¿Empiezan Los Problemas En La IA? — Transcript
Full transcript
- 0:01We continue with results from
- 0:03interesting companies. Today it was
- 0:06Broadcom's turn, as well as NetApp,
- 0:08Snowflake, and Brown-Forman. We are
- 0:12seeing significant moves in the
- 0:14aftermarket, so let's review the
- 0:16numbers, look at the accounting, and
- 0:18see if there might be good investment
- 0:20opportunities in these companies. And
- 0:24let's start today, our review of the
- 0:26day, with Broadcom's numbers, naturally
- 0:28. One of the largest semiconductor
- 0:31companies in the world. It is number
- 0:34two by market cap, right after Nvidia.
- 0:38Broadcom is number one in custom chip
- 0:40design, also chips for network
- 0:42connectivity, and it also has a very
- 0:45important virtualization software
- 0:47branch. Well, Broadcom semiconductors,
- 0:54in this case semiconductor design, is
- 0:57skyrocketing. Look at how in terms of
- 1:02revenue we are up 86%for the quarter.
- 1:08And if we go to adjusted net income, we
- 1:12are up 95%, with operating profit
- 1:15rising 92%. The company is growing
- 1:21tremendously, beating expectations, and
- 1:24certainly Broadcom's sales and profits
- 1:27are going at full speed, like most
- 1:30semiconductor companies in general.
- 1:35However, although sales and profits are
- 1:37doing very well, we are also noticing
- 1:39that in recent months the market is no
- 1:41longer as excited about good results.
- 1:45There has likely already been a lot of
- 1:48expectations priced into semiconductors
- 1:50, and right now for there to be a
- 1:52premium in semiconductors, there has to
- 1:55be a huge beat of expectations or a
- 1:57clear guidance upgrade, which we will
- 2:00see now hasn't happened with Broadcom.
- 2:03Even so, we will see the valuation now,
- 2:06right? But it is a great company, a
- 2:09chip company that is not very cyclical.
- 2:13As you can see, we have it at record
- 2:15sales, record profits, a very beautiful
- 2:17company in terms of its income
- 2:18statement, revenue growth, profit
- 2:20growth, and good margins. As I say, it
- 2:24is the undisputed leader in network
- 2:26chips and also custom chips. It
- 2:28manufactures chips for Google or Meta,
- 2:30for example. It doesn't manufacture
- 2:33them, it designs them; Taiwan
- 2:34Semiconductor manufactures them. There
- 2:36is no other manufacturer like Taiwan
- 2:37Semiconductor, and well, then there is
- 2:39Samsung and a few others, but really,
- 2:40Taiwan manufactures. The rest of the
- 2:44companies are semiconductor designers.
- 2:48Well, Broadcom is at record sales,
- 2:50record profits, continues to grow very
- 2:53strongly, continues to grow, and if
- 2:56sales continue to grow and profits
- 2:58continue to grow, then logically the
- 3:01stock will continue its upward trend.
- 3:04The interesting thing is that we are
- 3:07seeing recent drops; we are seeing some
- 3:10contraction across all semiconductors
- 3:13in recent months, and this could start
- 3:15to present opportunities. I think
- 3:18it’s generally still too early to buy
- 3:21chips. You could hold something like
- 3:24Taiwan Semiconductor, like Nvidia,
- 3:26which is something very, very good,
- 3:28very, very special, and at good prices,
- 3:31though they must keep growing strongly,
- 3:33but they are very special assets.
- 3:35Nvidia and Taiwan Semiconductor. I also
- 3:38like the case of Broadcom, but it isn't
- 3:41as special a company as Nvidia. But
- 3:45look, we’re seeing drops of 23%from
- 3:48the highs. It’s not a bad drop at all
- 3:50; it’s a significant decline. In fact
- 3:53, it’s one of the largest drops in
- 3:55its history. It’s hard to believe,
- 3:58but we have some chip companies seeing
- 4:00historically significant declines.
- 4:03It’s true that 23%isn't a massive
- 4:06crash, but notice that in other periods
- 4:09of history, it has also had them, like
- 4:13during COVID, in 2023, 2016, and 2018,
- 4:17declines that hovered between 20 and 30
- 4:19%. Maybe at some point, at the peak of
- 4:232020, right? At the maximum drop, we
- 4:27could reach a decline of over 40%.
- 4:31Obviously, it recovered quickly, but a
- 4:3423%drop isn't bad; it's a relevant
- 4:36decline for Broadcom's history, and
- 4:38well, it’s a company that I think is
- 4:41very good and worth holding in a
- 4:44portfolio. I don’t own it, but well,
- 4:47I’m considering perhaps, why not?
- 4:51Getting in. We’ll see if I end up
- 4:53buying or not. In terms of multiples, I
- 4:56think we’re in a decent spot.
