YT YouTube2Text

YouTube2Text → Transcripts → AI over-reaction and earnings for LITE, FN, AMD, QCOM, and COHR

AI over-reaction and earnings for LITE, FN, AMD, QCOM, and COHR — Transcript

by Mark Meldrum · 18,845 words · 2,768 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:12We have a lot to get through today. We
0:17had midterms last week and I am more
0:20than happy to admit that my uh suspicion
0:26of how Trump would behave after the fact
0:28was wrong. Although having gone back and
0:30listened uh to the previous market
0:33outlooks that I gave, I did give one
0:35scenario which would make it difficult
0:37for Trump to um use the court system is
0:41that you would need a Democratic victory
0:44so large that it can't be questioned. He
0:47would then just call the elections
0:48rigged. But he didn't.
0:51Hats off to him on this one. He didn't
0:53call it rigged. So I am more than happy
0:57for that. However, the outcome, we had
1:01one ugly Thursday, so we may as well we
1:04may as well have had the result that I
1:07suspected because I think we could have
1:09got the same red Thursday. But fingers
1:12crossed, November 8th, we have to get to
1:16all the way to to January 3rd, the
1:19swearing in of the new Congress and the
1:20new Senate. We got to get all the way
1:22there
1:24uh uh without anything happening. He's
1:26got two months basically to change his
1:28mind about this. So fingers crossed that
1:30he just says I can't I can't win this
1:33one. We already see that he's throwing
1:35other people under the bus. He's saying
1:37we didn't win. I wasn't on the ballot. I
1:39didn't lose. These losers lost. And
1:42looks like he's willing to give up some
1:44of his people as well. Hegs being first.
1:47We'll talk a little more about that
1:49later. Let's try to make sense of last
1:51week as we go through um the the screams
1:54that we're going through because it
1:56seemed rather excessive to me. It seemed
1:59the concerns seemed justified.
2:02The big one regulatory
2:05Democrats won in key areas where there
2:09is data center uh um plans and data
2:13center construction. What's going to
2:14happen there? Is there going to be a
2:16moratorum on new data centers in blue
2:18blue states? Are they going to have to
2:20move to red states and move to red
2:22states or other countries? They will.
2:24Uh, is there going to be a moratorium on
2:26AI at the congressional level of which
2:29Trump would just veto anyways? Valid
2:32concerns. Is it setting something up for
2:342028?
2:36The second, which has been lingering
2:39since the Neoclouds decided to show
2:41their true colors and people decided to
2:43open their eyes finally and say, "Hey,
2:46that doesn't work." is the cost and
2:49availability of capital to fund the AI
2:52buildout and the necessity of even using
2:55the Neoclouds which Microsoft late in
2:58the week did question
3:02and we'll uh we'll go through some of
3:03the facts on that one but Microsoft did
3:06cancel uh some deals only because they
3:09were delayed and it was really only with
3:11one company and they are looking at
3:12other companies saying you no longer
3:16have a a a a preferential access to
3:20power over what we have is the end
3:23result. So, we don't need you, which uh
3:27I think everybody could have seen that
3:28coming. The Neocloud's existence was
3:31predicated on the fact that you have
3:33access to power now. That's something I
3:36don't have. Fine, we'll use you. But
3:39once that access to power became a level
3:41playing field, I don't think the
3:42neoclouds have any reason for survival.
3:46These are the two big things, the
3:48regulatory and the access to capital.
3:51Access to capital can be solved if we
3:54start seeing significant ROIs on on
3:57these investments, but so far it's been
3:59elusive or it's been here and there. The
4:02regulatory I think will be solved over
4:05time. They'll just extract something
4:08from uh the hyperscalers who want to
4:10build in certain communities. But it
4:12would be foolish
4:14for a governor of a state to say nothing
4:16in this state. No, we do not want that
4:18tax base in our state. We do not want
4:20those jobs in our state. We do not want
4:23to even find a solution to the problem.
4:25In fact, we might even go so far as to
4:27ban AI within the state. Who knows? Uh I
4:32think that's a step too far. But I think
4:33at some point somebody along the way and
4:36a group of adviserss will say, "Look,
4:37these are construction jobs.
4:39uh these are uh local data center jobs
4:42as well. They pay they pay fairly well
4:44once they're done. We don't know how
4:45much expansion there is and think of the
4:47tax base associated with it. We have a
4:49budget hole to fill. So I think they'll
4:51get somewhere on that. So against this
4:54backdrop, we have a good economy in the
4:56US
4:58and you have solid earnings. So we'll go
5:01through uh a number of earnings reports
5:03and again when it comes to the data and
5:07AI side strong and in fact uh uh
5:11increasing
5:12and the economic data in the US is
5:15pretty good. Let's start with ISM
5:18manufacturing which we got last week.
5:20PMI 55.5
5:22expanding from 54.5
5:25expectation was 54.6. So that is a nice
5:28uh uplift. Employment uh in expansion
5:31territory uh call that unchanged. This
5:33is a survey. So uh I wouldn't say that
5:36that statistically is any different.
5:38Construction spending was a negative
5:401.9. However, this is not part of ISM.
5:44It's just data that showed up in here.
5:45And this was for September. A little bit
5:47backward looking. New orders uh 55.5
5:51still in expansion territory. Uh 553.
5:54Call that unchanged. not growing but not
5:57shrinking and prices paid October.
6:00Shocker shocker 779 expected 77 came in
6:03much lower 752.
6:06So we're looking at this we're saying
6:08okay well there's nothing there's
6:09nothing here that suggests that anything
6:12is falling apart. It's holding up nicely
6:14in expansion territory. Okay, job done.
6:18Uh let's jump down to ISM services 545
6:22uh previous 54 uh 549
6:26and uh the expectation was 548. Call
6:29that basically unchanged. You're in the
6:31neighborhood. Business activity that is
6:33a step down 538 from 565.
6:37Employment
6:38uh more into expansion area 514 versus
6:4250.1. New orders down quite a bit here.
6:45594 to 559
6:47and prices a little bit uh down. All the
6:52key variables are in expansion
6:54territory.
