ofa negotiating tactic to get some
concessions from Taiwansemiconductor.
So the reports are that there'sthree options on the table.
Number one build morefacilities in the US.
(12:27):
They built one in Arizona.
They have issued statementssaying that they are happy with
its performance.
Taiwan Semiconductor I've saidin the past, David, they are one
of the most relentlessnegotiators in the world.
You're going to have themagainst Trump.
It might make our heads explodein two sides that just want to
negotiate, but you saw, withTaiwan Semiconductor, when they
(12:49):
were building their fab inArizona, everything was US
workers are lazy.
This is not great.
Well, they're, you know, againlooking for more subsidies, more
concessions in the moment.
And it's done, it's oh yep,we're performing great here and
the PR shifts.
So you know there is a questionof whether or not, we could get
more facilities, and ones thatproduce kind of the most leading
(13:12):
edge technologies, which theyhave said they're open to.
So it's building morefacilities in the US.
There's another plan on thetable them forming a joint
venture with Intel, and thiswould effectively force some
technology transfer to thecompany.
And then there's a third optionthat I've seen reported on,
where they could take ordersfrom companies like Apple, send
(13:33):
them back to the US forpackaging by Intel.
As we've talked about a lot ofthe value kind of add with where
semiconductors are going.
It's moving from actually thesteps like ASML's lithography to
how you can package these chipsup, so that would give US more
expertise in the area thatTaiwan Semiconductor leads in.
(13:55):
So, David, I don't know wherethis is going to end up, but
again it kind of just circleswhat we talked at the beginning,
that chips and the AI race.
It's moving from somethingthat's probably a niche interest
for investors into the front ofgeopolitics, and for a lot of
these companies it's going to beinescapable that the short-term
(14:16):
movements are really going tobe controlled by what's in the
headlines around, what worldleaders are looking to do.
I think at the end of the day,though, that will mostly be
positive, as we're going to seea lot more countries wanting to
put their investment and securethis kind of as a strategic area
.
So overall bullish.
But for any specific news, thisspecific news is slightly
(14:41):
negative for TaiwanSemiconductor, slightly or
potentially very positive forIntel.
Speaker 1 (14:47):
Okay, all right, let's move on to the
hyperscalers.
We've had some earnings, sothat means we've got some
commentary from some of thesecompanies.
In particular, let's jump intoMeta and Amazon.
These are two companies thatyou recently added to the
portfolio in a big way, putting$100,000 behind each.
These are the largest positionsin our portfolio that weren't
there before.
Last episode, you talked aboutsome of the rationale behind
(15:10):
adding a company like Meta andAmazon, again benefiting from if
they have to spend a lot, theycan win, or if the cost profile
comes down, they have enormousdistribution, so you felt
comfortable using them as ahedge to the rest of the
portfolio.
So what did you hear from thesecompanies?
Are you more excited?
Are you regretting putting$100,000 on any of these
(15:33):
companies based on what youheard?
So what are your thoughts?
Speaker 2 (15:35):
post earnings David, I know it's been 84 years since
we last talked, but coming outof DeepSeek, the big question
was on everyone's mind are wegoing to see hyperscalers adjust
what their planned spend levelis for 2025?
We're heading into the year andthat means they're issuing,
(15:55):
basically, growth projections.
They're giving guidance for2025.
So this was a very pivotalearning season for seeing what
this AI story looks like.
Now, if you go back to theNovember 8th podcast for the
aficionados that have catalogedall the different AI investor
podcasts, we did walk throughexpectations for every single
(16:16):
hyperscaler and what their spendper Wall Street expectations
was expected to be this year,and that was Amazon $85 billion.
Microsoft $84 billion.
Google $57 billion.
Meta $48 billion.
So we've gotten reports frommost of those companies now.
From Amazon, they gave a numberthat isn't implied about $105
(16:38):
billion, so that's 24% aboveexpectations from just about
three months ago.
We'll call it Google $75billion in planned capital
expenditures this year.
That's 32% above expectations aquarter ago.
Meta $60 to $65 billion that's30%.
(17:00):
So, david, here's the bigpicture.
As a group, this collection ofcompanies is now expected to
spend more than $320 billionthis year.
That's above what the 2026expectations were just months
ago.
So we'll talk later on howmaybe some of this bullishness
(17:20):
that you would expect to reallyhave risen in a lot of the data
center stocks hasn'tmaterialized and maybe some of
the potential reasons for thatof the data center stocks hasn't
materialized and maybe some ofthe potential reasons for that.
But as far as what this lookslike for the continuing
investment across AI datacenters, this is pretty bullish
stuff.
Now you mentioned I'm adding$100,000 in Meta and Amazon and
(17:45):
we don't really get to dive intothe details on that last
episode.
So I want to take just a momentto say why these two companies
in particular are the ones Iwould pick from this larger
group.
If you're picking from kind ofthe magnificent seven, which
ones do you want to allocatemore into?
For Meta, number one, thevaluation is becoming less
(18:05):
favorable.
We were talking earlier.
I believe they are now slatedtoday, friday February 14th, to
be up for the 20th consecutiveday, which is a new S&P 500
record.
Clearly the story is out.
They're up now.
I believe it's 22 or 23% yearto date.
So you know it's becoming lessand less attractive.
(18:27):
But there's a reason for thisstory coming out and the bottom
line is no one's made more moneyfrom AI, as meta AI has made
their products and advertisingsignificantly better.
The estimates I read itcontributed close to two thirds
of the revenue growth last year,and that's from content
(18:49):
recommendations, more time onapps.
It's from ad targeting.
It's also from AI optimized ads.
One thing generative AI andkind of transformer models is
really good at is understandingthe context of why a specific ad
works, which in the past, withold recommendation engines, you
didn't have, and it's a realgame changer for efficiency.
(19:12):
Moving on to the next generationof AI, meta, as you mentioned
earlier, they've got a leadingopen source model.
They are in the game, at thecutting edge.
They are able to kind ofdetermine their own future.
They have a natural entry pointfor agents and the rise of
agents via messenger and David.
They've got a sea ofproprietary data.
(19:33):
What are the most powerfulassets in the world Once AI
rises, if it reaches thepotential?
We think?
Well, it's unique data and it'sdistribution.
What company has both betterthan AI?
What?
What company has both betterthan ai?
Um, if you know, if ai enableslimitless content, you, you want
(19:54):
to be the central distributionpoint that people are going to
share their creations, andthat's exactly what meta is.
So you know, david, I mentionedthis is a portfolio hedge in
part two that if spending goesdown, um, it only it, meta.
The reason that they wouldprobably drop the rest of the
year is because they get tooaggressive in their spending.
That would be the number onethreat to the stock.
(20:15):
So I am keeping it as a bit ofa hedge and, as I know, I'll
invest 100,000, but I willdecrease these positions as well
, as we're probably makinginvestments, because we are
getting pretty close, aftermaking these large investments,
to being fully allocated on the$500,000.
I don't know if you've got anymore follow-ups on Meta or if
(20:36):
you just want to move on toAmazon.
Speaker 1 (20:38):
Let's jump into Amazon real quick, and then we
can kind of put a bow on both ofthem.
Speaker 2 (20:42):
Yeah, and Amazon.
You look at their upcomingearnings expectations and it's
just very smooth growth.
It's $68 billion this year, $82the next, $104, the year after
that, $128 after.
And the reason is theiroperating efficiency has been
outstanding.
David, I pulled the numbers.
They basically have two profitengines right now their
(21:05):
advertising business and AWS.
So if you take out ads and youtake out their operating profits
from AWS in 2021, they madenegative 18 billion.
In 2022, negative 48 billion.
That's when the stock hit itsnadir.
Since then they've kind ofturned the course around 2023,
(21:26):
negative 34.7, 2024.
This past year, about negative30.
So they're getting a lot moreefficient.
So I think one thing too isI've talked I'd like to add more
robotic stocks to the portfolio.
It's just when that market'sgone.
You can't force it and investtoo early if there's not
(21:47):
investable ideas.
I think Amazon has more to gainthan any big tech company from
robotics and the efficiency thatthat's going to lead to.
It already has led toefficiency, but the kind of
inflection point they're goingto see is going to be incredible
in terms of drivingprofitability and making
e-commerce as big a profitengine in some ways as there are
(22:08):
other business units.
Speaker 1 (22:09):
So you're saying that 30 billion loss.
You're basically taking out AWS, taking out advertisers, You're
looking mostly at the retailbusiness.
So you're saying you can takethat 30 down to 15.
That's essentially like addinga $15 billion profit business.
Speaker 2 (22:24):
Yeah, yeah, I mean, and it is just Amazon there they
have.
You know, it's when you look atthese companies and you go, how
do you get bigger?
You're a $2.4 trillion company,which is what Amazon is, and
it's like, yeah, but if youactually have a really good path
to grow, sales maybe only eightto 10%, but profits 20% plus,
(22:45):
you can continue, uh, you cancontinue being a great
investment from a PE in the 30s,which is what Amazon is going
to be capable of.
And, david, the other thing isAWS.
We think the biggest and mostintelligent models they're going
to live in the cloud.
We think AWS was behind.
(23:06):
They're catching up.
They've got a pretty aggressivegrowth map.
And the other side of this ispost-DeepSeek.
If AI models become commodities, the number one winner from
that is probably Amazon Because,again, they haven't been
involved.
Aside from some investments incompanies like Anthropic, they
haven't been involved in theincreasing spend to build their
(23:26):
own models.
So if everything becomes acommodity, that just makes AWS
even more valuable.
So you know, david, that's kindof the pitch on these two
companies Outside the AIportfolio.
I've said this many times sincewe talked about them.
I hadn't originally added thembecause I thought a lot of
people already own them.
We're looking for smalleropportunities, but I do in my
(23:48):
own personal investing.
These have been two of mylargest stocks for years and I
like the way that they balanceout the portfolio, so these
position sizes will probably getsmaller, but that's why we're
starting at a larger positionand, as I noted, we'll sell this
down to get them.
It wouldn't surprise me if, insix months, we had something
closer to 50K of each of thesestocks.