- 4:59Although, look here, we have a forward
- 5:01P/E of 23 which seems high, forward 23
- 5:04versus a median forward P/E of 14. Sure
- 5:07, one might say, it’s far from the
- 5:09median, I’d like to buy at a forward
- 5:10P/E of 14. Yes, but a forward P/E of 14
- 5:14corresponds to a time for Broadcom that
- 5:16will likely not return. Broadcom didn't
- 5:19used to generate as much revenue as it
- 5:22does now, nor did it have the growth
- 5:24expectations it has today. The chip
- 5:27design aspect, a new business area with
- 5:32artificial intelligence, largely new,
- 5:33which has allowed them to significantly
- 5:35increase their growth rates. We also
- 5:38have the acquisition of VMware, which
- 5:40has provided all that virtualization
- 5:41growth. So, Broadcom’s decade leading
- 5:46up to 2022—that previous decade with
- 5:48lower multiples and much more moderate
- 5:51growth—cannot be extrapolated to the
- 5:54era or decade we are starting now, with
- 5:56the explosion of artificial
- 5:58intelligence, where I believe multiples
- 6:01will be slightly higher. Even so, there
- 6:04could be issues, and buying at or below
- 6:07the median always gives you more
- 6:08comfort, but this forward P/E median of
- 6:1114 falls short for me. It seems like a
- 6:14low median, way too low. If we look at
- 6:18the most recent period with Broadcom's
- 6:20new businesses, we would actually be
- 6:23hovering around a median since 2022 of
- 6:25about a 23-24 forward P/E, and we are
- 6:28right there. So, I think Broadcom is a
- 6:31very good company. Of course it is.
- 6:34Sales and profits are through the roof,
- 6:36cash flow is through the roof, a record
- 6:37, a super historical record for cash
- 6:39flow. And the multiple is fine. The
- 6:45multiple is fine. I don’t mean to say
- 6:48it’s a steal, because it isn't.
- 6:50It’s not cheap, or it might be in
- 6:53line with a reasonable price, depending
- 6:58on how generous you want to be. I mean,
- 7:00if you assign it higher growth, then it
- 7:03will be cheap. If you assign it lower
- 7:04growth, it will be more expensive. It
- 7:06also depends on the price, or rather,
- 7:08the growth one assigns to a company.
- 7:10But here, we are going to grow at 66%
- 7:12this year, then 63, then 34, then 20,
- 7:14then 21. There is good growth in sales
- 7:18and there is also good growth in
- 7:21profits. We have net income growth of
- 7:2570%, 67, 33, 17, 14. There is good
- 7:29average growth; we will be at 20-30%.
- 7:32In other words, I think the P/E is
- 7:34reasonable, even a bit cheap. If you
- 7:37are generous, or if you are demanding,
- 7:39you will end up more in line with what
- 7:41is currently a fair price. Eh, right
- 7:45now you can see that the stock is
- 7:46falling 0.8%in after-hours trading. The
- 7:50results were good; it beat on sales and
- 7:52earnings, but there was no reward. As I
- 7:55say, we are not seeing a reward after
- 7:57results in the semiconductor sector,
- 7:58but that is because there was a very
- 8:00positive scenario already priced in.
- 8:03And when you price in a great scenario,
- 8:05you then have to beat it by a lot for
- 8:07the market to keep rewarding you,
- 8:08because you’ve already taken the
- 8:10prize. Broadcom has risen since '25, it
- 8:13has had a good revaluation, so well,
- 8:16now it is finding it a bit harder and
- 8:18it had a not-so-serious problem, but
- 8:21they lowered their guidance a little
- 8:23bit. We are seeing it in some
- 8:27semiconductor companies, uh, some
- 8:30aren't accelerating, they are even
- 8:32decelerating. And this, at any moment,
- 8:36if the market gets nervous that data
- 8:38center investment is easing, slowing
- 8:41down, capex is dropping and fewer chips
- 8:43are being bought or there is a chip
- 8:45surplus for some reason, which could
- 8:47happen, obviously because we are seeing
- 8:50overcapacity or overinvestment and that
- 8:52will lead to overcapacity; there is
- 8:54very likely overinvestment, clearly,
- 8:57with generally very high semiconductor
- 8:59prices. Well, we'll see. How does
- 9:04everything arrive? Notice that for the
- 9:07next quarter we had guidance of 34.8
- 9:10billion and the previous figure was 35
- 9:13billion. We’ve had a small decrease
- 9:18in guidance for the next quarter. This,
- 9:22well, hasn't been a very sharp drop,
- 9:24but it’s what the market must not
- 9:26have liked very much and there has been
- 9:28a small decline. Maybe the stock will
- 9:30go up tomorrow, but honestly, with such
- 9:33good results, such good growth, and
- 9:34having a reasonable price, the market
- 9:36could have rewarded it and it didn't.
- 9:39Also, another issue with Broadcom. The
- 9:44uh, the buybacks. We are at reasonable
- 9:48prices and there are no buybacks.