6:56Uh when we have a 50 uh a 559 from 599
7:00that means you're still expanding more
7:02more uh respondents are uh answering the
7:06same or growing month over month. So
7:08this was a growth over the previous
7:10month. This is growth over this month
7:12just not at the same uh not at the same
7:14strength but still growth over those
7:16months. So job done on ISM services as
7:19well. When you start to see these with a
7:21four handle under 50 on both of them and
7:24decreasing that's something to be
7:26worried about. We don't have this. Um
7:29we'll see that we had jobs numbers uh
7:31this week as well that were very good
7:34and the revisions were very good on top
7:36of GDP. we already got over 3%.
7:40So if we're looking for the economy as
7:43being the source of the angst last week,
7:47it's the wrong place to look. Balance of
7:50trade, uh this is uh quite backwards
7:52looking for September. Canada and the
7:54US. Canada uh increased its uh lead here
7:585.35 billion up from 4.95 billion. Uh
8:03exports higher, imports lower.
8:06That's a positive for September GDP in
8:10the US. Let's call that unchanged 105.7
8:12to 105. The expectation was for 108. So
8:15it came in lower than expectations, but
8:18still you're uh you know 100 billion a
8:21month, 1.2 trillion a year on the trade
8:24balance. Exports were up uh but uh so
8:27were imports.
8:29Bank of England raised rates to 4%.
8:33We'll see on the next screen. Mexico cut
8:35rates
8:37Friday, jobs day. Um, big numbers out of
8:42the US, not so great out of Canada.
8:45Unemployment rate in October for Canada
8:47up to 6.7%.
8:504.7,000
8:52jobs were lost. you multiply that by
8:56about nine to get what it would be
8:58equivalent on a US basis if the US had
9:01printed somewhere around 43 44,000 uh
9:04job loss uh that is uh just to put the
9:08size difference into effect uh lost on
9:11full-time gained on part-time
9:13uh which is sort of the wrong way to go
9:15but uh you have to account for job
9:17losses right would have been better to
9:19have full-time gains with a lot of
9:20part-time losses but uh because
9:23full-time uh typically pays more. You
9:25see this over here in the hourly average
9:28hourly wages uh 2% year-over-year down
9:30from 2.3%. That is uh well you know the
9:35the two the three two of the three
9:38measures that the Bank of Canada looks
9:40at uh came in at 1.9 and 1.8 and the
9:44other one I think was 2.2.
9:46So we have wages 2% that appear to be
9:49growing at the rate of uh if you if you
9:53blend those three you probably get to
9:55about 2%. Growing at the rate of
9:57inflation uh not exceeding the rate of
9:59inflation. So you have flat real wages
10:02year-over-year participation rate
10:04unchanged.
10:06US very different another A+ paper
10:10165,000 job gain uh revisions to August
10:1453,000 higher revisions to September
10:1731,000 higher so 84,000
10:22jobs were found in the revisions on top
10:24of the 165
10:27unemployment rate dropping to 4.1
10:30surprising because you have
10:31participation rate increasing to 619
10:35So you have more people in the labor
10:37force yet yet you have a drop uh in the
10:41unemployment rate. That's making that's
10:45uh accounting for the 84,000 plus the
10:48165.
10:50Average hourly earnings last month. 264
10:54uh this month 369. That is a big
10:57increase in one month.
11:00um annualized it's probably at the high
11:03end of the range of where the uh Fed
11:06feels that um labor is not contributing
11:11uh to inflationary pressure. Um so still
11:14in there average hourly earnings
11:16year-over-year 3.1 3.2 this is kind of
11:19where inflation is if not a little bit
11:22higher at the headline because consumers
11:25face headline inflation. So, while
11:27Canada appears to have flat real wages
11:30year-over-year, it does appear that the
11:32US is experiencing
11:34um negative real wages year-over-year at
11:373.1%.
11:39Unchanged on the average weekly hours,
11:42taking out government, your private
11:43payrolls were 159. Government didn't
11:46make up a lot of it. Usually, health
11:48care makes up quite a bit. Healthcare
11:50was 55, only made up 33% of the job
11:53gains. healthcare had in some months
11:56were 100% of all the job gains. 80 70 80
11:5990 100% of the job gains only a third of
12:02the job gains. Your diffusion indexes
12:05for your total industries 576. Dusion
12:08tells you um which industries are
12:11expanding or contracting in terms of
12:12employment total 576 up from last month.
12:15Manufacturing alone 625
12:18up from last month. again looking
12:21somewhere to uh you know deduct marks
12:23off of this uh off of this paper and
12:26it's hard to find where you're going to
12:27deduct marks. Household 463
12:31breaking down in terms of uh lesson high
12:35school um losers here. High school 479
12:40lesson BA uh 251 BA 422. What's the
12:43message here? Don't be a quitter. Don't
12:45quit high school. Don't quit college if
12:47you quit. It says something on your
12:48resume. No, you quit. You're out. 16 to
12:5219. Uh, you're too young. Uh, lost
12:5515,000, but gains everywhere else. 20 to
12:5724, 25 to 54, and surprise surprise, the
13:0155 plus gaining 245,000
13:06full-time. This is what you want to see.
13:08Big full-time job gains and uh part-time
13:11losses. So, some of the part-time losses
13:13might have been converted into full-time
13:14gains. Overall, an A+ paper on this one.
13:18Hard to find uh hard to find fault. I'm
13:20sure if my job was to find fault and I
13:23dug into the categories I oh this
13:25category lost and this category lost. I
13:27think information technology lost a
13:29thousand jobs.
13:31Uh but other categories gained. So it
13:34was hard to see uh if there was a
13:36pattern in there somewhere. And this I
13:38think is uh most most likely AI caused.
13:42We did get challenger gray and Christmas
13:44earlier this week as well. I didn't put
13:46it up here, but uh layoffs were h well
13:49in line of what uh what was normal. I
13:51think planned hirings exceeded planned
13:53layoffs, so that was good.
13:57So, as far as GDP goes, um over 3%. As
14:01far as the job market goes, you look at
14:03the jobs that were gained, the
14:04revisions, and you look at how low
14:06initial claims have been over the last
14:08couple months. Solid there. ISM
14:11manufacturing, ISM services,
14:13productivity, unit labor costs, all
14:15going in the right direction.
14:18So, I don't know that we can say it's
14:20the economy uh that led to uh the
14:23selloff that we saw on Thursday. I think
14:26it really comes down to the regulatory
14:29uh and the uh issue with capital.
14:33uh dovetailing off the jobs report. Uh
14:36this is a paper I came across really
14:38interesting from the NBER working paper
14:41series AI premium.