Speaker 1 (24:11):
Okay, so let's move on to those are the hyperscalers
.
Again, it feels like theconsensus has been very positive
.
You talk about 20 straight daysof positive gains.
I don't know what Amazon is,but I know Amazon over the past
month.
You know it's still.
It's positive.
It traded up.
Even after DeepSeek it didn'tseem as affected as obviously
(24:32):
NVIDIA had the massive sell-offand a lot of our other kind of
data center plays.
We've also gotten earnings fromsome of those smaller companies
in the data center where, again, it's been a wild couple of
weeks for them.
They had massive, massivesell-offs after that news.
Then we saw some jump back.
I can't recall the company youtalked about last episode that
was down 40% and thenimmediately made it all back in
(24:54):
two weeks.
We've had some of that, butwe've had some of these
companies that have sold off andhave certainly not recovered to
where they were pre the deepseek story.
So what have we heard fromthose companies?
It does seem like the consensusis not as cheery in that corner
of the market, but also not asthe world is ending and these
companies all growth prospectsare dead.
(25:15):
So what have you heard fromthese smaller companies in the
portfolio that had been rocketships but have hit a little bit
of a stumbling ground now.
Speaker 2 (25:22):
Yeah, in the short term, most of its stocks,
whether or not it does good orbad, isn't going to be driven by
results per se.
It's what you call narrativeand it's whether or not it's
multiples on earnings or salesor contracting or expanding.
So I think what we have rightnow is, with DeepSeek, you had a
(25:43):
very risky an event that showed, maybe you know, these stocks
are very risky and you've had alot of people want to kind of
branch out to different areas.
Now the question is, is that ashort-term bump or is that a
longer-term concern?
So let's drill in on.
You know, if I was talkingabout our data center stocks,
(26:03):
you know, in optical we've gotCoherent.
You know, in optical we've gotCoherent, lumentum, fabrinet,
we've got companies like Sienna,then we've got, you know, other
plays Credo, vertiv, schneiderElectric.
So we have a good collection ofthese stocks and, as you noted,
they are probably the aspectweighing down the portfolio the
(26:23):
most right now because they aretrading about where they were
after deep seek.
So let's zero in on coherence alittle bit more of a bellwether
, especially in the opticalspace.
They report earnings.
The earnings were up 252percent from last year.
I I trust you agree, david,that's pretty good.
(26:45):
Um, well, just just just, uhbeat expectations by a tad.
I think it was something like93 cents a share of earnings
versus 69 cents.
Wall Street was expected.
They were up 14% the nextmorning, which is kind of
expected reaction when you havethose kinds of earnings and
you're beating outlook for nextquarter.
(27:05):
They got to about $100 pershare.
But, david, what we've seensince then is shares have kind
of just every day just kind ofdropped 2% or 3% to where it's
trading almost exactly kind ofmid-80s back where it was
post-Deep Seek.
So the question is what's goingon?
There's a lot of factors here,but I think one issue is right
(27:29):
now NVIDIA is just such a largepercentage of the spend around
these kind of AI data centersand what was expected was that
NVIDIA was going to see a verymassive ramp of a specific
product called GB200.
And this is the one that's beenin our production.
Now, david, if you wanted tofollow every single bit of
(27:49):
speculation and people sayingthat there's engineering
problems or this and that withGB200, you'd go insane
scuttlebutt coming out of thesupply chain.
(28:10):
But it does seem like what'sgoing on right now.
Is that the size of the rampfor this specific product is
being curtailed, expectationsare coming down and a lot of the
demand is moving to the backhalf of the year because NVIDIA
was very aggressive aboutcreating a successor product,
the 300 series.
That is a vast improvement andisn't a significant weight.
So I think what's going on isthat's causing a lot of chop and
(28:33):
it's causing a belief thatthere's going to be kind of an
air pocket for a lot of thesecompanies and I think investors
are kind of trying to get out oflooking stupid in front of some
near-term disappointment, forlack of a better word.
So my bottom line, david, Istill think that these are good
companies.
I still like their kind ofpotential, especially across the
(28:57):
entire year.
I think they're just goingthrough some chop right now,
especially related to thespeculation One area that's
doing great we've talked aboutin the past.
We're going to have all thesebig data centers.
You're going to need to getthem to do kind of coherent
training together.
You're going to need to networkthem all.
Sienna is a big winner fromthat, but also Coherent itself
(29:19):
and Fabrinet are winners.
That's how the business isdoing great.
It's just a lot of the businessthat's kind of blowing through
from Nvidia is seeing a bit of apause, so I did add Fabrinet.
Right now it's trading forabout 20 times earnings ex cash.
It's taking a blow from maybelosing a little content in
Nvidia, but it's got some newcustomers, including Sienna,
(29:42):
that are going to be great forthe long-term story.
So, again, these stocks aredown right now, but if this was
a part of the portfolio youhadn't added to, I would say
it's a good entry point for alot of these stocks, taking a
bit of a breather from how hotthey had gone.
We added Fabrinet.
That one continues to be one ofthe cheaper options in terms of
(30:03):
PE, but I haven't seen anythingthat changes my long-term
opinion.
I think we're just kind ofworking our way through a bad
narrative.
Now there are some sectors ofthe AI space with good
narratives, which is what we canget to next right.
Speaker 1 (30:18):
Yeah for sure.
Yeah, I was just going to note30 minutes into the podcast, I
don't think we've really talkedabout NVIDIA, which must have
been a record.
And you're talking about allthe spend from Meta, amazon,
again, everyone, all the spend,not all of it, but a huge chunk
going to NVIDIA.
We should note NVIDIA stockobviously had the huge sell-off.
It's back to flat year to date.
(30:39):
It's recovered a little bitfrom the deep seek story.
It's performing a little bitbetter than some of these
coherence or some of thesesmaller plays.
So again, what we've talkedabout since we've started the
show is that there might bethese moments of air pockets and
shop and it seems like themarket reaction is appropriate.
If NVIDIA is flat, it makessense that a coherent is maybe
(31:01):
down 15%.
It's not the same size ofstocks, same size of company.
So I don't want to get intoNVIDIA.
We'll talk about it.
We've talked about NVIDIAenough and I think we have their
earnings coming in a couple ofweeks, is that?
Speaker 2 (31:15):
right, yeah, they'll get some coverage.
Speaker 1 (31:17):
We'll get some coverage then.
So any more earnings you wantto hit on?
We had Cloudflare, which is anew company that we added to the
portfolio.
Reported earnings Positivereaction there.
Is there anything you want tohit on Cloudflare there?
Speaker 2 (31:29):
Yeah, and it just kind of shows how much in the
near-term narrative drives shareprice.
We had Cloudflare before itsearnings report.
You know there's anotherpodcast I'm sure people out
there listen to it Bill Simmons.
They do the guess the linewhere you say a football game
and you try and guess who'sgoing to be favored.
I was doing a guess the line.
When I saw Nets earnings kindof the numbers break I said ah,
(31:51):
it's down 10%.
I believe they are now up 20%and it is just partially.
This company has just such aphenomenal story right now.
It's.
You know, david, if you know.
Once again we'll kind of archivepast episodes of the podcast.
But in one of the earlier onesyou would ask me what's a stock
(32:13):
that you love but it's just alittle too pricey right now to
add to the portfolio.
I said Palantir because we hadstarted last year publishing a
report on AI.
We had picked out Palantir asone of our favorite
opportunities and just keptgoing up and up and around the
time that we had recorded thatepisode it was trading closer to
I believe it was about 40 timessales.
(32:35):
I said it's just a bit too upthere for me to touch at this
point.
It's tripled since then.
So I wish we would have added.
I still personally hold it.
I haven't made any changes to mypersonal holdings, but the next
best story in terms of howcleanly you can take a company's
(32:55):
business and kind of draw it totrends in AI after Palantir is
probably Cloudflare.
So this is an expensive stock.
I think when we added it it waswhen we were before earnings.
When we were looking to add it,it was trained closer to about
20, 25.
It was trained closer to 25times this year's sales.
(33:16):
It's even more expensive now.
And the reason, dave, isCloudflare is what you call an
edge network.
I think it's not like a greatbusiness.
You probably remember Fastly,which was doing content delivery
.
Unfortunately, I do.
Yes, everyone loved it in 2020,but you know again, it just it
(33:36):
wasn't a great business becauseif you start, you know what
you're doing is kind of buildingservers that are located closer
to people.
If you're not able to buildmore value-add services on top
of that and you know, do theclassic land and expand, it's
just not going to be a greatbusiness.
But that's exactly whatCloudflare has done.
They have tremendous execution.
(33:56):
They've moved from contentdelivery and what you'd call
denial of service attacks tobuild a much bigger security
business.
And what they did was theybuilt a developer platform which
you can run services closer tothe consumer, which was a pretty
niche concept until AI took offand, all of a sudden, training
(34:19):
you can go do a training runanywhere.
It's not that big of a deal ofneeding locations closer to
consumer With inference latencymatters a ton.
So all of a sudden, thesedeveloper services that are
offering a network that'sextremely close to users has
incredible value.
They call it workers AI andthat's the service, once again,
(34:42):
for running inference throughCloud Flare.
The other thing that you have issecurity.
People have asked us to talkabout security.
Security is really taking off.
You look at the charts of allthese security stocks.
It is up and to the right rightnow and that's because of a
realization that security is anarea where being a leader maybe
gets its next advantage.
(35:02):
Because, again, we've talkedabout unique data being the
biggest key to AI and the moreattacks that you're seeing by
being the leading vendor, themore data you have to be able to
build a better product.
So it's kind of aself-reinforcing cycle people
are now buying into.
And finally, with Cloudflare youknow, david, again with our
business.
When AI first started, when AIwas really taking off after
(35:27):
NVIDIA's famous earnings reportin early 2023, when I was
building out my initial AIportfolio, one of the first
stocks I went to was Cloudflare,because I was like, if there's
all this traffic from all theseAI agents and also at a certain
point you're going to need toprotect your data against
basically being trained for freeby these models, the number one
(35:48):
company to fill that gap isactually Cloudflare and that's
what they're moving into rightnow.