- 9:51They’ve removed them. Look at how we
- 9:53have it here under Repurchases of
- 9:55Common Stock. Absolutely zero. What do
- 9:57you think of that? Absolutely zero in
- 10:00this last quarter. Why? Nvidia is
- 10:05buying back strongly. Nvidia has a
- 10:08similar multiple to Broadcom and
- 10:10similar growth to Broadcom, and Nvidia
- 10:12is buying back strongly. Nvidia is
- 10:15buying back. Why isn't Broadcom buying
- 10:18back? They are relevant, large-scale
- 10:22companies, among the world leaders in
- 10:24semiconductors. Well, they should
- 10:27follow each other a bit. It’s true
- 10:29that Nvidia competes with AMD, it
- 10:30doesn't compete with Broadcom, but in
- 10:32terms of size, Nvidia is more in the
- 10:34style of Broadcom in size, but it is
- 10:36very large, it is one of the largest in
- 10:39the world. And then there is Nvidia as
- 10:43the biggest of all. Further down we
- 10:47would have the case of AMD, the case of
- 10:49Intel, the case of Marvell, the case of
- 10:51many others, but there have to be
- 10:55buybacks, there should be buybacks this
- 10:56quarter. Well, they’ve removed them
- 10:59and they’ve removed them, furthermore
- 11:01, without a justification that they are
- 11:03going to make investments, that they
- 11:04are going to reduce, well, they are
- 11:06reducing some debt. They are reducing
- 11:09debt, it is true, but so far this year
- 11:11they have reduced less debt than last
- 11:13year. So, well, they are reducing debt,
- 11:18but it's not like they have enormous
- 11:21debt either. Right now, the balance
- 11:24sheet shows 1.1 debt and 10 times
- 11:27EBITDA. So it's a solid company, sales
- 11:31and profits are firing on all cylinders
- 11:33, cash flow is great, but buybacks are
- 11:35being eliminated. There is a dividend,
- 11:39but since the multiples aren't low, it
- 11:40won't yield much. The dividend will be
- 11:43around 1%, maybe even a bit less. Less
- 11:46than 1%, it'll be a 0.7%dividend, plus
- 11:48the buybacks. In fact, they’ve never
- 11:53been impacted by SBCs. Broadcom also
- 11:56pays high stock-based compensation to
- 11:59its employees, partly due to the recent
- 12:02VMware acquisition, which was a
- 12:05software company that, well, brought in
- 12:08employees with SBCs, which is typical
- 12:11in software. So, look, there is
- 12:14dilution. There is dilution. So, look,
- 12:18it's a good company, it's a great
- 12:20business, a powerhouse, of course, with
- 12:23dips, and it's well-priced, and I think
- 12:25you can buy if you like it; I think
- 12:28Broadcom could be an opportunity. The
- 12:31buybacks don't add up for me; I don't
- 12:33know why they're cutting them when they
- 12:35have plenty of cash and controlled debt
- 12:36. And I think the multiple is
- 12:40reasonable, it's priced about right. So
- 12:45, I’d like to see, regarding chips,
- 12:48if AI finally explodes and we see a
- 12:50crash and the bubble bursts, that’s
- 12:52when we can really load up on chips. So
- 12:56, I think it's still too early to start
- 12:58buying chips, beyond Nvidia and Taiwan
- 13:00Semiconductor, which I find very
- 13:02special and well-priced. For the rest,
- 13:04I would wait a little bit. Even so, as
- 13:07I say, Broadcom is at a reasonable
- 13:09price; it's doing well in sales, doing
- 13:11well in profits, cash flow is great,
- 13:13but there are a few things that don't
- 13:16quite convince me, so I’m going to
- 13:18postpone the idea of buying Broadcom,
- 13:20though I don’t rule it out in the
- 13:23future if things improve or shares fall
- 13:25further. It’s a very good company,
- 13:29and in the long term, it has everything
- 13:31to keep prospering in its business
- 13:32areas. Of course, the custom chip and
- 13:36networking chip branches are super
- 13:37important for connectivity and the
- 13:39internet; it is the world leader in
- 13:41connectivity chips. So, well, an
- 13:45interesting company. Moving on to the
- 13:46topic of artificial intelligence. Today
- 13:50we had results from NetApp, a company
- 13:53specializing in hybrid cloud, a company
- 13:56that also deals with servers and
- 13:58physical data storage at the hardware
- 14:01level, a company that spends heavily on
- 14:04memory to provide memory to its clients
- 14:07, logically, and is currently
- 14:09experiencing, as we will see, issues
- 14:12regarding memory. In fact, NetApp, even
- 14:16though sales and profits are going full
- 14:18speed ahead, notice that after-hours
- 14:21NetApp is falling by 8%. The market is
- 14:26punishing NetApp in the after-hours
- 14:27session despite incredible results and
- 14:29improved guidance, but there is an
- 14:31issue with inventories that we will
- 14:33look at now. Well, notice how NetApp
- 14:36had 30%revenue growth. Superb. Sales
- 14:42are going very well. In general,
- 14:44everything involving artificial
- 14:46intelligence or that benefits from AI
- 14:48infrastructure investment is doing well
- 14:50. The problem is that this is already
- 14:53priced into the stock market. We have
- 14:55already had significant rallies, and
- 14:57until major drops occur, opportunities
- 15:00in the infrastructure industry probably
- 15:02won't be widespread. There have been
- 15:04drops in software, and it has been
- 15:06possible to buy. It will have recovered
- 15:10, but the key is always the drops,
- 15:12sharp drops, to be able to take
- 15:14advantage of the best opportunities.