14:44Um it uses data provided by open router.
14:47Let's look at the abstract. Uh it was
14:49the abstract that got my attention. Uh I
14:51read I thought okay I got to read this
14:53paper now. It's um it's long. It's about
14:5660 70 pages but it ends at around page
14:5930. So, it's really not longer than any
15:01other social science paper. Uh, but it
15:04has a lot of stuff in the appendices.
15:07Let's see what they did here. Using 380
15:10trillion tokens of realized AI
15:13consumption across more than 400 large
15:16language models.
15:18I thought, 400? Are there really that
15:20many? uh from the licensed proprietary
15:24open router data set covering
15:26approximately 2% of current global
15:28monthly AI token consumption
15:31380 trillion tokens is 2%.
15:34We analyze how AI affects firms,
15:38markets, and workers. Leveraging the
15:41unprecedented size, scope, and
15:43granularity of this data, we construct
15:46the AI factor from growth in tokens,
15:49dollars, and users, estimate firm level
15:53AI betas from stock return co- movement,
15:56and characterize the AI premium. For
15:59anyone who has been through CFA level
16:02one and level two quant uh regression,
16:05if you uh understand multiple
16:07regression, this is nothing more than a
16:08multiple regression.
16:11First, we build a highfrequency AI
16:13factor and decompose it into salient
16:16components.
16:18Second, we show that firms whose returns
16:21co-vary more positively with the AI
16:24factor, those being high AI beta firms,
16:27earn higher subsequent returns and the
16:31AI premium is large and heterogeneous.
16:34A value weighted long short strategy
16:38earns 64.1 basis points, wait for it,
16:42per week. 64.1 basis points per week.
16:48and you multiply that by 52, you're well
16:50over 30% a year. The premium is large
16:54for loadings on the intensive frontier
16:58oriented margin of AI consumption. This
17:01is important here. What do they mean by
17:03frontier oriented margin of AI
17:05consumption? Closed source models. Uh
17:09this is your uh open AI uh claude gemini
17:14open-source models paying and seasoned
17:18users. So not the free ones, the ones
17:20who are paying and the seasoned users
17:24uh and long prompts,
17:27closed source models, paying and
17:29seasoned users and long prompts.
17:34premium is large for loadings on these
17:36but not on casual or openw weight use.
17:42That would mean a lot for this
17:44discussion of but the Chinese models are
17:47really cheap. All these openweight
17:48models are really cheap to use. If if
17:51their data if you go through their data
17:53and you say okay I I found I find that
17:56that that your data is sound your uh uh
17:59uh your variables make a lot of sense to
18:02me. If we find that there is um
18:06insignificant
18:08premium attached to usage of those
18:10models but more premium attached uh to
18:13the frontier models that means there is
18:16something qualitatively different that
18:19can be measured.
18:22Third, the premium reaches beyond
18:24technology firms into consumerf facing
18:27and capital heavy parts of the economy
18:30but is absent in emerging markets
18:33including China. Interesting.
18:36Fourth, the AI exposure is more positive
18:38in non-rine interactive work. I think I
18:42think anybody could have come up with
18:43that conclusion. And more negative in
18:46analytical, scientific and operations
18:48control skills. An occupation with one
18:51standard deviation higher in interaction
18:54and communication content has a.36
18:58standard deviation higher market implied
19:00AI exposure.
19:02Additionally, we provide early evidence
19:04on the rise of the agentic economy.
19:07So if you want to search for it, it is
19:09uh it's available on the internet
19:11because it's a working paper uh NBER and
19:14the title is AI premium. Here are your
19:17authors. Here's the affiliation uh where
19:20is the here's the affiliation of the
19:22authors uh of what uh uh where they're
19:25at looking at rates and yields. A good
19:28week uh for yields. Capital market curve
19:31showing a bull flattener. Rates coming
19:34down. That is bullish for bond pricing.
19:37And you have a flatter curve week overw
19:40week both in the capital market curve
19:42the two to the 10 by three basis points
19:45in the money market that the capital
19:47market curve 11 basis points.
19:50Um look down here. This is uh rather
19:53interesting. I've got a uh the level of
19:55our public debt our well the US is
19:58public debt. I feel I'm part of the US
20:00market because I'm in it so much. So I
20:02say our we and I don't really mean it in
20:04that sense. Public debt 32.7 trillion
20:07intergovernment debt 7.7 40.556
20:10trillion. We had hit 40 trillion in
20:13August and uh that was less than 90 days
20:16ago and we've already added another 556
20:19billion of that. The US debt will hit 41
20:22trillion late at this rate late January,
20:26early February, and hit 42 trillion
20:29sometime next summer, July, uh late
20:31July, early August. Well, more like mid
20:34July, I think, early to mid July will
20:36hit uh 42 trillion. the new uh structure
20:41of the government uh with the two
20:44houses, Congress and Senate, swinging
20:46wildly towards uh Democrat. I don't
20:49think that changes spending at all. Uh
20:52in 2026, the average interest rate on
20:56the US debt 3.531%
20:59uh 1.36 trillion in interest expense.
21:04October which is the first month of the
21:06next fiscal year for the government.
21:09Their year ends September 30th. So we
21:11have one month already in interest
21:14expense is up to 3.58%
21:19and uh the October interest expense
21:22clocked in at 115 billion at the 3.58%.
21:27If we look at where things could be
21:29issued, this is the current uh level of
21:33rates and yields, there is no way to
21:36issue any debt at all that is cheaper
21:38than 3.58%. So any rolling over of the
21:42debt, which is what's going on now,
21:44simply just keeps pushing this higher.
21:46and 3.58
21:48uh uh times 40.566
21:52uh gives you an interest expense that is
21:54greater than what 2026 is. Uh if you're
21:58going to be hitting half a trillion
22:01every 90 days, um this is this is at
22:05this point an unsustainable pathway. I
22:09don't know that I would look for
22:10continual uh flat or continual bullish
22:14uh moves on the long end of the curve.
22:16given this trajectory unless you can
22:20maintain a real growth rate greater than
22:233.58%
22:25in other words real GDP has got to come
22:27in at 3.6%
22:29every quarter or higher uh you're simply
22:33going to have a debt to GDP uh ratio
22:36that continually increases
22:39um at 3.6% 6% you're growing at the same
22:42rate as the interest and you're throwing
22:44the interest onto the debt. If you're
22:45also running a deficit, this number has
22:47to be bigger and that's not happening.