They're just starting it where,essentially, you can protect
your content, you can charge arate to be able to be ingested
into models, and that's whatCloudflare's next big business
opportunity is going to be.
So, david, it's an extremelyexpensive stock but, again, this
(36:10):
is probably the cleanest storyI would call it in terms of how
their products really lead intothe growth of AI right now.
You saw Palantir the past year.
Them seeing I don't know whatit is it's probably like 600%
growth from when they were at$20 near the beginning of 2024
to today.
They didn't require going from50% growth to 150% growth rates.
(36:34):
They grew their top line about20% to 36%.
I think Cloudflare as we see alot of growth in inference from
the emergence of these reasoningmodels.
We'll see a similar expansionon top line and I think it's
going to be a stock that peoplemost closely associate to the
growth of AI.
(36:54):
So I don't think we want aportfolio full of these
companies, right, let's?
Let's call a spade a spade, astock selling for a sales
multiple.
This high is risky.
For a sales multiple.
This high is risky.
But I also think, if you arewanting to take some chances on
having stocks that kind ofdefine this generation, I think
(37:14):
Cloudflare has a space in yourportfolio, which is why we added
15,000.
And I think there's anotherhuge narrative in AI that we're
about to add software stocktoday that I think will also be
a leader in one of the spacesthat's going to attract
significant media attention thisyear.
Speaker 1 (37:35):
Yeah, I want to get to the new company.
But just to your point on highvaluation multiples, high price
to sales, high PE.
If there's any lesson learnedover the last couple of years
with some of these companies inthe AI space, it's be careful.
Judging only by those valuationmetrics is right before an
inflection point on the top lineor even on earnings.
I don't know what they're goingto get to.
In today's show we had Applovin, everyone's favorite named
(37:58):
company report earnings Again acompany that, when you look at
just the amount of cash they'rebringing to the bottom line
because of AI, it's juststaggering the growth that some
of these companies can see.
I think Applovin was up 700%last year and it was up 33% in a
day this week.
So very risky.
(38:21):
Just to say it's expensive,I'll never buy it.
You can often get paired withexplosive growth on the business
side, which very quicklydoesn't make it quite as
expensive.
Speaker 2 (38:25):
So I do want to talk about Applo, and so let's come
back to it, but let's do thestock we're actually buying
first.
Speaker 1 (38:31):
Sure, let's get into the purchase.
We are adding a new company tothe portfolio.
Let's kick it off.
Speaker 2 (38:35):
Yeah, the company I'd like to add today is GitLab.
I think we'll follow a $15,000purchase In terms of some of the
leaders in companies at theforefront of software trends.
That will give us a cloudplayer and you know security and
inferencing.
It will also give us Snowflakeand we'll have GitLab beyond
(38:57):
also other softwareopportunities like Synopsys and
Cadence.
So we are starting to build outkind of the software side in
addition to, you know, many ofthe data center plays we had.
Now the storyline here is, davidI'm sure you've seen it right
there is a chart making therounds on social media right now
that shows developer jobopenings from 2022 to today, and
(39:21):
it's absolutely fallen off acliff, right, and you know the
the reason many will attributeto this.
There's several reasons, butright now there are coding tools
utilizing AI that make peoplephenomenally more efficient, do
some stuff.
That is honestly kind ofshocking.
You know, david, we met in LasVegas, as we talked about
(39:43):
earlier, and, talking with thedevelopers on our team, their
embrace of AI has completelychanged the way that they do
their jobs.
Not only that, but another oneof our co-workers he was showing
me yesterday that what he callsvibe coding.
You know if you get in too much, you'll see a lot of people
talk about this vibe coding now,where you're working more with
(40:07):
ideas than the code itself.
He had taken some that we mightpreviously had to pay someone on
Upwork maybe six figures, maybe$100,000, to build some
concepts, and he had just doneit all in a spare time with a
week knowing almost no, havingalmost no ability to code.
Let's just call it is what itis.
So the preeminent productthat's getting most of the
attention in Silicon Valley issomething called Cursor.
(40:29):
It's a code editor that set thenew record, as the PR would go
and going from 1 million to 100million annual run rate in 12
months.
I've seen people say that, uh,when they're talking with
classes at in Stanford, uh,computer sciences, if they ask
how many are using cursor, it'sa hundred percent people.
Speaker 1 (40:49):
Wow.
Speaker 2 (40:50):
Um, if you look at Y common air one of the best
incubators for new startups ahundred percent usage uh, cursor
right now.
So if you're at the bleedingedge right now, the adoption for
Cursor is maybe at the highestlevels of any tech product ever.
So, being able to code with AI,it's a big deal, and these
(41:12):
reasoning models where theyperform best is areas like
coding, because that chain ofthought gives them objective,
true or not true statements andit's really going to turn coding
into a superpower extremelyfast via AI.
So I think the growth rate inthese areas is going to be
absolutely incrediblegrowth-raising capabilities.
(41:35):
So there's a space named DevOpsand that's for code management.
It's a great first place tolook in terms of where you could
see, you know, some gains fromthis trend.
There's a company named GitHub.
It's very strong, but it'sowned by Microsoft and you know
Microsoft's very big.
It's just a small piece of itsbusiness.
So we're looking at their topcompetitor, which is GitLab.
(41:57):
They have a lineup named Duo.
It's not just coding assistants, but it's agents that monitor
across the entire workflow ofall company code.
It can not only function areaslike security, but also code
optimization.
So Cursor it's going to be usedheavily by startups, first
movers.
(42:18):
But if you're in a heavilyregulated space and you're
looking for a lot of theseefficiencies, a company like
GitLab is exactly where you'regoing to look, and they have a
ton of differentiation here.
So you think about Bank ofAmerica.
Gitlab actually specializes inbeing self-hosted, so you don't
need to go to an LLM out in thecloud or untrusted spaces.
(42:40):
You can actually keep thingslocal and use GitLab.
So, again, you think about allthe industries you could think
about healthcare, finance,energy, national security.
This is exactly the kind ofproduct they're looking for.
And when it comes to actuallyhaving relationships in the
cloud, well, I mentioned theirmain competitor, github, owned
(43:01):
by Microsoft.
Well, that's an opportunitybecause it means GitLab can
partner more strongly withAmazon and Google.
So, David, this is a companythat's been growing among the
higher growth rates in the SaaSspace.
It was 31% last quarter.
It's expected to decelerate to25% this next year.
If there's a big push to lookfor solutions in being able to
(43:26):
find agentic coding assistanceacross your code base, they're
going to be the first companythat a lot of people look to.
If they bump up that rate from25 expected back to 30, it's
very likely going to be awinning investment.
They're trading at 12 timestheir expected sales this year.
That's high, but it's also notoutside a band of SaaS companies
(43:49):
.
That one with one of the bestopportunities is a problem.
So there's some risk here,david.
It is a little more expensive.
It does have good competition,but in terms of getting a
company that's aligned with whatI see as one of the most
exciting trends and somethingthat gets a big push from the
recent breakthroughs andreasoning that maybe the market
(44:10):
has internalized, I really likethis play and I think there's a
good chance that we look back ina year or two and GitLab is one
of the stronger portfoliocompanies, right.
Speaker 1 (44:21):
Yeah, when you talk about Microsoft and GitHub, you
have a little bit of PTSD fromhearing Microsoft as a
competitor if you were a Zoomshareholder, or just the ability
for Microsoft to come in andjust kind of dominate a
competitor.
So is that the biggest outsideof valuation?
Is that the biggest risk?
(44:41):
Is that they just get squeezedon pricing, or is that what I
should be most worried about ifI'm thinking about this company?
Speaker 2 (44:48):
I think so.
If I needed to say what was thesingle biggest risk that
concerns me when I look at theircustomer acquisition, it's been
a little weaker in recentquarters.
And if you want to, you knowyou're fundamentally in SaaS to
land and expand and if you'renot landing you can't get people
on these tools that we thinkcould be the growth sweetener
for the company.
Hard time landing in thiscompetition with Microsoft,
(45:18):
which doesn't allow kind of thisproduct suite to give that
incremental growth we thinkmakes this a winning investment.
Speaker 1 (45:23):
Okay, so that was GitLab.
You said 15,000 routing earlierthere.
Speaker 2 (45:28):
Yeah, and one note, just a minor programming note,
because I have been battling thefeared Las Vegas flu and
traveling.
We haven't been able to do theSK Hynix purchase.
Everything else that we'vetalked about has been purchased.
We haven't been able to do that.
Next week should have an update.
I will make sure that thespreadsheet we track all this I
(45:51):
need to get updated Again.
That's just been a recentcasualty of not coming out of
bed for several days in a row.
But we will make sure to getthat update.
And you know again, we'llpublish more portfolio review
pieces, I believe in the comingweeks, since you know we're
trying to catch up with a lotright now.
Speaker 1 (46:10):
We fully realize that , yeah, that sounds good.
Yeah, and we actually we have afew listener questions.
I think we're going to bumpthose.
Maybe we'll do special YouTubevideos on answering these
listener questions becausethey're running a little bit
long on time.
And I did want to circle backto Applovin, again a company
that is probably more onpeople's radar than it was, I'm
guessing, 18 months ago.
It was on no one's radar andnow this is a $200 billion
(46:31):
company.
So I did want to get your quickthoughts on Applovin.
What should we make of it?
Again, this is a company thathas been utilizing AI in a
different way.
It's not a data center stock,it's not NVIDIA, but it's been
the most successful kind of techcompany over the last 14 months
now.
So what were your thoughtscoming out of just monster
earnings from this company thisweek?
Speaker 2 (46:49):
Yeah, and just quickly on the listener
questions, there's one fromSaltyCuddy707.
He left in a five-star review.
That's amazing.
Anyone leave a five-star review?
We'll answer your question,salty.
We will put that on our YouTubequestion.
You wanted our thoughts onPhotronics, which is a photo
mask manufacturer.
We also had one from JohnMcGuire on our YouTube channel
(47:10):
Asked about SteraLabs.
He asked on YouTube.
We'll put it there Nowlovin.
I think this is just fun.