- 15:18But well, we will see when AI
- 15:20infrastructure starts to fall, and I
- 15:21will be the first one to start buying
- 15:23things. Of course, absolutely, but
- 15:25let's look at NetApp's data in terms of
- 15:27valuation. Well, look at sales growing
- 15:30at 30%. And adjusted net profit growing
- 15:35at 64%. The company is improving
- 15:40margins and doing very well; what can I
- 15:42say, beating expectations, and NetApp
- 15:44is doing superb in terms of sales and
- 15:47profits; we have nothing bad to say. It
- 15:50is a company that can be somewhat
- 15:51cyclical, as you can see here due to
- 15:53the hardware aspect, but right now we
- 15:54are at record sales, record profits,
- 15:56and with good estimates. In fact, they
- 15:58have even improved their guidance. We
- 16:00will discuss that now. It is true that
- 16:04NetApp is a more traditional company in
- 16:06the industry and that it had been flat
- 16:09in sales for some years, trying to
- 16:11improve the cost structure to increase
- 16:13profits without growing revenue. This,
- 16:17evidently, has now been surpassed by
- 16:19the emergence of artificial
- 16:20intelligence. Record sales and record
- 16:23profits. Regarding NetApp, we mostly
- 16:25talked about it when it was in the $
- 16:27100 range, do you remember? It was
- 16:31trading at a P/E of 10, it was
- 16:32extremely cheap, just like Dell, which
- 16:34we discussed the other day, it was also
- 16:36around $ 100 at a P/E of 10 or 9.
- 16:40NetApp has quadrupled, well, it
- 16:43practically doubled, now it would be at
- 16:4660%after the aftermarket drops. But as
- 16:50you can see, we have a company that
- 16:52came from ridiculously cheap multiples,
- 16:55it was at a P/E of 10 and now we have
- 16:57it at a P/E of 20. The P/E has doubled
- 17:00and it is one of the highest P/E ratios
- 17:02in its history. Possibly NetApp will
- 17:04grow more in the future than it has
- 17:06grown in the past. It could happen, but
- 17:08it's a high multiple for its history
- 17:10and that isn't usually as appealing.
- 17:12Also, NetApp doesn't have great
- 17:13competitive advantages. It is a fairly
- 17:16traditional hardware company, of memory
- 17:20, of hybrid cloud, of data storage, it
- 17:23doesn't have huge advantages like
- 17:25Broadcom might have or Nvidia might
- 17:28have, of course, in AI infrastructure.
- 17:33Then in servers we have Dell, we have
- 17:35HP, which like NetApp, won't have huge
- 17:37competitive advantages. It is true that
- 17:39NetApp has a software branch, which
- 17:41provides a bit more advantages, but
- 17:43there isn't much advantage in the
- 17:44hardware. In software yes, but in
- 17:47hardware there isn't much. In
- 17:49generalist software, not in specialist
- 17:51software, not in specialized fields
- 17:53like Nvidia, right? Well, what is
- 17:56NetApp's problem, why are the shares
- 17:58falling 8%in the aftermarket and also
- 18:01let me tell you if it could perhaps be
- 18:03an opportunity to enter NetApp at these
- 18:06prices. Well, you can really see that
- 18:08even though there has been an 8%drop,
- 18:11we are still coming off a very
- 18:13important rise, so we aren't in total
- 18:15collapse mode or anything like that.
- 18:19But NetApp's problem, although sales
- 18:20and profits were very good, I want you
- 18:22to see here in the income statement how
- 18:24we have free cash flow falling 35%. A
- 18:3035%drop in free cash flow and this was
- 18:32not well received at all. And why do we
- 18:35have this drop in free cash flow? There
- 18:38has been a lot of capex, right? It is
- 18:39not a capital-intensive company. What
- 18:42there were were short-term investments,
- 18:44not long-term CAPEX, but short-term,
- 18:46which affected operating cash flow,
- 18:48which is what dropped and in turn
- 18:49dragged down free cash flow. As you see
- 18:53here in the cash flow, we have a figure
- 18:56that catches our attention, which is
- 18:58these 176 million in the quarter for
- 19:00inventory. These are payments for
- 19:02inventory that have not yet been
- 19:04collected. What does it mean? Inventory
- 19:07is piling up. They are paying in
- 19:09advance for inventory that they are not
- 19:12yet selling. What? Why is this
- 19:14happening? Because memory is so
- 19:17extremely expensive that these
- 19:20companies that live off reselling
- 19:22memory, or reselling systems with
- 19:25memory included, are hoarding inventory
- 19:28so they don't run out; they are pulling
- 19:32forward orders, and this is a major
- 19:34risk for the industry. Since everyone
- 19:39is pulling forward orders at the same
- 19:40time, there will eventually be a
- 19:42surplus of chips. They are
- 19:45over-accumulating inventory before
- 19:50selling it. They are generating
- 19:53overcapacity at high prices.