22:50So at this point right here, right now,
22:53this is unsustainable.
22:55Let's see if the conversation begins to
22:58turn more towards some kind of fiscal
23:00prudence somewhere.
23:04If the um if the new Congress wants to
23:08remove powers of taxation or I shouldn't
23:12say taxation, tariffs from the
23:13president, tariffs right now is a source
23:16of revenue. You kill you kill tariffs.
23:19You're going to have to find some other
23:21source of revenue or you increase this
23:23deficit even faster. One of the things
23:27discussions that uh that I've heard that
23:29are out there is the removing of the
23:31debt ceiling.
23:33Uh if you remove that then you never
23:35have this this this uh confrontation
23:38every now and then about look uh we're
23:41spending a little bit too much. Once you
23:43remove the debt ceiling then you never
23:45have to revisit that conversation.
23:49Um we did not have a good week in tech.
23:52Look at that. Down 5% week overw week.
23:54Well into correction territory. IWM as
23:58well in correction territory. S&P still
24:00holding up. not in correction territory.
24:03Look at the winner this year of the ones
24:04I track at least. Uh Staples, big loser
24:07last year. Uh best performing uh on the
24:10board here of the ones I have up. Uh
24:13Staples beating uh Triple Q, beating
24:16small cap, beating S&P, uh beating the
24:19banking sector, Staples beating IYR,
24:21beating XLU.
24:23Uh TLT still down for the year, but uh
24:26with the yield that you would have
24:28gotten, you're you're basically dead
24:30money. Slightly down for the year, but
24:33basically dead money.
24:35Uh Friday's job report, Thursday's
24:38selloff really uh took some rate hikes
24:41out of the equation and then Friday's
24:43job report added them back in. But
24:45still, look where we are
24:48from from three possible rate hikes down
24:51to two possible rate hikes. And I say
24:53two and seven even though there's eight
24:55meetings cuz one meeting we're told one
24:58meeting will be purely on policy with no
25:02decision being made at all.
25:06So the takeaway here is this is still a
25:08problem. Government finances are still a
25:12problem and a growing problem. And if
25:14you do get a rate hike in December, then
25:17this is going to go up uh again and the
25:21interest expense is going to get uh is
25:23going to go higher obviously, which
25:25means the debt will accelerate faster
25:27because well, you're running a deficit.
25:31Something some conversation has to come
25:34in into the fold on that one.
25:38Here's the breakdown.
25:40Um and I'll tell you a problem I had.
25:43I caught myself using Google. I went to
25:46uh open up a browser and I typed in
25:49current composition of Congress in the
25:52Senate and I got the URL links because
25:54it was an unsafe question to ask because
25:57it leaned towards politics. So I was
25:59told it was unsafe unsafe
26:02asking for how many people are sitting
26:04in what's unsafe question
26:07guard rails a little too tight on that.
26:08So I started clicking on them mindlessly
26:11and I and I got some numbers from one
26:14then I clicked on another different
26:15numbers and I clicked on another
26:17different numbers again. I'm thinking
26:18okay well I got to find a source of
26:19truth here. So after about 10 minutes I
26:21thought what am I doing clicking on all
26:23these URLs what am I doing here? Went to
26:26perplexity and I got it right away.
26:29So Google, you used to be something. I
26:32think years from now, we will look back
26:36on Google and say, remember when
26:39remember when Google used to be
26:41something? Um Google already reported
26:44we're already seeing some cracks in
26:45search. I I really don't even understand
26:48why anybody even uses Google search
26:51anymore, especially when you can ask an
26:54innocuous question like what's the
26:56breakdown of Congress between Democrat
26:58and Republican? And it answers back that
27:00it's an unsafe question.
27:04Wow.
27:05Congress, you need 218 for a majority.
27:07Democrats 254, Republicans 162. I guess
27:11uncalled. There are 19 that are uncalled
27:14as of now, but it doesn't really matter.
27:16254 over 218. Dems clearly control
27:20Congress. The Senate, you need 51 uh for
27:24a majority. 50/50 still gives the
27:26majority to the Republicans because the
27:28vice president is the tiebreaker. Uh
27:31with 51, uh you'll have 5150. The vice
27:35president can't possibly there can never
27:36be a tie if the Dems all stay uh united.
27:40Democrat 54, Republican 44. Vice
27:44President is meaningless in the Senate
27:45at this point. Uncalled are two. I think
27:48it was Mississippi and Alaska that are
27:50still uncalled. Uh with Democrats
27:53leading by very very small margins, so
27:55you'll probably have recounts for a
27:57while. But if they win, that would be a
28:005644.
28:04uh still not enough in the Senate to get
28:06any kind of conviction on any kind of
28:08impeachment unless of course Republicans
28:10are willing to go along look for a Heg
28:13Seth probably to be the one unless he
28:16resigns but they'll still be still be
28:20questions asked.
28:23Um this is this is this Congress this
28:26was expected that the Dems would take
28:28Congress. this was iffy if they would
28:30take Senate. And I've heard a lot of uh
28:34decomposition of why this was the case.
28:38Um with the economy being pointed out as
28:41as the big thing is people are facing
28:44higher costs everywhere they go. Uh and
28:48the economy usually does it. I if you
28:50have a good economy, usually it favors
28:52the incumbent. If you have a bad
28:54economy, uh it favors the opposition. It
28:57certainly did. But but this is this is
29:00more than just the economy. And of the
29:03ones that I have listened to, um only a
29:06couple have really said listen, you
29:08know, they were just tired of of the
29:11lies, the the grifting, the
29:15corruption. Only a couple really said
29:17that to have this kind of sway. This is
29:20more than just the economy.
29:23uh the economy would have resulted, I
29:27think, in in this leaning, but this is
29:30not a leaning. This is a shellacking.
29:32This is a a a a throwing out of of the
29:37Republican uh um administration
29:40and their handling of affairs
29:42altogether.
29:43So I don't think that we can discount
29:45that some of the vote
29:49maybe some a significant part of the
29:52vote was not an embracement
29:56of democratic ideals as much as a
30:00rejection
30:01of Republicans.