So I want to talk about thisbecause it kind of shows how
kind of emerging ideas function.
Applovin, you had noted this isone of the story stocks of the
past year, when we were talkingabout companies with unique data
(47:36):
and sometimes the pivots thatthey're able to make and how the
market rewards that.
Apple 11 was a company I hadowned.
To be honest, I had sold, inNovember, I think, around $330
because I thought they were kindof reaching the end of their
opportunity in the video gamespace.
They just reported earnings areup 30%.
(47:57):
They're close to $500 per share, because now they're trumping
an e-commerce kind of push.
So you know, I don't know ifthat will be persistent, but the
point is, the market reallylikes what they're doing.
Now.
There might be anotheropportunity here.
David, though, and I kind ofwant to walk through it.
You know, I just I was justlooking at this on Thursday, so
(48:20):
this is like real time thinkingout loud for the audience.
I know people out here probablylove trade ideas, so, again,
applovin up 30%.
What I've said about Applovin inthe past is they're kind of
like Google if they didn't giveevery employee free dry cleaning
and filet mignon dinners everysingle night.
(48:41):
You know, this is just acompany that takes dollars on
the top line.
They convert to profits on thebottom line at an incredible
pace and they have unique kindof mobile gaming data.
They've utilized an aggressivepush into AI ad tech, and
they've done that to getphenomenal growth.
And you might say how themarket missed this, that the
(49:05):
company is up more than athousand bolts since the start
of 2024.
And the reality is the marketmight not have fully missed it.
Their greatest competitor is acompany named Unity that was
worth close to $100 billion in2021.
But, david, the opposite isUnity.
They gave away a lot of freedry cleaning, a lot of filet
(49:27):
mignons.
This is just a company that hadincredible expense levels, so
they weren't able to kind ofachieve the same areas as
ZappLove and also their mainbusiness is ad tech, but they
make a game engine and they justdidn't make the same pivot to
focus on applying a lot of thesebreakthroughs in AI to their
(49:49):
advertising technologies.
And, david, what's interestinghere is you look at the track
record.
We talked about Meta earlier.
They were using CPUs instead ofGPUs from NVIDIA.
They did a huge build-out to doall their content
recommendations through GPUsinto 2021.
They started flipping theswitch with all these new data
(50:12):
centers they built in late 2022.
If you go and you look at stockcharts and they're close to a
thousand percent since thebottom app loving was a fast
follower.
They were doing press releasesabout how they were, you know,
moving their um ad tech to youknow again, these, these newer
transformer technologies in in2023 and they're up more than
(50:35):
1,000%.
Unity today is worth $8 billion, so about four times sales.
Apple opens at 34 times sales.
So what's interesting here,david?
A company that clearly hasn'tbeen able to keep up with
competitors but this week,buried in the news, they're
doing a massive reorganizationwhere they are reorganizing
(50:56):
entirely around this ad tech andthey're going to have an
opportunity to kind of capturethe same thing that's produced
those almost generationalreturns in those other companies
.
So do I think?
Do I think Unity is going tonecessarily be an automatic 5X
in the next year?
Not necessarily, but they havethis unique data that we've seen
(51:18):
other people utilize to getthese kind of generational
returns.
They're making the shifts intheir business and this is a
company that, again now I'mwondering if the time is, you've
been a loser if you've beenbuying any time in the past few
years, but this might be thepivot point where you are
healthily rewarded for it.
So you know, this is where wesee an app loving and a company
(51:41):
that's producing crediblereturns, and maybe we won't buy
that.
But by following the story andunderstanding it, it might lead
you to discover some other gems.
Speaker 1 (51:51):
So I thought that was oh, go ahead, David yeah.
I was gonna say I think thiscould be a very interesting case
study of there's a company thatlaid out the path to success
with AI, and it's a question ofexecution, of how hard is it to
go and execute these ideas.
It's easy for you know someoneto get on a conference call and
say we're going to do that too,but now let's see if they
actually do.
(52:11):
I think it'll be fascinating tofollow.
So that is a story withApplovin and Unity.
I think that's all we got fortoday, right, anything else?
Speaker 2 (52:19):
Yeah, like you said, we'll answer the questions on
our YouTube channel.
If you aren't subscribed there,go ahead and subscribe to 24-7
Wall Street on YouTube.
We'll get those handled.
If anyone else has anyquestions, send them our way.
And, like we said, there's beenenough news.
We would have liked to havedone a show earlier, but maybe
(52:42):
I'll return to my sleep holeafter we finish recording Go get
some rest For Eric Flieger.
Speaker 1 (52:47):
I don't buy or sell based solely on what you hear.
On today's episode of the AI Investor Podcast from
24-7 Wall Street, we're checkingin on earnings from tons of
companies inside of the $500,000AI portfolio and adding a new
one to the mix.
All of that and more is nextWelcome back to the show.
I am David Hanson, joined byEric Bleeker.
Eric, it's been two weeks sinceour last episode, which, in AI
(00:24):
dog years, it feels like it'sbeen about 13 years.
There's so much we've missed.
I'm happy to be back.
We did a couple of weeks wherewe had it weekly.
Then we went to buy it weekly.
We were traveling last week sowe didn't have a chance to get
the show up.
So apologies if anyone wasexpecting a show last week, but
we're here now, excited to jumpin.
We've had some time to digestthe big DeepSeek story which we
(00:47):
talked about last episode.
We've had a few weeks to digestit.
Hear from some other companies,so we're going to get into some
earnings.
But I wanted to start with justa couple other news stories.
Deepseek obviously dominatedthe headlines a couple weeks ago
and everyone was wondering well, does that mean all this
capital spending is going tolighten up?
We haven't necessarily seenthat a lot, and one of the
stories that caught my eye wasyou know, we had the big
Stargate announcement, thenDeepSeek and now Europe is
(01:07):
getting.
Remember Europe, that continentwhich at one point dominated
the world, which had a rough run.
So Europe is getting in the mix.
So walk us through.
What is the AI announcementcoming out of Europe this week?
Speaker 2 (01:16):
Yeah, you know you talk about how long it's been.
I feel like that Titanic memewhere it's like it's been 84
years.
Uh, yeah, and, as you mentioned, we were traveling.
I got this thing called thevegas flu.
That's, that's not a flu.
It doesn't show up on tests orcovid tests.
But you know, everyone outthere be be be careful, there is
(01:38):
something weird going around.
It has laid laid me up.
I'm glad I can even talk.
So, yeah, you're up just aboutout of every single conversation
in technology for about thepast decade, aside from seems
like passing some sort of newregulation with relatively
sweeping consequences.
But, like you said, the USbegins kind of the narrative in
(02:00):
2024 with the Stargateannouncement and before that the
CHIPS Act.
In 2024 with the Stargateannouncement and before that the
CHIPS Act, we've got DeepSeekcoming out of China and, as we
noted in the last podcast,deepseek was paired with a very
large funding bill.
I believe it's like $137billion worth of funding for AI
from China.
And now we have Europeannouncing it's France, if we
(02:24):
want to be specific a $113billion plan.
It was kind of funny.
The prime minister, thepresident of France, macron, he
had I'm sorry I'm saying thatwrong as well.
He had done his version ofdrill baby drill with plug baby
plug, promoting their nuclearpower.
(02:46):
So you know, david, I do thinkthat this shows, you know, as we
head into 2025 and we're goingto be continuing to follow this
kind of storyline and givingeveryone this kind of front row
seat to this storyline of therise of AI, geopolitics go front
and center.
(03:07):
We've seen Europe, which hasbeen hesitant, now a major
country on the continent,announcing a very sweeping and
ambitious plan, and I'm surethis isn't going to be the last
area.
We see Areas like the MiddleEast, even if they're not
getting the headlines, they'rein the background, providing
funding to areas like Stargate.
We have, you know, japan doinga lot of funding for their
(03:31):
semiconductor industry andtrying to, you know, bring
funding back there, and alsowe'll talk more about this later
, I'm sure but we also now haveIntel rising this week on kind
of speculation that the nextmajor push from the Trump
administration is going to begiving them a lifeline.
So this, this stuff reallymatters, and I think it shows
(04:00):
that AI is going to move from abig investing story to the pre.
You know, companies in the AIspace.
You're going to need to payattention to developments
because, as AI's impacts ramp,how to make sure that its
benefits accrue across societybecomes the number one issue
facing the world, and it's goingto be a mad scramble to figure
(04:21):
out how to do that.
Speaker 1 (04:23):
Now, this is a recently announced plan and,
just like Stargate, we don'thave all of the details of what
exactly they're doing.
But is it in the same realm ofa Stargate, of building out
larger data centers, the typesof companies that would benefit
from a Stargate type project?
Is this the same type ofinvestment that France is
talking about here?
Speaker 2 (04:38):
Yeah, I think people have called it Le Stargate, so
it's smaller and absolutefigures.
But yeah, it's similar.
It's trying to get companies tobuild large data centers,
potentially using France'savailable nuclear power.
You do look at how geopoliticshas gone.
A big export market of France'snuclear power is Germany, which
(05:01):
turned off its nuclear power,and now France is offering this
for AI.
So you see some of the dominoesfalling, but smart move for
France, in my opinion, becausethey have a great asset in their
amount of nuclear power andthey're going to make it kind of
the European hub for AI.
Speaker 1 (05:18):
Well, as an investor who has unfortunately had maybe
too much of his portfolio tiedto Europe over the last couple
of years I'm excited to at leasthave maybe some I don't know if
it's good news, but just someexcitement or growth prospects
for some of our European friendsover there.
So that's exciting.
Again, we talked about DeepSeeka lot last episode.
(05:38):
We've had a few weeks to digestit and I wanted to get your
take on some of the companiesbuilding models here in the US.
So that's OpenAI, anthropicGrok what have we been hearing
from those companies?
Have they changed their tune atall?
Again, the big discrepancy wasokay, deepseek does this for
less and OpenAI Sam Altman hasbeen out there for over a year
(05:59):
now saying I need trillions ofdollars Podcast bro, if we
recall of him going aroundtrying to drum up all this
enthusiasm for more spend.