- 19:56Furthermore, if the time comes when
- 19:58they have too much inventory to sell
- 20:00and have to lower prices because there
- 20:02is once again an excess of chips on the
- 20:04market, prices will fall, as I say, and
- 20:06there will be problems here. That is
- 20:09what the market is starting to see, and
- 20:11I don't know if we are already at that
- 20:13turning point or if there is more to go
- 20:17, but we saw it in Broadcom and a
- 20:19slight drop in NetApp. inventories with
- 20:24advance payments to see if they sell or
- 20:28not. So, the artificial intelligence
- 20:32situation—of course it is a great
- 20:34technology and in the long term
- 20:36artificial intelligence is going to
- 20:38change everything, obviously—but in
- 20:41the short term there will be
- 20:42bankruptcies, there will be problems,
- 20:44there will be businesses ruined by
- 20:47artificial intelligence. There will be
- 20:50AI model companies that will go
- 20:52bankrupt, native AI companies that will
- 20:54go bankrupt, NeoCloud companies that
- 20:55will go bankrupt, businesses that will
- 20:57have many problems; they will have
- 20:59excess inventory, they will have excess
- 21:00chips, they will have to lower prices,
- 21:02memory companies that will have
- 21:04overcapacity and will manufacture too
- 21:05many chips, and the market will not
- 21:07absorb them. Hyperscalers will lower
- 21:10their capex at some point to try, well,
- 21:12to manage to make profitable everything
- 21:15they have invested and eventually
- 21:17generate free cash flow. Obviously, and
- 21:19for that, CAPEX will have to be reduced
- 21:21somewhat. Otherwise, this pace is not
- 21:23sustainable; the debt will have to be
- 21:25repaid. AI models will have to start
- 21:28making money, right? To be able to
- 21:31continue being sustainable, right?
- 21:35Without depending on external investor
- 21:37capital. So, we have a situation that
- 21:40at some point I think is going to
- 21:42explode, and that many investors are
- 21:44going to suffer, and that many
- 21:46companies are going to suffer as well.
- 21:48So, that is why it is time to be
- 21:50cautious. That is why I told you that I
- 21:53still don't see, even though there are
- 21:55falls in some infrastructure companies,
- 21:57that it is the time to enter, outside
- 21:58of Nvidia and Taiwan Semiconductor and
- 22:03if you push me, Shin-Etsu, the rest.
- 22:07Well, I don't think it's time yet to
- 22:09start building more positions in chips
- 22:12or artificial intelligence
- 22:13infrastructure. But, well, NetApp's
- 22:17results have been extraordinarily good
- 22:19and, on top of that, the company has
- 22:21slightly improved its guidance. We have
- 22:24previous guidance, we have it right
- 22:26here, right? Previous guidance, um, for
- 22:32the fiscal year is at almost 8 billion
- 22:35for the entire fiscal year 27 and that
- 22:38was below 8 billion. We were at 7
- 22:41billion and change. So, there has
- 22:42definitely been an improvement in
- 22:43guidance here. Of course, that's why
- 22:49the company is investing so much in
- 22:51inventory, because it expects to sell
- 22:53it, but this has to play out. It
- 22:56remains to be seen. You pay a lot up
- 22:58front and then you have to sell it. So,
- 23:03it's a company that, well, I think it
- 23:05can do well in the long term,
- 23:07undoubtedly, as it was very cheap at $
- 23:09100, but now at $ 160 it's not as
- 23:13appealing. Even so, I must say that I
- 23:15don't think it's a bad price either, I
- 23:17don't think it's a bubble, we are at a
- 23:19forward P/E of 20 and, well, for this
- 23:21year they are improving guidance and
- 23:23there will be more growth than what's
- 23:24written here. So, I think we are going
- 23:27to be around fair value. I mean, I
- 23:29don't see NetApp at a bad price, but
- 23:31the same thing happens to me as with
- 23:32Broadcom. Hm. I don't see a margin of
- 23:39safety and in the current environment,
- 23:42to pay a forward P/E of 23 or 20, I
- 23:45would prefer to think about it a bit
- 23:47more. I would prefer other options at
- 23:51lower multiples, as there are many
- 23:53things at lower multiples and we have
- 23:55been able to buy at much lower
- 23:57multiples. Of course, we've come from
- 24:01buying software at 10, 12, 7, 8, 15
- 24:03times earnings. So, well, let's go now
- 24:08to another company also related to
- 24:11artificial intelligence that had
- 24:14earnings today, which is Snowflake.
- 24:19Well, before moving on to Snowflake,
- 24:21let me tell you that NetApp has also
- 24:23slightly reduced its share buybacks,
- 24:25which I don't like either. I also
- 24:28wanted to tell you, in terms of cash
- 24:30flow, it is also important to mention
- 24:31that the repurchase of common stock has
- 24:33decreased compared to last year. Well,
- 24:37we see that in software there is
- 24:39constant repurchasing, while in AI
- 24:41infrastructure they are scaling back
- 24:44buybacks. What does that mean? That the
- 24:46businesses don't look like they're at
- 24:47such good prices either. If they take
- 24:49away buybacks, Broadcom has stopped
- 24:51buybacks, NetApp is reducing buybacks.
- 24:54They must see that they are near the
- 24:56top of the cycle. It's possible, except
- 25:00for Nvidia, which is increasing them.
- 25:03Nvidia is increasing buybacks. Look at
- 25:08Nvidia's case, we have buybacks rising
- 25:09here in the cash flow. Nvidia does buy
- 25:13back. Nvidia looks like it's at good
- 25:15prices, it buys back more and more, but
- 25:17NetApp does not. And Broadcom doesn't
- 25:19either; buybacks are going down. For
- 25:22NetApp we have buybacks falling, and
- 25:24for Broadcom we have buybacks falling.