30:03So hopefully, and I've heard this
30:06stated, and I really do believe this,
30:09uh, and Bill Maher uh, uh, on Friday
30:12said this, and I agreed wholeheartedly
30:14with him, hopefully the Democrats don't
30:17interpret this as they love us. They
30:20really, really love us and what we want
30:22to get done, and they rembrace that
30:24crazy woke because they'll lose
30:262028 if they do it.
30:30when they decompose this win, hopefully
30:32they'll say, "Look, it's not necessarily
30:36true that they embrace everything we
30:38believe in. They're just rejecting that
30:40other stuff. So, let's lean into what
30:42they're rejecting and let's fix that for
30:44them without going full crazy again.
30:48Hopefully, they can do that and and if
30:51they can, they set themselves up well
30:53for 2028.
30:55The danger is as as uh as Mara had
30:59pointed out is that they think this is
31:01all about us. They love us. They really
31:03love us. They want more of this stuff.
31:06Let's lean into more of what we truly
31:08believe in as opposed to what they were
31:11rejecting
31:14after the election.
31:16Um, man, I'd heard this from uh um quite
31:20a few of the u de decomposition podcasts
31:24that happened after looking at what
31:26happened and quite a few of them said
31:28the same thing I'm about to say. Uh
31:31Trump had said, uh yeah, I'll accept the
31:33outcomes and said it is what it is. It
31:35was like that was the biggest shocker to
31:38me is like you got to be kidding.
31:41You know, I I seem to think that the the
31:44win was so decisive and so big that even
31:48if you were going to try to
31:51deny the outcome, it would it would be
31:54it would be a lot. It would be difficult
31:56to overcome this. Um, but I am I was
32:01wrong on that call. I thought at 801 for
32:03sure he would start saying that this is
32:06rigged, that's rigged.
32:08That being said,
32:11uh as of yesterday, he has directed uh
32:14the Department of Justice to open
32:16investigations
32:18into any Democrat calling for
32:20investigations.
32:23So,
32:24we still have 60 days. We still have to
32:27get to January 3rd. So far so good.
32:29Knock on wood, everything is fine in
32:32terms of the strength of the
32:34institutions.
32:36I hope I hope it continues on. I am more
32:39than happy to be wrong on this one. I
32:42hope I continue to be wrong on this one.
32:45Real yields a little bit mixed. The
32:48five, the seven, and the 10 all a little
32:50higher, but the 20 and the 30 a little
32:52bit lower. Break evens really came down.
32:56Look at this. from the 10 outwards lower
32:59than uh from the beginning of the year
33:02in uh in you could say it's the built-in
33:04inflation expectation visa v the nominal
33:07yield versus the real yield.
33:11Gold, silver, platinum, palladium had a
33:13good week in a bad year. Still negative
33:16quite a bit year to date. Copper bad
33:19week in a good year and aluminum uh you
33:22know treading water. Oil bad week in a
33:25good year. natural gas, good week and a
33:27bad year. And uh food prices coming down
33:30this week. Look at hogs and uh and
33:33cattle. Hogs down 35 35% cheaper than at
33:37the beginning of the year. Even meat
33:40year to date down seven well meat live
33:43cattle uh down 7% year to date. Things
33:47starting to move into the right
33:48direction. If you can get um Trump to
33:52sail his boats out of the Gulf, um I
33:56think you can have probably a good end
33:59to the year in terms of of inflation.
34:02We have an FOMC coming up 31 days. We'll
34:05get summary of economic projections.
34:08This may be the last SCP we see. It may
34:11be, I don't know.
34:13um 76% on a hike up from 70% 76% and I
34:19think the jobs number uh had a lot to do
34:21with that mortgage rates just having a
34:24bad time of it an all right week but uh
34:28year to date even including the dividend
34:31on these you are you got a 7.18%
34:36uh mortgage rate over 7% that's got to
34:39come down quite a bit before we start to
34:41get a lot of activity
34:43in this uh in this sector. Reverse repo
34:46almost at zero. Repo's at zero. Reserve
34:49still sitting uh hugging the $3 trillion
34:51level. TGA936.
34:53So for a little stressed got up to 3.97
34:57on November 3rd.
35:00Um this was the uh election day and then
35:04it's come down since then but 3.92
35:06usually is pretty much on target with
35:10where uh effective federal funds rate is
35:13at 388 but 3.92
35:15and the S is still sitting at about 6.3
35:18trillion. Uh I have this up on the
35:20screen. This is a uh a screenshot uh
35:24from the mobile app uh for Interactive
35:26Brokers uh uh listing all your trades uh
35:29for the last 90 days. This is from my my
35:31big account for the last 90 days. So
35:34this is August 9th to November 6th.
35:36There is the realized P&L 1.3 1.3
35:40million. It says 1791 trades, but you
35:44have to take that into uh into context.
35:47That's not 1791
35:50names. Usually when I buy something, I
35:53buy, I buy, I buy, I buy, I buy, it
35:56starts going up and going up. I'll sell
35:58a little, I'll sell a little, I'll sell
36:00a little. So, one name uh might have as
36:04many as 15, 20 trades associated with
36:07it, but it is still just one trade. And
36:10especially when I'm hedging with
36:11options. So, it sounds like a lot, but
36:14it really isn't a lot. But this sets up
36:15a problem. And you should be doing this
36:17into the end of the year is having a
36:19look at what your um what your realized
36:22gains are for the year. The number is a
36:25little a little higher than this. Uh but
36:28going into the end of the year, you'll
36:30always find in your portfolio uh that
36:32you're carrying some some names uh that
36:36uh are underwater that maybe you've just
36:38had enough of. And usually it's around
36:40this time of year that you start uh
36:43because you still have time to sell the
36:45at the money calls for November. And if
36:47they don't get called away from you, you
36:49can then go to December, sell deep in
36:52the money calls because they'll get
36:53called away from you December 18th
36:55enough for tax season and you can start
36:58to work uh work down the amount of gains
37:00you had. This by the way uh is the
37:03result of around mid August onwards uh
37:07or earlier than that. I uh decided that
37:09I was my portfolio was going to be a
37:11beta uh alpha portfolio which is long
37:14beta and you just take a futures
37:16position. I took NQ and then outside the
37:19portfolio you short very highvalued
37:21stocks. Well, it just so happened that
37:24worked out nicely. I was expecting the
37:27market to continue climbing and I was
37:30going to add um every week add to my
37:33short position, but it did what it did.