Stargate have they changedtheir tune at all since the in
these last two weeks, or whathave we?
What have you heard from thosecompanies?
Speaker 2 (06:13):
Yeah, it's interesting.
You're seeing kind of abifurcation of kind of companies
talking their book, in a way,where out of China you've got
Alibaba saying the Americancompanies are doing all wrong
and that's a great narrative forthem because it works in their
favor and they're up 40% inrecent weeks as people kind of
look to stocks that have franklybeen stuck in the mud for years
(06:34):
and now have kind of catalystto get out of it.
But you look on the Americanside and it's kind of business
as usual.
In fact I would say it'sdoubling down on kind of the
storylines that we've had.
You know, david, the key ideawe've been following since
(06:58):
December has been thisbreakthrough in these new types
of AI models that use reasoningkind of a bull on AI and believe
that we're going to seecontinuing acceleration this
year, which I am A big reason iswe are still so early onto this
path of harnessing thesereasoning models and seeing
their capabilities skyrocket.
So, from OpenAI, sam Altman,who's the CEO of the company he
(07:22):
dropped a big blog post and heemphasized that intelligence of
a model it's still CEO of thecompany.
He dropped a big blog post andhe emphasized that intelligence
of a model it's stillproportional to the resources
used to train it.
If that's true, it's great newsfor all the data center
companies that we've talkedabout.
And he emphasized which wetalked about in our DeepSeek
episode that the cost of AIdropping by phenomenal levels,
(07:44):
which is what caused the DeepSesea panic is nothing new.
He had quoted their GPT modelsbetween 2023, mid 2024, 18
months, 150 fold drop in costperformance.
And he noted Moore's law, whichwas the fundamental centerpiece
of technology, being kind ofthe most transformative force on
(08:05):
earth.
That was 2x every 18 months.
So 150x versus 2x is striking,and he also emphasized that they
see no reason for investment tonot continue growing at an
exponential rate.
In interviews he was alreadytalking about you know, hey, we
just did Stargate with 500billion, the next one's 5
trillion.
Now again talking his book,right, talking his book.
(08:28):
But you know nothing.
Nothing has changed materiallyand in fact, with open AI, they
released a product named deepresearch.
David, have you played aroundwith deep research?
So it's on.
It's still under $200 tier.
They are going to move it to befree pretty soon.
So they've said I got a chanceto play around with it.
Really impressive Did someindustry reports in some areas
(08:54):
that we were looking at for ourbusiness.
It used novel ways of findingdata, it put it into tables,
making kind of unusualconnections that if I had
received this report and it wascoming from some management
consultant charging $50,000 forit, it wasn't far off from the
quality that you would expect.
(09:15):
So a lot of people have beencalling deep research kind of a
little bit of an iPhone momentfor the capabilities of AI, and
we're still near the front of it.
The other thing I would mentiontoo, the other companies.
We're about to see a lot ofleapfrogging because Anthropic.
They are presumably readyingtheir next model.
They came out with a statementkind of directed towards
(09:38):
politicians around the world,saying that by 2026 or 27, no
later than 2030, we're going tohave basically what they call a
nation state of intelligentbeings, which is AI agents.
And Elon Musk spoke at aconference I believe it was in
Dubai where he said Grok 3 willbe released before the end of
the month and that it, in theirtesting, outperforms all other
(10:01):
models.
So, david, we'll get to thehyperscalers the people actually
spending the money later inthis podcast, but from these
companies building the models,there's been almost for the size
of the deep seek reaction.
You have expected, maybe somekind of change their tune, but
in fact, if anything you knowthey are like I said earlier,
(10:23):
they are doubling down and Ithink what we're going to see in
the months ahead is kind of afurious pace of releases.
That, of course, will be funfor us to follow on this podcast
, right.
Speaker 1 (10:34):
Yeah, unfortunately, you know, openai, anthropic,
grok, those are.
They're not public companiesright now, so we can kind of
only assume how the market wouldmaybe be reacting.
We do have, obviously, meta,who's not a direct competitor to
those but is obviously buildingmodels as well.
We're going to get to Meta in alittle bit but, agreed, it does
seem like the vibes for lack ofa better term have got a little
(10:57):
more positive over the lastcouple of weeks, kind of what we
said in the DeepSeek episode,where there is something here,
but the overall story is notreally changing and in a way, as
long as it opens up more usecases, more use of AI, it could
be more bullish for thesecompanies, despite spending
billions and billions of dollarsBefore getting to hyperscalers
and companies that have madetons of money.
(11:18):
I wanted to touch briefly on acompany that loses money and
that's Intel, a company we'vetalked a little bit about.
You said I don't know what thefuture here is.
Is this company to break up?
Is it going to need a bailout?
What's going to happen here?
The stock is up around 20% thelast couple of weeks.
Are we getting some sort ofoutcome here or do you have any
speculation on what might begoing on with Intel.
(11:38):
This isn't a company in theportfolio.
I think you said you don'treally have an interest in
bringing a portfolio, but onethat we've just talked a little
bit about.
Speaker 2 (11:45):
Yeah, you know, the thing with Intel is it's it's
all speculation right now.
So you know, this is this is aat the moment, it's a gambler
stock Right and the thespeculation of what's happening
right now.
Donald Trump he has floated outtariffs and you know, it
(12:05):
depends how you want to look atTrump, but the kind of, I'll say
, industry consensus is thesetariff ideas for Taiwan.
It is largely the beginning ofa negotiating tactic to get some
concessions from Taiwansemiconductor.
So the reports are that there'sthree options on the table.
Number one build morefacilities in the US.
(12:27):
They built one in Arizona.
They have issued statementssaying that they are happy with
its performance.
Taiwan Semiconductor I've saidin the past, David, they are one
of the most relentlessnegotiators in the world.
You're going to have themagainst Trump.
It might make our heads explodein two sides that just want to
negotiate, but you saw, withTaiwan Semiconductor, when they
(12:49):
were building their fab inArizona, everything was US
workers are lazy.
This is not great.
Well, they're, you know, againlooking for more subsidies, more
concessions in the moment.
And it's done, it's oh yep,we're performing great here and
the PR shifts.
So you know there is a questionof whether or not, we could get
more facilities, and ones thatproduce kind of the most leading
(13:12):
edge technologies, which theyhave said they're open to.
So it's building morefacilities in the US.
There's another plan on thetable them forming a joint
venture with Intel, and thiswould effectively force some
technology transfer to thecompany.
And then there's a third optionthat I've seen reported on,
where they could take ordersfrom companies like Apple, send
(13:33):
them back to the US forpackaging by Intel.
As we've talked about a lot ofthe value kind of add with where
semiconductors are going.
It's moving from actually thesteps like ASML's lithography to
how you can package these chipsup, so that would give US more
expertise in the area thatTaiwan Semiconductor leads in.
(13:55):
So, David, I don't know wherethis is going to end up, but
again it kind of just circleswhat we talked at the beginning,
that chips and the AI race.
It's moving from somethingthat's probably a niche interest
for investors into the front ofgeopolitics, and for a lot of
these companies it's going to beinescapable that the short-term
(14:16):
movements are really going tobe controlled by what's in the
headlines around, what worldleaders are looking to do.
I think at the end of the day,though, that will mostly be
positive, as we're going to seea lot more countries wanting to
put their investment and securethis kind of as a strategic area
.
So overall bullish.
But for any specific news, thisspecific news is slightly
(14:41):
negative for TaiwanSemiconductor, slightly or
potentially very positive forIntel.
Speaker 1 (14:47):
Okay, all right, let's move on to the
hyperscalers.
We've had some earnings, sothat means we've got some
commentary from some of thesecompanies.
In particular, let's jump intoMeta and Amazon.
These are two companies thatyou recently added to the
portfolio in a big way, putting$100,000 behind each.
These are the largest positionsin our portfolio that weren't
there before.
Last episode, you talked aboutsome of the rationale behind
(15:10):
adding a company like Meta andAmazon, again benefiting from if
they have to spend a lot, theycan win, or if the cost profile
comes down, they have enormousdistribution, so you felt
comfortable using them as ahedge to the rest of the
portfolio.
So what did you hear from thesecompanies?
Are you more excited?
Are you regretting putting$100,000 on any of these
(15:33):
companies based on what youheard?
So what are your thoughts?
Speaker 2 (15:35):
post earnings David, I know it's been 84 years since
we last talked, but coming outof DeepSeek, the big question
was on everyone's mind are wegoing to see hyperscalers adjust
what their planned spend levelis for 2025?
We're heading into the year andthat means they're issuing,
(15:55):
basically, growth projections.
They're giving guidance for2025.
So this was a very pivotalearning season for seeing what
this AI story looks like.
Now, if you go back to theNovember 8th podcast for the
aficionados that have catalogedall the different AI investor
podcasts, we did walk throughexpectations for every single
(16:16):
hyperscaler and what their spendper Wall Street expectations
was expected to be this year,and that was Amazon $85 billion.
Microsoft $84 billion.
Google $57 billion.
Meta $48 billion.
So we've gotten reports frommost of those companies now.
From Amazon, they gave a numberthat isn't implied about $105
(16:38):
billion, so that's 24% aboveexpectations from just about
three months ago.
We'll call it Google $75billion in planned capital
expenditures this year.
That's 32% above expectations aquarter ago.
Meta $60 to $65 billion that's30%.
(17:00):
So, david, here's the bigpicture.
As a group, this collection ofcompanies is now expected to
spend more than $320 billionthis year.
That's above what the 2026expectations were just months
ago.
So we'll talk later on howmaybe some of this bullishness
(17:20):
that you would expect to reallyhave risen in a lot of the data
center stocks hasn'tmaterialized and maybe some of
the potential reasons for thatof the data center stocks hasn't
materialized and maybe some ofthe potential reasons for that.
But as far as what this lookslike for the continuing
investment across AI datacenters, this is pretty bullish
stuff.
Now you mentioned I'm adding$100,000 in Meta and Amazon and
(17:45):
we don't really get to dive intothe details on that last
episode.
So I want to take just a momentto say why these two companies
in particular are the ones Iwould pick from this larger
group.