- 25:30You see it here, right? Buybacks going
- 25:31down, not up in the last quarter. So,
- 25:34well, it's another reason to be
- 25:36cautious. Right, let's look at
- 25:38Snowflake. It is a database software
- 25:42company. Basically, companies use it to
- 25:46store data at the software level and
- 25:48draw conclusions about their business.
- 25:51It is an observability-style company
- 25:54and although it is more of a modern,
- 25:59advanced database company. Well,
- 26:03Snowflake has presented extraordinary
- 26:06results and artificial intelligence is
- 26:08a great advantage for software. AI
- 26:12boosts software. If it boosts anyone,
- 26:16it's the data layer software. Data is
- 26:21essential to use AI well, and it is
- 26:23going to be the new trend. Once the
- 26:26infrastructure investment has been made
- 26:28, we must invest in software. I think
- 26:31the era of hardware is ending and the
- 26:34era of software is arriving. Hardware
- 26:38has gone up a lot, everything is
- 26:40generally at expensive prices. There
- 26:43are also things at reasonable prices,
- 26:46but hardware has gone up a lot,
- 26:48everything is near highs. Not software,
- 26:52software has had a hard time, it has
- 26:54suffered, but its time is coming. The
- 26:58hardware era ends and the software era
- 27:02arrives. It will change later because
- 27:05these are cycles; hardware and software
- 27:07alternate, but I think the software
- 27:09stage is here and I am very well
- 27:10positioned, right? In Snowflake,
- 27:14because I would have never bought
- 27:16Snowflake as it is a company that has
- 27:18no profits because SBCs are still very
- 27:20large, but it is growing a lot. In fact
- 27:23, Snowflake has already recovered quite
- 27:25a bit. Notice that Snowflake software
- 27:28was really coming from massive
- 27:30collapses of 60%. We had talked about
- 27:38software, right? In the months of April
- 27:41, May, June, brutal months to buy
- 27:44software. Snowflake too, if you liked
- 27:46the company. As I say, it doesn't
- 27:50convince me because it has no profits,
- 27:52but we were able to buy many other
- 27:54software companies in that period,
- 27:56around the middle of this year. There
- 28:01has been a strong recovery and many
- 28:03software companies can no longer be
- 28:05bought, but some still can, and we have
- 28:07already done our homework on Snowflake,
- 28:10with 23%gains after-hours following the
- 28:12good results. And this one has already
- 28:15slipped away. You can't buy Snowflake
- 28:17anymore. Snowflake cannot be bought. In
- 28:20the Data Layer, perhaps the only one
- 28:23left is Dynatrace. We'll talk about
- 28:25that one now. Salesforce also has a
- 28:28data layer with Informatica. So, I
- 28:30think I’d lean toward Salesforce,
- 28:33which has a database segment. But hey,
- 28:37notice that in Snowflake, we have
- 28:39brutal growth in both sales and
- 28:40earnings. Sales have gone from 1,144
- 28:44million to 1,546. Here we have growth
- 28:48of over 30%, around 35%. Very good
- 28:52revenue growth, clearly beating
- 28:54expectations, and while profit is
- 28:57negative, if we subtract SBCs and look
- 29:00at net profit adjusted for normal
- 29:02business operations, we see it went
- 29:05from 128 million to 235. That is an 80%
- 29:09growth. In other words, Snowflake is
- 29:14really growing at full speed, and if we
- 29:16look at the forecasts, they have also
- 29:21improved. We have guidance that
- 29:23improves for the entire year of '27. We
- 29:27have 6 billion in the product branch,
- 29:30to which you then have to add the
- 29:32service branch, which is small, but
- 29:35this is just the product, and it’s a
- 29:3736%year-over-year growth for fiscal
- 29:40year '27, when 31%was previously
- 29:42expected. So, it's rising, it’s
- 29:47accelerating, and we are seeing this in
- 29:49software companies that are picking up
- 29:51speed. Software is starting to
- 29:54accelerate at Snowflake. Sales are
- 29:56beginning to accelerate. Profit, we
- 29:59don't have profit, there isn't any. It
- 30:01has losses because the SBCs are
- 30:03enormous. We have SBCs that are even
- 30:08higher than the free cash flow, and
- 30:10although there are some buybacks, they
- 30:13don't fully correct them, and above all
- 30:16, they might correct the dilution, but
- 30:19not the costs. The cost is the cost of
- 30:24the SBCs. They might correct it later
- 30:26with a buyback, but you still take the
- 30:28hit on the cost. So, it is a company
- 30:34that is young, it already has a
- 30:35considerable size, but it is a young
- 30:37company, and notice that there is
- 30:38dilution, there really is dilution.