37:35I'm basically almost out of short
37:37positions to have because everything now
37:40the one the names that were on my short
37:41list now look like a buy. So what am I
37:44doing here? Um Capri I'm starting with
37:46Capri here. I've given it enough time.
37:49Um it's cheap and it has a nice yield.
37:52And you look at this and you say well
37:53why would you get rid of it now? Uh I
37:56have about across all the accounts I
37:58have maybe 90,000 close to 90,000 shares
38:02of Capre. sold 6,500 shares on Friday
38:06and I booked the loss. Uh the loss
38:09already shows up in here. So you see I
38:11have a ways to go. Why' I do it? Their
38:14earnings were mediocre.
38:17Um certainly nothing I'd look at and
38:19say, "Well, that's great." But nothing
38:21I'd look at and say, "Oh, it's falling
38:22apart." It was like, given the
38:24environment you're in, it's not bad. Uh
38:27you appear to have NOI growth at about
38:29the rate of inflation.
38:32So, real earnings, your real earnings
38:34are flat year-over-year. Eh,
38:38um, they are at about a 43% discount to
38:41net asset value. And you had a Bank of
38:43Canada that raised rates. Um, I don't
38:46know if they're going to raise rates
38:48again. I seem to think the rate increase
38:51uh was useless, but it is not helpful in
38:53the REIT space for rates to be going up.
38:55So, that's a bit of a headwind. The
38:57yield is 4.92%.
39:00$155 a year dividend at on a $31.50
39:04stock. You're getting a $4.92% yield.
39:08However,
39:10um where it matters, there is an over
39:12supply of condos in uh in the Toronto
39:15area and same with the Vancouver area
39:17and a few other uh key markets as well.
39:19Just a big over supply of condos. Well,
39:22they're going to have to be rented out
39:25uh in in uh you know, while they're
39:27waiting for a buyer. Uh also, new
39:31immigration ball policy numbers had uh
39:34come out uh during the week when uh we
39:36got the uh updated budget uh from the
39:40Liberals, their new budget. Um numbers
39:43are not changing. There was a belief
39:46that the numbers were going to increase.
39:48They're not changing. And if you don't
39:50have these numbers changing, when they
39:52arrive at the country, they're more
39:55often renters than buyers. They don't
39:58just show up in a new country and buy a
40:00house. You rent for a period of time.
40:02They're more often renters, especially
40:04on the student side and the uh and the
40:06work visa side. They are renters. You're
40:09not increasing it. So there's no new
40:10source of demand. You have an economy
40:13that for three months now has shed jobs.
40:15I ask myself over over the next 30 to 60
40:19days, do I see a 20% rally uh in this
40:23name? No, I do not. So, I can get out of
40:26it now. I can begin to solve the tax hit
40:30uh with this by taking the loss and I
40:32can get into something that I think will
40:34rally by the end of the year and then in
40:37January I can revisit Capri and decide
40:40whether I want to get in. My assessment
40:42at this point is that a two-handle on
40:44cap rate seems to be in the works that
40:48that at some point it is going to be a
40:50$29 stock and then a $28 stock. It is
40:55the biggest of the Canadian REITs. Uh I
40:58don't think that it is uh a target right
41:01now for private equity. private equity
41:04has taken out in the last few years.
41:07Three of the Canadian REITs
41:11actually Inter Mento are still trading
41:13uh on the market. Uh they haven't they
41:16haven't actually closed and there's
41:18about five of them left.
41:21Uh well, four of them four of them
41:23really left. I think Killllum is
41:25probably the most obvious target. It
41:27seems to be the best run in terms of the
41:30numbers that it produces. But cap rate I
41:32don't think is going to be a target for
41:34an acquisition in the next 60 days. And
41:37I don't see it rallying 10 or 20% in the
41:40next 60 days given where we are right
41:42now. Uh and that we're coming into the
41:45end of the year. There is more value in
41:47taking the loss and not paying the tax
41:50on that realized gain than there would
41:52be uh in in holding it. I would
41:55encourage you to do the same thing in
41:57your own portfolio is have a look at
41:59where you are in terms of your realized
42:02gains or losses. If you're sitting on
42:05realized gains and you have in your
42:08portfolio some names that you've been
42:10just dragging along and looking at,
42:13there is an immediate tax savings by
42:15taking the loss. and ask yourself over
42:18this period of time, the next 30 or 60
42:21days, uh, will it rally 20%. Because in
42:24the IT space, in the AI space, I can
42:27find names that I think will rally 20,
42:2830, 40% by the end of the year.
42:32So, why hang out in a name when I could
42:35get a tax savings immediately and get
42:38into something that has the potential
42:40for upside more than this? And right
42:42now, I mean, I have I have stuck with
42:45Capri these last two years. And uh I'm
42:49just not seeing any signs of life and
42:52after their last earnings,
42:54mediocre earnings in the backdrop of a
42:58Bank of Canada that's raising rates
43:01uh and an administration that is not
43:03changing its immigration numbers, I
43:05don't see the catalyst for 10 or 20%
43:07upside to Capri. Before we get into the
43:10uh earnings, let's have a look at uh
43:12Lummentum and Marll.
43:15They announced uh they're getting
43:17married on Monday.
43:19Um which didn't make any sense to me at
43:22all of why you would um want to buy Lum
43:28such a high valuation when you could
43:30have just uh entered into a joint
43:33venture to get something done without
43:35having to do this. Since that
43:37announcement, both of them have been cut
43:39almost in half. 50% loss, almost 50%
43:43loss week overw week,
43:45the deal has to be approved by both
43:47Lentum shareholders and Marll
43:49shareholders. And I seem to think if I
43:51was a Lumenum shareholder, it would be a
43:53resounding no. I can't see why any
43:56Lumenum shareholder would vote in favor
43:58of this. Once it was announced, Lmentum
44:02had to uh issue its uh its uh numbers
44:05for the quarter. It was scheduled to
44:07report on Thursday. It issued it Monday.
44:10No press conf uh no press conference. No
44:12um no investor call. I can't find any
44:15investor call on their website at all.