If you're picking from kind ofthe magnificent seven, which
ones do you want to allocatemore into?
For Meta, number one, thevaluation is becoming less
(18:05):
favorable.
We were talking earlier.
I believe they are now slatedtoday, friday February 14th, to
be up for the 20th consecutiveday, which is a new S&P 500
record.
Clearly the story is out.
They're up now.
I believe it's 22 or 23% yearto date.
So you know it's becoming lessand less attractive.
(18:27):
But there's a reason for thisstory coming out and the bottom
line is no one's made more moneyfrom AI, as meta AI has made
their products and advertisingsignificantly better.
The estimates I read itcontributed close to two thirds
of the revenue growth last year,and that's from content
(18:49):
recommendations, more time onapps.
It's from ad targeting.
It's also from AI optimized ads.
One thing generative AI andkind of transformer models is
really good at is understandingthe context of why a specific ad
works, which in the past, withold recommendation engines, you
didn't have, and it's a realgame changer for efficiency.
(19:12):
Moving on to the next generationof AI, meta, as you mentioned
earlier, they've got a leadingopen source model.
They are in the game, at thecutting edge.
They are able to kind ofdetermine their own future.
They have a natural entry pointfor agents and the rise of
agents via messenger and David.
They've got a sea ofproprietary data.
(19:33):
What are the most powerfulassets in the world Once AI
rises, if it reaches thepotential?
We think?
Well, it's unique data and it'sdistribution.
What company has both betterthan AI?
What?
What company has both betterthan ai?
Um, if you know, if ai enableslimitless content, you, you want
(19:54):
to be the central distributionpoint that people are going to
share their creations, andthat's exactly what meta is.
So you know, david, I mentionedthis is a portfolio hedge in
part two that if spending goesdown, um, it only it, meta.
The reason that they wouldprobably drop the rest of the
year is because they get tooaggressive in their spending.
That would be the number onethreat to the stock.
(20:15):
So I am keeping it as a bit ofa hedge and, as I know, I'll
invest 100,000, but I willdecrease these positions as well
, as we're probably makinginvestments, because we are
getting pretty close, aftermaking these large investments,
to being fully allocated on the$500,000.
I don't know if you've got anymore follow-ups on Meta or if
(20:36):
you just want to move on toAmazon.
Speaker 1 (20:38):
Let's jump into Amazon real quick, and then we
can kind of put a bow on both ofthem.
Speaker 2 (20:42):
Yeah, and Amazon.
You look at their upcomingearnings expectations and it's
just very smooth growth.
It's $68 billion this year, $82the next, $104, the year after
that, $128 after.
And the reason is theiroperating efficiency has been
outstanding.
David, I pulled the numbers.
They basically have two profitengines right now their
(21:05):
advertising business and AWS.
So if you take out ads and youtake out their operating profits
from AWS in 2021, they madenegative 18 billion.
In 2022, negative 48 billion.
That's when the stock hit itsnadir.
Since then they've kind ofturned the course around 2023,
(21:26):
negative 34.7, 2024.
This past year, about negative30.
So they're getting a lot moreefficient.
So I think one thing too isI've talked I'd like to add more
robotic stocks to the portfolio.
It's just when that market'sgone.
You can't force it and investtoo early if there's not
(21:47):
investable ideas.
I think Amazon has more to gainthan any big tech company from
robotics and the efficiency thatthat's going to lead to.
It already has led toefficiency, but the kind of
inflection point they're goingto see is going to be incredible
in terms of drivingprofitability and making
e-commerce as big a profitengine in some ways as there are
(22:08):
other business units.
Speaker 1 (22:09):
So you're saying that 30 billion loss.
You're basically taking out AWS, taking out advertisers, You're
looking mostly at the retailbusiness.
So you're saying you can takethat 30 down to 15.
That's essentially like addinga $15 billion profit business.
Speaker 2 (22:24):
Yeah, yeah, I mean, and it is just Amazon there they
have.
You know, it's when you look atthese companies and you go, how
do you get bigger?
You're a $2.4 trillion company,which is what Amazon is, and
it's like, yeah, but if youactually have a really good path
to grow, sales maybe only eightto 10%, but profits 20% plus,
(22:45):
you can continue, uh, you cancontinue being a great
investment from a PE in the 30s,which is what Amazon is going
to be capable of.
And, david, the other thing isAWS.
We think the biggest and mostintelligent models they're going
to live in the cloud.
We think AWS was behind.
(23:06):
They're catching up.
They've got a pretty aggressivegrowth map.
And the other side of this ispost-DeepSeek.
If AI models become commodities, the number one winner from
that is probably Amazon Because,again, they haven't been
involved.
Aside from some investments incompanies like Anthropic, they
haven't been involved in theincreasing spend to build their
(23:26):
own models.
So if everything becomes acommodity, that just makes AWS
even more valuable.
So you know, david, that's kindof the pitch on these two
companies Outside the AIportfolio.
I've said this many times sincewe talked about them.
I hadn't originally added thembecause I thought a lot of
people already own them.
We're looking for smalleropportunities, but I do in my
(23:48):
own personal investing.
These have been two of mylargest stocks for years and I
like the way that they balanceout the portfolio, so these
position sizes will probably getsmaller, but that's why we're
starting at a larger positionand, as I noted, we'll sell this
down to get them.
It wouldn't surprise me if, insix months, we had something
closer to 50K of each of thesestocks.
Speaker 1 (24:11):
Okay, so let's move on to those are the hyperscalers
.
Again, it feels like theconsensus has been very positive
.
You talk about 20 straight daysof positive gains.
I don't know what Amazon is,but I know Amazon over the past
month.
You know it's still.
It's positive.
It traded up.
Even after DeepSeek it didn'tseem as affected as obviously
(24:32):
NVIDIA had the massive sell-offand a lot of our other kind of
data center plays.
We've also gotten earnings fromsome of those smaller companies
in the data center where, again, it's been a wild couple of
weeks for them.
They had massive, massivesell-offs after that news.
Then we saw some jump back.
I can't recall the company youtalked about last episode that
was down 40% and thenimmediately made it all back in
(24:54):
two weeks.
We've had some of that, butwe've had some of these
companies that have sold off andhave certainly not recovered to
where they were pre the deepseek story.
So what have we heard fromthose companies?
It does seem like the consensusis not as cheery in that corner
of the market, but also not asthe world is ending and these
companies all growth prospectsare dead.
(25:15):
So what have you heard fromthese smaller companies in the
portfolio that had been rocketships but have hit a little bit
of a stumbling ground now.
Speaker 2 (25:22):
Yeah, in the short term, most of its stocks,
whether or not it does good orbad, isn't going to be driven by
results per se.
It's what you call narrativeand it's whether or not it's
multiples on earnings or salesor contracting or expanding.
So I think what we have rightnow is, with DeepSeek, you had a
(25:43):
very risky an event that showed, maybe you know, these stocks
are very risky and you've had alot of people want to kind of
branch out to different areas.
Now the question is, is that ashort-term bump or is that a
longer-term concern?
So let's drill in on.
You know, if I was talkingabout our data center stocks,
(26:03):
you know, in optical we've gotCoherent.
You know, in optical we've gotCoherent, lumentum, fabrinet,
we've got companies like Sienna,then we've got, you know, other
plays Credo, vertiv, schneiderElectric.
So we have a good collection ofthese stocks and, as you noted,
they are probably the aspectweighing down the portfolio the
(26:23):
most right now because they aretrading about where they were
after deep seek.
So let's zero in on coherence alittle bit more of a bellwether
, especially in the opticalspace.
They report earnings.
The earnings were up 252percent from last year.
I I trust you agree, david,that's pretty good.
(26:45):
Um, well, just just just, uhbeat expectations by a tad.
I think it was something like93 cents a share of earnings
versus 69 cents.
Wall Street was expected.
They were up 14% the nextmorning, which is kind of
expected reaction when you havethose kinds of earnings and
you're beating outlook for nextquarter.
(27:05):
They got to about $100 pershare.
But, david, what we've seensince then is shares have kind
of just every day just kind ofdropped 2% or 3% to where it's
trading almost exactly kind ofmid-80s back where it was
post-Deep Seek.
So the question is what's goingon?
There's a lot of factors here,but I think one issue is right
(27:29):
now NVIDIA is just such a largepercentage of the spend around
these kind of AI data centersand what was expected was that
NVIDIA was going to see a verymassive ramp of a specific
product called GB200.
And this is the one that's beenin our production.
Now, david, if you wanted tofollow every single bit of
(27:49):
speculation and people sayingthat there's engineering
problems or this and that withGB200, you'd go insane
scuttlebutt coming out of thesupply chain.
(28:10):
But it does seem like what'sgoing on right now.
Is that the size of the rampfor this specific product is
being curtailed, expectationsare coming down and a lot of the
demand is moving to the backhalf of the year because NVIDIA
was very aggressive aboutcreating a successor product,
the 300 series.
That is a vast improvement andisn't a significant weight.
So I think what's going on isthat's causing a lot of chop and
(28:33):
it's causing a belief thatthere's going to be kind of an
air pocket for a lot of thesecompanies and I think investors
are kind of trying to get out oflooking stupid in front of some
near-term disappointment, forlack of a better word.
So my bottom line, david, Istill think that these are good
companies.
I still like their kind ofpotential, especially across the
(28:57):
entire year.
I think they're just goingthrough some chop right now,
especially related to thespeculation One area that's
doing great we've talked aboutin the past.
We're going to have all thesebig data centers.
You're going to need to getthem to do kind of coherent
training together.
You're going to need to networkthem all.
Sienna is a big winner fromthat, but also Coherent itself
(29:19):
and Fabrinet are winners.
That's how the business isdoing great.
It's just a lot of the businessthat's kind of blowing through
from Nvidia is seeing a bit of apause, so I did add Fabrinet.
Right now it's trading forabout 20 times earnings ex cash.
It's taking a blow from maybelosing a little content in
Nvidia, but it's got some newcustomers, including Sienna,
(29:42):
that are going to be great forthe long-term story.