- 30:40They were only able to buy back in 2024
- 30:42, but look, right now it is a company
- 30:46that, if you own it, congratulations, I
- 30:48think it's a good business, but I
- 30:50wouldn't buy it at these prices. We are
- 30:54at a forward P/E of over 100, it
- 30:56doesn't make money, but well, I mainly
- 30:59wanted to bring it to you because
- 31:01software is largely—if we look at
- 31:03data companies—doing very well, and
- 31:06one of them, for example, is ServiceNow
- 31:08. ServiceNow was at very good prices,
- 31:12we were able to buy it. That is one I
- 31:14do own. It is up 3%after-market. It's a
- 31:19bit pricey now, but it could have been
- 31:20bought at very good prices in the $ 90
- 31:21range. We also have the case of
- 31:26ServiceNow, which is also ITSM, for
- 31:28technical support, but it has a data
- 31:30layer as well. And the same with
- 31:34Dynatrace; it also has its
- 31:35observability layer and is also
- 31:36recovering. At a forward P/E of 24,
- 31:39we're not doing so bad. For ServiceNow,
- 31:42the forward P/E is already 30, or
- 31:45otherwise, I would also mention the
- 31:47case of Salesforce, which is primarily
- 31:50CRM (customer relationship management),
- 31:53but also has a data management branch
- 31:56with Informatica, and in fact,
- 31:58after-hours we're seeing a small rise
- 32:01due to Snowflake's good numbers. So
- 32:05that's another one, I really like this
- 32:06one. At a forward P/E of 17, there has
- 32:09been a recovery because Salesforce
- 32:11comes from a P/E of, well, it comes
- 32:13from a P/E of 10 at 150. We are at 250.
- 32:20Here with Salesforce, we were able to
- 32:22get in well in that 150 range and I'm
- 32:24very happy. I really like Salesforce
- 32:28and I have it in my portfolio, just as
- 32:30I also have ServiceNow. I would have
- 32:34never bought Snowflake because it
- 32:35doesn't meet the requirements to enter
- 32:37my portfolio. And we are going to
- 32:40finish with alcoholic beverages to
- 32:41change the subject a bit. We have
- 32:45talked a lot about AI and the companies
- 32:47that were presenting on it today, which
- 32:48are some of the most relevant ones
- 32:50reporting today, but we also have today
- 32:51...results from Brown-Forman, one of
- 32:56the world leaders in alcoholic
- 32:57beverages, the owner of Jack Daniels.
- 32:59Above all, together with Pernod Ricard
- 33:03and Diageo, it is the oligopoly of
- 33:06spirits. I wouldn't include beer. There
- 33:09is also Constellation Brands, but
- 33:10that's more beer. Spirits are
- 33:13Brown-Forman, Pernod Ricard, and Diageo
- 33:17. Well, bad results. Alcoholic
- 33:21beverages are still doing poorly, with
- 33:23sales falling 1%. And operating profit,
- 33:28which would be the most adjusted to
- 33:30business operations—because net
- 33:33income includes some expenses from
- 33:35layoffs and employment readjustments
- 33:38that happened last year—we have a 3%
- 33:41drop. In other words, we remain in
- 33:44decline; it's not a severe decline, but
- 33:46alcoholic beverages are in decline.
- 33:48This is bad. This is something we don't
- 33:50like. I don't hold Brown-Forman shares,
- 33:53but I do hold Pernod Ricard, for
- 33:55example, and this is not good. We
- 33:59remain in decline and, furthermore,
- 34:01it's expected for the full year '26,
- 34:03which is their fiscal year '27. Uh,
- 34:09decline as well; we really have a
- 34:11forecast that it will take time for
- 34:14growth to return to alcoholic beverages
- 34:17. I want you to see how Brown-Forman
- 34:22has had a huge drop from its highs, a
- 34:2462%drop from its highs. It is a
- 34:27historic collapse, but it is also true
- 34:30that this historic collapse is deserved
- 34:33because Brown-Forman had never had as
- 34:36many problems as it is currently having
- 34:39. It is a very sharp drop, the
- 34:41historical trend had been very nice and
- 34:43Jack Daniels has been for decades,
- 34:45right? The best-selling American
- 34:47whiskey, the best-selling American
- 34:48whiskey in the world. It also has other
- 34:54brands, of course, but Jack Daniels is
- 34:57the most relevant and it has been a
- 34:59stable, growth-oriented, nice company,
- 35:02but in recent years it has been
- 35:04plummeting. It could be an opportunity
- 35:07if you trust that the industry will
- 35:09recover. Yes, but of course, it all
- 35:11remains to be seen. Notice that profits
- 35:14have been stagnant since 2016. Yes,
- 35:18there was a peak in 2022 after COVID
- 35:20because people took to drinking alcohol
- 35:22, but we have fallen back down, right?
- 35:25So we are there in a profit stagnation.