44:17So, usually you cancel those when you
44:20intend to do this. You say, "Look, we're
44:22we're getting this done. We intend to do
44:24it." But you got to get through a
44:25shareholder vote. And I can't see it
44:27happening
44:29for Marll. I would seem to think, yeah,
44:31let's vote in favor of this. It can only
44:33help us. But for Lmentum, I think it can
44:35only hurt. We won't get any voice on
44:39this until December 1st when Marll
44:41reports. Um, and then there could be
44:44some questions there as to why this, why
44:46now, what's going on? What do you see
44:48that nobody else sees? Maybe there's
44:50some deal in the works uh that is going
44:53to make uh uh both of these companies
44:57incredibly profitable and incredibly
44:59rich and and that there's great
45:02synergies in working together and we
45:04just don't know it yet. I don't know
45:05because it would seem to me there was an
45:08interview on CNBC with the CEO of uh
45:11Lumenum and he said, "We never even
45:12thought about anything but an all stock
45:14deal."
45:16What? you never thought of anything.
45:18Like to me it just seems rather odd that
45:21this is going on. You almost think,
45:23okay, okay, somebody has naked pictures
45:26of somebody somewhere, right? Cuz who
45:28would ever agree to this? Let's just uh
45:30look at Lummentum's quarter
45:33uh and their current valuation which is
45:36compelling at this point
45:39because their stock price is being cut
45:41in half and their um results showed uh a
45:45significant improvement. So the
45:48fundamentals significantly improving and
45:50the price being cut in half makes the
45:52valuation compelling enough that I would
45:54say this is not a short candidate
45:56anymore. But Marll still is cash of
45:59almost 2 billion, debt of about 1.4,
46:01market cap of uh 55 billion gives you
46:05enterprise of 5437. I'll tell you how I
46:08calculate EBIDA and EPS in a little bit.
46:10But uh Q2 run rate on EBIDA 2.348
46:14billion gives enterprise uh value at
46:17EBIDA run rate of 23.16. That's at the
46:19high end because this is a run rate, not
46:21a growth rate. So it's just assuming Q2
46:24is looks like this instead of having
46:26growth to it. So this would be the worst
46:29condition 23.16
46:31EPS of 1620. This is a run rate for the
46:34year. So a PE of 33.35 a little on the
46:37higher end.
46:40Um but again just a run rate.
46:44They do have operating leverage um
46:46year-over-year 38.6. You can't really
46:49count that because you're using 6.7
46:51million against 360 quarter over quarter
46:541.13. I think we can go with that. Let's
46:58look at what they did on revenue
46:59year-over-year 136%.
47:02Quarter over 25.5. That's a run rate of
47:05over 100% a year. Uh Q4 uh let's look at
47:10their margins. Uh cost of sales 47.48.
47:13Last year was 62.34. So year-over-year
47:16they improved by almost 1,500 basis
47:18points. Q4 was 50.67.
47:22So quarter over quarter their gross
47:24margins incre increased by 319 basis
47:27points. That's a positive. Gross profit
47:3051%.
47:32Uh because you have your amortization in
47:34here 51% up from 34%. Q4 was 47.73.
47:39So another good uh uh number for this
47:42quarter. Here is a question here. R&D uh
47:46went from 81 million to 191 135.6%
47:51year-over-year
47:53uh in line with your revenue growth 136%
47:57but look at quarter over quarter you had
47:5925% here 83% quarter over quarter within
48:03this number sharebased compensation was
48:05105.2
48:07million of the 191.8 date big sharebased
48:11compensation within there's something
48:13going on in R&D.
48:15So it sort of uh uh um uh skews their
48:20their costs. Uh it it could have been it
48:23could have been a lot better. I think if
48:25you look at total operating expenses for
48:26Q4 was 19.67 jumped to 28.5 because of
48:31something here. However, it did come in
48:34better than last year. 32.5
48:3632.75% of revenue down to 28.5, which is
48:40nice. But quarter over quarter. Ouch. Uh
48:43net income 28.53% of revenue versus 79.
48:48So that is good. And we're using this as
48:50our operating leverage here. Our share
48:52count increased 30%. And that is because
48:54of convertibles that happened in the
48:56previous quarter in Q4 of last fiscal
48:58year. 101.7
49:01um diluted $3.54
49:05up from 5
49:07quite a bit. Q2 guide revenue guiding
49:10the 1.6 billion which is 26.7%
49:14quarter over quarter. This quarter was
49:1625% over the last quarter. They're
49:18guiding to pretty much the same if not a
49:21little bit more for the next quarter.
49:23And they love to give non-GAAP. This
49:26just makes the work a little bit harder
49:27because you have to go to the schedules
49:29and and start with the GAP and say,
49:31"Okay, well, how did you get to non-GAAP
49:33and then I got to turn their non-GAAP
49:35into GAP?" Diluted EPS uh 605. If I take
49:39their GAP net income to their non-GAAP
49:42uh uh net income, uh GAP net income is
49:45rough running around 67% of non-GAAP.
49:47So, I'm going to multiply this by 67.
49:50Get to about $45
49:52on a GAP basis. We're at 354 for this
49:55quarter. Giding 26% higher on revenue.
49:59Going to 405 doesn't seem unreasonable.
50:01That's where I get the 1620 run rate. I
50:03multiply that by four. Operating margins
50:06non-GAAP 45.5.
50:09Uh Q1 the non-GAAP over here. Uh the
50:13non-GAAP operating margin for this
50:15quarter was 423. So they are guiding the
50:17higher operating margins which does make
50:19sense because you do have uh you do have
50:22some operating leverage in here.
50:25Uh so if we take our non-GAAP
50:28uh you have to add back sharebased
50:30compensation and the amortization of
50:33acquired intangibles to get to gap. Uh
50:36so let's calculate Ebid duh. If
50:39sharebased compensation and acquired
50:42intangibles gets you to uh a gap ebbit,
50:45then you have to add depreciation. So
50:47we'll take the 1.6 billion multiply by
50:49what we think the operating margin is
50:52and then subtract out uh the sharebased
50:55compensation because this 45.5 doesn't
50:58include it. Subtract out the uh
51:00amortization of acquired intangibles and
51:03I'm just using Q1's numbers and just
51:05saying they'll be the same for Q2. add
51:07back depreciation I get 606.3
51:10million for EBIDA for Q2
51:14another way of doing it because we have
51:16operating leverage is just to look at uh
51:19the Q1 EBIDA which is 359.8 adback
51:23depreciation uh and amortization you get
51:25the 436.2 2 uh for Q1 um we have
51:30operating leverage of 1.13 and revenue
51:33growth of 26.7% on the guide. So if
51:37revenue is going to increase by 26% then
51:39we have 1.13 times leverage. Multiply
51:42these two * 436 you get the 567
51:46versus 606. We're in the neighborhood.