So, again, these stocks aredown right now, but if this was
a part of the portfolio youhadn't added to, I would say
it's a good entry point for alot of these stocks, taking a
bit of a breather from how hotthey had gone.
We added Fabrinet.
That one continues to be one ofthe cheaper options in terms of
(30:03):
PE, but I haven't seen anythingthat changes my long-term
opinion.
I think we're just kind ofworking our way through a bad
narrative.
Now there are some sectors ofthe AI space with good
narratives, which is what we canget to next right.
Speaker 1 (30:18):
Yeah for sure.
Yeah, I was just going to note30 minutes into the podcast, I
don't think we've really talkedabout NVIDIA, which must have
been a record.
And you're talking about allthe spend from Meta, amazon,
again, everyone, all the spend,not all of it, but a huge chunk
going to NVIDIA.
We should note NVIDIA stockobviously had the huge sell-off.
It's back to flat year to date.
(30:39):
It's recovered a little bitfrom the deep seek story.
It's performing a little bitbetter than some of these
coherence or some of thesesmaller plays.
So again, what we've talkedabout since we've started the
show is that there might bethese moments of air pockets and
shop and it seems like themarket reaction is appropriate.
If NVIDIA is flat, it makessense that a coherent is maybe
(31:01):
down 15%.
It's not the same size ofstocks, same size of company.
So I don't want to get intoNVIDIA.
We'll talk about it.
We've talked about NVIDIAenough and I think we have their
earnings coming in a couple ofweeks, is that?
Speaker 2 (31:15):
right, yeah, they'll get some coverage.
Speaker 1 (31:17):
We'll get some coverage then.
So any more earnings you wantto hit on?
We had Cloudflare, which is anew company that we added to the
portfolio.
Reported earnings Positivereaction there.
Is there anything you want tohit on Cloudflare there?
Speaker 2 (31:29):
Yeah, and it just kind of shows how much in the
near-term narrative drives shareprice.
We had Cloudflare before itsearnings report.
You know there's anotherpodcast I'm sure people out
there listen to it Bill Simmons.
They do the guess the linewhere you say a football game
and you try and guess who'sgoing to be favored.
I was doing a guess the line.
When I saw Nets earnings kindof the numbers break I said ah,
(31:51):
it's down 10%.
I believe they are now up 20%and it is just partially.
This company has just such aphenomenal story right now.
It's.
You know, david, if you know.
Once again we'll kind of archivepast episodes of the podcast.
But in one of the earlier onesyou would ask me what's a stock
(32:13):
that you love but it's just alittle too pricey right now to
add to the portfolio.
I said Palantir because we hadstarted last year publishing a
report on AI.
We had picked out Palantir asone of our favorite
opportunities and just keptgoing up and up and around the
time that we had recorded thatepisode it was trading closer to
I believe it was about 40 timessales.
(32:35):
I said it's just a bit too upthere for me to touch at this
point.
It's tripled since then.
So I wish we would have added.
I still personally hold it.
I haven't made any changes to mypersonal holdings, but the next
best story in terms of howcleanly you can take a company's
(32:55):
business and kind of draw it totrends in AI after Palantir is
probably Cloudflare.
So this is an expensive stock.
I think when we added it it waswhen we were before earnings.
When we were looking to add it,it was trained closer to about
20, 25.
It was trained closer to 25times this year's sales.
(33:16):
It's even more expensive now.
And the reason, dave, isCloudflare is what you call an
edge network.
I think it's not like a greatbusiness.
You probably remember Fastly,which was doing content delivery
.
Unfortunately, I do.
Yes, everyone loved it in 2020,but you know again, it just it
(33:36):
wasn't a great business becauseif you start, you know what
you're doing is kind of buildingservers that are located closer
to people.
If you're not able to buildmore value-add services on top
of that and you know, do theclassic land and expand, it's
just not going to be a greatbusiness.
But that's exactly whatCloudflare has done.
They have tremendous execution.
(33:56):
They've moved from contentdelivery and what you'd call
denial of service attacks tobuild a much bigger security
business.
And what they did was theybuilt a developer platform which
you can run services closer tothe consumer, which was a pretty
niche concept until AI took offand, all of a sudden, training
(34:19):
you can go do a training runanywhere.
It's not that big of a deal ofneeding locations closer to
consumer With inference latencymatters a ton.
So all of a sudden, thesedeveloper services that are
offering a network that'sextremely close to users has
incredible value.
They call it workers AI andthat's the service, once again,
(34:42):
for running inference throughCloud Flare.
The other thing that you have issecurity.
People have asked us to talkabout security.
Security is really taking off.
You look at the charts of allthese security stocks.
It is up and to the right rightnow and that's because of a
realization that security is anarea where being a leader maybe
gets its next advantage.
(35:02):
Because, again, we've talkedabout unique data being the
biggest key to AI and the moreattacks that you're seeing by
being the leading vendor, themore data you have to be able to
build a better product.
So it's kind of aself-reinforcing cycle people
are now buying into.
And finally, with Cloudflare youknow, david, again with our
business.
When AI first started, when AIwas really taking off after
(35:27):
NVIDIA's famous earnings reportin early 2023, when I was
building out my initial AIportfolio, one of the first
stocks I went to was Cloudflare,because I was like, if there's
all this traffic from all theseAI agents and also at a certain
point you're going to need toprotect your data against
basically being trained for freeby these models, the number one
(35:48):
company to fill that gap isactually Cloudflare and that's
what they're moving into rightnow.
They're just starting it where,essentially, you can protect
your content, you can charge arate to be able to be ingested
into models, and that's whatCloudflare's next big business
opportunity is going to be.
So, david, it's an extremelyexpensive stock but, again, this
(36:10):
is probably the cleanest storyI would call it in terms of how
their products really lead intothe growth of AI right now.
You saw Palantir the past year.
Them seeing I don't know whatit is it's probably like 600%
growth from when they were at$20 near the beginning of 2024
to today.
They didn't require going from50% growth to 150% growth rates.
(36:34):
They grew their top line about20% to 36%.
I think Cloudflare as we see alot of growth in inference from
the emergence of these reasoningmodels.
We'll see a similar expansionon top line and I think it's
going to be a stock that peoplemost closely associate to the
growth of AI.
(36:54):
So I don't think we want aportfolio full of these
companies, right, let's?
Let's call a spade a spade, astock selling for a sales
multiple.
This high is risky.
For a sales multiple.
This high is risky.
But I also think, if you arewanting to take some chances on
having stocks that kind ofdefine this generation, I think
(37:14):
Cloudflare has a space in yourportfolio, which is why we added
15,000.
And I think there's anotherhuge narrative in AI that we're
about to add software stocktoday that I think will also be
a leader in one of the spacesthat's going to attract
significant media attention thisyear.
Speaker 1 (37:35):
Yeah, I want to get to the new company.
But just to your point on highvaluation multiples, high price
to sales, high PE.
If there's any lesson learnedover the last couple of years
with some of these companies inthe AI space, it's be careful.
Judging only by those valuationmetrics is right before an
inflection point on the top lineor even on earnings.
I don't know what they're goingto get to.
In today's show we had Applovin, everyone's favorite named
(37:58):
company report earnings Again acompany that, when you look at
just the amount of cash they'rebringing to the bottom line
because of AI, it's juststaggering the growth that some
of these companies can see.
I think Applovin was up 700%last year and it was up 33% in a
day this week.
So very risky.
(38:21):
Just to say it's expensive,I'll never buy it.
You can often get paired withexplosive growth on the business
side, which very quicklydoesn't make it quite as
expensive.
Speaker 2 (38:25):
So I do want to talk about Applo, and so let's come
back to it, but let's do thestock we're actually buying
first.
Speaker 1 (38:31):
Sure, let's get into the purchase.
We are adding a new company tothe portfolio.
Let's kick it off.
Speaker 2 (38:35):
Yeah, the company I'd like to add today is GitLab.
I think we'll follow a $15,000purchase In terms of some of the
leaders in companies at theforefront of software trends.
That will give us a cloudplayer and you know security and
inferencing.
It will also give us Snowflakeand we'll have GitLab beyond
(38:57):
also other softwareopportunities like Synopsys and
Cadence.
So we are starting to build outkind of the software side in
addition to, you know, many ofthe data center plays we had.
Now the storyline here is, davidI'm sure you've seen it right
there is a chart making therounds on social media right now
that shows developer jobopenings from 2022 to today, and
(39:21):
it's absolutely fallen off acliff, right, and you know the
the reason many will attributeto this.
There's several reasons, butright now there are coding tools
utilizing AI that make peoplephenomenally more efficient, do
some stuff.
That is honestly kind ofshocking.
You know, david, we met in LasVegas, as we talked about
(39:43):
earlier, and, talking with thedevelopers on our team, their
embrace of AI has completelychanged the way that they do
their jobs.
Not only that, but another oneof our co-workers he was showing
me yesterday that what he callsvibe coding.
You know if you get in too much, you'll see a lot of people
talk about this vibe coding now,where you're working more with
(40:07):
ideas than the code itself.
He had taken some that we mightpreviously had to pay someone on
Upwork maybe six figures, maybe$100,000, to build some
concepts, and he had just doneit all in a spare time with a
week knowing almost no, havingalmost no ability to code.
Let's just call it is what itis.
So the preeminent productthat's getting most of the
attention in Silicon Valley issomething called Cursor.
(40:29):
It's a code editor that set thenew record, as the PR would go
and going from 1 million to 100million annual run rate in 12
months.
I've seen people say that, uh,when they're talking with
classes at in Stanford, uh,computer sciences, if they ask
how many are using cursor, it'sa hundred percent people.
Speaker 1 (40:49):
Wow.
Speaker 2 (40:50):
Um, if you look at Y common air one of the best
incubators for new startups ahundred percent usage uh, cursor
right now.
So if you're at the bleedingedge right now, the adoption for
Cursor is maybe at the highestlevels of any tech product ever.
So, being able to code with AI,it's a big deal, and these
(41:12):
reasoning models where theyperform best is areas like
coding, because that chain ofthought gives them objective,
true or not true statements andit's really going to turn coding
into a superpower extremelyfast via AI.