- 35:30Sales are rising a little, marketing
- 35:32stuff and so on, but they are logically
- 35:34contracting margins. It seems sales
- 35:38were growing a little bit again in the
- 35:40last quarter. Now we are stagnant again
- 35:42. I'm not saying it's a bad company
- 35:45because it's not a bad company. But I
- 35:48don't know if this is going to grow
- 35:49again. And we are at a forward P/E of
- 35:5216, which perhaps, well, if the company
- 35:54grows a little bit again, it's not a
- 35:56bad price at all. It is the lowest
- 36:00multiple, one of the lowest multiples
- 36:01in its history since the year 2008,
- 36:03year 2000. But of course, we need to
- 36:06have more alcoholic beverage sales and
- 36:08I don't know if the trends of young
- 36:10people and new generations are towards
- 36:12higher alcohol consumption. It is true
- 36:14that Brown-Forman works internationally
- 36:16and in emerging markets there is still
- 36:18a lot of alcohol consumption, perhaps
- 36:20it is more the West, Europe, the United
- 36:22States, right?,that consumes less, but
- 36:24in emerging countries, Latin America,
- 36:26alcohol consumption is growing, of
- 36:28course, in Asia too. So, well, I think
- 36:34it is a company that if you trust in
- 36:39the recovery of the industry, it can be
- 36:41held, but I don't see it as especially
- 36:43cheap either, right? In any case, a
- 36:46reasonable price. 27, I don't think it
- 36:50is cheap. Within this industry, the one
- 36:53I like the most, if I had to choose one
- 36:55, is Heineken. Heineken is not pure
- 37:00alcoholic beverages as such, because
- 37:02Heineken also has non-alcoholic ones.
- 37:07If you're dealing with beer after all,
- 37:10it's a beverage that differs from
- 37:12spirits and has different consumption
- 37:15habits than a high-alcohol drink. Beer
- 37:21has different consumption
- 37:22characteristics, and it has also had a
- 37:24hard time. Heineken, of course, has
- 37:27also had a hard time, but not that hard
- 37:29. In fact, it set a record in sales and
- 37:32profits recently, from 2024 to 2025.
- 37:35And well, we've had a small decline
- 37:38that is slight and with growth
- 37:41estimates. Furthermore, Heineken is a
- 37:44historic, family-owned company, it's a
- 37:47company that is cutting costs, that is,
- 37:49well, doing things as they have always
- 37:52said. Breweries, well, they have their
- 37:56problems, uh, consumption cycles,
- 37:58margins that aren't very high, capital
- 38:00costs, and Heineken has always been the
- 38:02best managed in the entire industry,
- 38:04without a doubt. Better than others
- 38:10like Anheuser, better than others like
- 38:12Constellation Brands and so on, right?
- 38:14I mean, Heineken is also better than
- 38:17Carlsberg; well, Heineken has always
- 38:19been the best, without a doubt, the
- 38:21best managed. And the expectations are
- 38:23good, and above all, the P/E ratio is
- 38:2513. When you have a P/E of 13, it's not
- 38:28the same as 16 like in Brown-Forman. In
- 38:31Heineken, it's a P/E of 13, and a P/E
- 38:33of 13 discounts practically no growth.
- 38:36It discounts a little bit of growth as
- 38:37long as this goes at the rate of
- 38:38inflation. We are going to be fine. We
- 38:41are going to be fine. And well, we have
- 38:46a dividend that isn't bad and a few
- 38:48buybacks, so everything adds up too. I
- 38:52would say that Heineken is the most
- 38:53interesting of all, and that you can
- 38:57hold it, and it would be where I'd be
- 38:59most comfortable because then Pernod
- 39:01Ricard or Diageo in spirits, the same
- 39:03thing happens to me as with
- 39:04Brown-Forman. The accounting is much
- 39:07more deteriorated. In fact, at Pernod
- 39:10Ricard, look at that massive drop in
- 39:11sales, drop in profits, uh, quite a bit
- 39:13more severe. It's true that a P/E of 11
- 39:18is cheap. It is cheap. I like Diageo a
- 39:22little less. I've always liked Diageo a
- 39:24little less, but, well, it's also cheap
- 39:26. Pernod Ricard is at a P/E of 11,
- 39:28which is among the lowest in its
- 39:30history since 2008. So, I think that
- 39:33sooner or later there will be a
- 39:34recovery. I don't think it's the time
- 39:37to sell and get scared about alcoholic
- 39:39beverages at all. I think that with
- 39:42patience these companies will recover,
- 39:45but they don't seem like bargains to me
- 39:47, they don't seem like they are super
- 39:50cheap and businesses with good
- 39:52accounting, no. In fact, I am waiting
- 39:55for them to recover to sell Pernod
- 39:58Ricard and Yes, it will surely go back
- 40:02to 100, 100 and something, and that is
- 40:04where I will try to exit. I am much
- 40:07more relaxed with Heineken; you can
- 40:09leave it for the long term without any
- 40:11problem. I think it is a company that
- 40:14is worth letting work. There are good
- 40:15forecasts. Already for this year. We
- 40:18have earnings growth forecasts due to
- 40:21cost adjustments. 8%in earnings. Then a
- 40:257, a 6, a 7 at a forward P/E of 13. We
- 40:28are doing well, we are doing well. So
- 40:31anyway, how do you see today's results?
- 40:33What do you think about Brown-Forman?
- 40:35What do you think about Snowflake? What
- 40:37do you think about NetApp? What do you
- 40:39think about Broadcom? Which of these
- 40:40companies do you hold? Are you buying
- 40:43anything? Well, as always, I'll read
- 40:46your comments. Remember that none of
- 40:49this is investment advice, and I'll see
- 40:51you in the next video. A big hug.
About this transcript
This page contains the full transcript of ¿Empiezan Los Problemas En La IA? by Alex Dito - Invertir en Valor, generated from the public captions YouTube serves with the video. The transcript has 6,103 words across 921 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.
What you can do with it
Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.
Free YouTube transcript tool
YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.