So I think the growth rate inthese areas is going to be
absolutely incrediblegrowth-raising capabilities.
(41:35):
So there's a space named DevOpsand that's for code management.
It's a great first place tolook in terms of where you could
see, you know, some gains fromthis trend.
There's a company named GitHub.
It's very strong, but it'sowned by Microsoft and you know
Microsoft's very big.
It's just a small piece of itsbusiness.
So we're looking at their topcompetitor, which is GitLab.
(41:57):
They have a lineup named Duo.
It's not just coding assistants, but it's agents that monitor
across the entire workflow ofall company code.
It can not only function areaslike security, but also code
optimization.
So Cursor it's going to be usedheavily by startups, first
movers.
(42:18):
But if you're in a heavilyregulated space and you're
looking for a lot of theseefficiencies, a company like
GitLab is exactly where you'regoing to look, and they have a
ton of differentiation here.
So you think about Bank ofAmerica.
Gitlab actually specializes inbeing self-hosted, so you don't
need to go to an LLM out in thecloud or untrusted spaces.
(42:40):
You can actually keep thingslocal and use GitLab.
So, again, you think about allthe industries you could think
about healthcare, finance,energy, national security.
This is exactly the kind ofproduct they're looking for.
And when it comes to actuallyhaving relationships in the
cloud, well, I mentioned theirmain competitor, github, owned
(43:01):
by Microsoft.
Well, that's an opportunitybecause it means GitLab can
partner more strongly withAmazon and Google.
So, David, this is a companythat's been growing among the
higher growth rates in the SaaSspace.
It was 31% last quarter.
It's expected to decelerate to25% this next year.
If there's a big push to lookfor solutions in being able to
(43:26):
find agentic coding assistanceacross your code base, they're
going to be the first companythat a lot of people look to.
If they bump up that rate from25 expected back to 30, it's
very likely going to be awinning investment.
They're trading at 12 timestheir expected sales this year.
That's high, but it's also notoutside a band of SaaS companies
(43:49):
.
That one with one of the bestopportunities is a problem.
So there's some risk here,david.
It is a little more expensive.
It does have good competition,but in terms of getting a
company that's aligned with whatI see as one of the most
exciting trends and somethingthat gets a big push from the
recent breakthroughs andreasoning that maybe the market
(44:10):
has internalized, I really likethis play and I think there's a
good chance that we look back ina year or two and GitLab is one
of the stronger portfoliocompanies, right.
Speaker 1 (44:21):
Yeah, when you talk about Microsoft and GitHub, you
have a little bit of PTSD fromhearing Microsoft as a
competitor if you were a Zoomshareholder, or just the ability
for Microsoft to come in andjust kind of dominate a
competitor.
So is that the biggest outsideof valuation?
Is that the biggest risk?
(44:41):
Is that they just get squeezedon pricing, or is that what I
should be most worried about ifI'm thinking about this company?
Speaker 2 (44:48):
I think so.
If I needed to say what was thesingle biggest risk that
concerns me when I look at theircustomer acquisition, it's been
a little weaker in recentquarters.
And if you want to, you knowyou're fundamentally in SaaS to
land and expand and if you'renot landing you can't get people
on these tools that we thinkcould be the growth sweetener
for the company.
Hard time landing in thiscompetition with Microsoft,
(45:18):
which doesn't allow kind of thisproduct suite to give that
incremental growth we thinkmakes this a winning investment.
Speaker 1 (45:23):
Okay, so that was GitLab.
You said 15,000 routing earlierthere.
Speaker 2 (45:28):
Yeah, and one note, just a minor programming note,
because I have been battling thefeared Las Vegas flu and
traveling.
We haven't been able to do theSK Hynix purchase.
Everything else that we'vetalked about has been purchased.
We haven't been able to do that.
Next week should have an update.
I will make sure that thespreadsheet we track all this I
(45:51):
need to get updated Again.
That's just been a recentcasualty of not coming out of
bed for several days in a row.
But we will make sure to getthat update.
And you know again, we'llpublish more portfolio review
pieces, I believe in the comingweeks, since you know we're
trying to catch up with a lotright now.
Speaker 1 (46:10):
We fully realize that , yeah, that sounds good.
Yeah, and we actually we have afew listener questions.
I think we're going to bumpthose.
Maybe we'll do special YouTubevideos on answering these
listener questions becausethey're running a little bit
long on time.
And I did want to circle backto Applovin, again a company
that is probably more onpeople's radar than it was, I'm
guessing, 18 months ago.
It was on no one's radar andnow this is a $200 billion
(46:31):
company.
So I did want to get your quickthoughts on Applovin.
What should we make of it?
Again, this is a company thathas been utilizing AI in a
different way.
It's not a data center stock,it's not NVIDIA, but it's been
the most successful kind of techcompany over the last 14 months
now.
So what were your thoughtscoming out of just monster
earnings from this company thisweek?
Speaker 2 (46:49):
Yeah, and just quickly on the listener
questions, there's one fromSaltyCuddy707.
He left in a five-star review.
That's amazing.
Anyone leave a five-star review?
We'll answer your question,salty.
We will put that on our YouTubequestion.
You wanted our thoughts onPhotronics, which is a photo
mask manufacturer.
We also had one from JohnMcGuire on our YouTube channel
(47:10):
Asked about SteraLabs.
He asked on YouTube.
We'll put it there Nowlovin.
I think this is just fun.
So I want to talk about thisbecause it kind of shows how
kind of emerging ideas function.
Applovin, you had noted this isone of the story stocks of the
past year, when we were talkingabout companies with unique data
(47:36):
and sometimes the pivots thatthey're able to make and how the
market rewards that.
Apple 11 was a company I hadowned.
To be honest, I had sold, inNovember, I think, around $330
because I thought they were kindof reaching the end of their
opportunity in the video gamespace.
They just reported earnings areup 30%.
(47:57):
They're close to $500 per share, because now they're trumping
an e-commerce kind of push.
So you know, I don't know ifthat will be persistent, but the
point is, the market reallylikes what they're doing.
Now.
There might be anotheropportunity here.
David, though, and I kind ofwant to walk through it.
You know, I just I was justlooking at this on Thursday, so
(48:20):
this is like real time thinkingout loud for the audience.
I know people out here probablylove trade ideas, so, again,
applovin up 30%.
What I've said about Applovin inthe past is they're kind of
like Google if they didn't giveevery employee free dry cleaning
and filet mignon dinners everysingle night.
(48:41):
You know, this is just acompany that takes dollars on
the top line.
They convert to profits on thebottom line at an incredible
pace and they have unique kindof mobile gaming data.
They've utilized an aggressivepush into AI ad tech, and
they've done that to getphenomenal growth.
And you might say how themarket missed this, that the
(49:05):
company is up more than athousand bolts since the start
of 2024.
And the reality is the marketmight not have fully missed it.
Their greatest competitor is acompany named Unity that was
worth close to $100 billion in2021.
But, david, the opposite isUnity.
They gave away a lot of freedry cleaning, a lot of filet
(49:27):
mignons.
This is just a company that hadincredible expense levels, so
they weren't able to kind ofachieve the same areas as
ZappLove and also their mainbusiness is ad tech, but they
make a game engine and they justdidn't make the same pivot to
focus on applying a lot of thesebreakthroughs in AI to their
(49:49):
advertising technologies.
And, david, what's interestinghere is you look at the track
record.
We talked about Meta earlier.
They were using CPUs instead ofGPUs from NVIDIA.
They did a huge build-out to doall their content
recommendations through GPUsinto 2021.
They started flipping theswitch with all these new data
(50:12):
centers they built in late 2022.
If you go and you look at stockcharts and they're close to a
thousand percent since thebottom app loving was a fast
follower.
They were doing press releasesabout how they were, you know,
moving their um ad tech to youknow again, these, these newer
transformer technologies in in2023 and they're up more than
(50:35):
1,000%.
Unity today is worth $8 billion, so about four times sales.
Apple opens at 34 times sales.
So what's interesting here,david?
A company that clearly hasn'tbeen able to keep up with
competitors but this week,buried in the news, they're
doing a massive reorganizationwhere they are reorganizing
(50:56):
entirely around this ad tech andthey're going to have an
opportunity to kind of capturethe same thing that's produced
those almost generationalreturns in those other companies
.
So do I think?
Do I think Unity is going tonecessarily be an automatic 5X
in the next year?
Not necessarily, but they havethis unique data that we've seen
(51:18):
other people utilize to getthese kind of generational
returns.
They're making the shifts intheir business and this is a
company that, again now I'mwondering if the time is, you've
been a loser if you've beenbuying any time in the past few
years, but this might be thepivot point where you are
healthily rewarded for it.
So you know, this is where wesee an app loving and a company
(51:41):
that's producing crediblereturns, and maybe we won't buy
that.
But by following the story andunderstanding it, it might lead
you to discover some other gems.
Speaker 1 (51:51):
So I thought that was oh, go ahead, David yeah.
I was gonna say I think thiscould be a very interesting case
study of there's a company thatlaid out the path to success
with AI, and it's a question ofexecution, of how hard is it to
go and execute these ideas.
It's easy for you know someoneto get on a conference call and
say we're going to do that too,but now let's see if they
actually do.
(52:11):
I think it'll be fascinating tofollow.
So that is a story withApplovin and Unity.
I think that's all we got fortoday, right, anything else?
Speaker 2 (52:19):
Yeah, like you said, we'll answer the questions on
our YouTube channel.
If you aren't subscribed there,go ahead and subscribe to 24-7
Wall Street on YouTube.
We'll get those handled.
If anyone else has anyquestions, send them our way.
And, like we said, there's beenenough news.
We would have liked to havedone a show earlier, but maybe
(52:42):
I'll return to my sleep holeafter we finish recording Go get
some rest For Eric Flieger.
Speaker 1 (52:47):
I don't buy or sell based solely on what you hear.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
The Breakfast Club
The World's Most Dangerous Morning Show, The Breakfast Club, With DJ Envy, Jess Hilarious, And Charlamagne Tha God!