51:48Let's take an average and we'll assume
51:50that our next quarter uh ABITa comes to
51:53about 587 million annualize that. That's
51:56where I get the 2 point uh the 2.348.
52:00So we could do this in one of two ways
52:02is if we have operating leverage and our
52:04revenue growth, we can take this
52:08period's uh EBIT uh increase it uh by
52:12the uh revenue times the leverage amount
52:15and add back depreciation
52:18uh or
52:20um reverse it from the non-gap to the
52:22gap by making the adjustments.
52:25Uh, if they come in in at about the
52:27same, I seem to think that if you take
52:30an average, you're probably in the good
52:32in a good neighborhood.
52:34Let's look at Marll. Um, what's my
52:36assessment of of of this company right
52:39now? Is at this point I would think that
52:41well, it's off my short list. I'm not
52:42shorting this, not with this, not with
52:44this growth rate, not with not with
52:47positive operating leverage. Uh the
52:49multiple is simply not high enough to be
52:52on the short list uh on either one of
52:54these. Um based on last quarter's
52:57earnings and when it was $1,000 stock,
53:00uh their enterprise value to Ebida
53:02multiple was sitting around 85. Well,
53:06that's you don't grow into 85. They're
53:10down to 23, but they're half the price.
53:12So, the short was the right thing to do
53:14on on Lmentum.
53:16Um, at this point, I'd almost be almost
53:19be a buyer of Lmentum at this level
53:22given the guidance that they gave. I
53:24would have loved a press conference. So,
53:26not press conference, an investor call.
53:29I would have loved that just to get a
53:31little bit more uh a little bit more
53:33detail, but I didn't get it.
53:36This is an all stock deal. Uh, every uh
53:38share of Lum you have, you'll get 4.5
53:42uh Marll shares uh for it. Marll closed
53:46at $141.95
53:48times 4.5 is $638
53:52of Marll's stock. Lumenum's trading at
53:55$540.
53:57So, you would seem to think, well,
53:58there's a big discount between the two.
54:01An arbitrage play would be to uh short
54:044.5 shares of Marll and buy one share of
54:07Lmentum
54:09uh because this gap has to be closed
54:11somehow. If you think this deal is going
54:14to be done, I don't think it's going to
54:16be done. I don't think you're going to
54:18have the momentum shareholders voting in
54:20favor of this thing. Unless you do a lot
54:23of selling at the board level and
54:26management level to investors, I don't
54:28see why I would ever vote for this.
54:31Um, unless I'm unless I'm getting shares
54:34of a company that is massively
54:35undervalued. Marll has earnings December
54:381st. Uh, November 2nd, they had to say
54:40something. So they maintained their
54:42revenue guidance at the midpoint 3.15
54:45billion. Uh previously it was plus or
54:48minus 5% they narrowed it to plus or
54:50minus 2%. So they're guiding 15% quarter
54:53over quarter doing a quick run rate
54:56enterprise value to Ebida. They're at 54
54:59times enterprise value to EBITa. 54 uh
55:03versus uh 23 for Lum. Their PE run rate
55:06is 70 times uh versus 33.5 for Lmentum.
55:12Um they are more than double the price
55:16per dollar of income. Per dollar of
55:18income you're paying more than double
55:21for $1 of income from Marll versus $1 of
55:24income from Lmentum. And Lumenum is
55:27guiding to the faster the faster growth
55:30rate.
55:3215% for Marll, uh 26.7%
55:36for Lum. Why would you ever agree
55:40to give up a lower valued higher growth
55:44share for a higher valued lower growth
55:48share even though I'm getting 4.5 times
55:50more? It almost seems like Marll needs
55:53momentum
55:55uh to support to support a higher growth
55:57rate, but momentum will end up getting a
56:00collectively lower growth rate with
56:03Marll involved.
56:06So, it should be trading at a premium,
56:09you would think, as opposed to a
56:11discount. $540 for a Lum share is worth
56:16$600 almost $640 for Marll shares.
56:22It's a very odd uh very odd thing. There
56:24are a number of plays in here. If you
56:27think the deal is going to happen, a
56:29long short, this is uh this is merger
56:32arbitrage. Uh if it's going to be 4.5
56:35for one, you would short 4.5 Marvel
56:38shares to buy one Lum share. And that
56:40gap between the 638 and the 540 has to
56:43be closed somehow. So you don't care how
56:45it gets closed.
56:47You don't care if they both drop in
56:49price cuz you're long and short. And you
56:51don't care if they increase in price cuz
56:53you're long and short. You only care
56:54that that gap gets closed. That's if you
56:56think the deal is going to close. I
56:58don't think it's going to close.
57:02If I were a Lum shareholder, I certainly
57:04wouldn't be voting for this. Would you
57:06have a look at these numbers? Would you
57:07be voting for this? I don't think so.
57:10Unless there is some compelling reason
57:13that they haven't disclosed, which they
57:15kind of have to, but they haven't
57:17disclosed. Maybe they know something we
57:19don't know. Maybe there's something in
57:20there, I doubt it, that could be this
57:22good. And for the CEO of Lmentum to say,
57:25"We never even once considered anything
57:28but an all stock deal." Or is that just
57:31something you have to say because it's
57:32an all stock deal?
57:35for Marll to make an all stock deal
57:37makes sense because they're massively
57:39overvalued or I shouldn't say massively
57:42but they they have an overvalued stock.
57:45There's a reason I was shorting them at
57:47260 and 270 is because at 260 and 270
57:51these numbers were even worse. Once you
57:53get to uh an enterprise value uh to
57:56Ebida multiple of 90 to 100 times you
57:59cannot grow into that. That is always a
58:01short. And even even uh being down at
58:06142 from 260 270, you're talking almost
58:0950%, they're still at 54 times.
58:14Um from Arll's perspective, yeah, let's
58:16use our overvalued shares to buy stuff.
58:19Um but from Lumenum's perspective is why
58:22would you ever accept this?
58:26You know, I could see if it was the
58:27other way around that Lum decided to buy
58:30Marll. It's like, okay, well, I guess
58:33maybe, but this to me seems rather
58:36silly. So, we don't think it's going to
58:38happen. What's the play here?
58:41I seem to think the plate is uh a
58:44rebound in Lmentum stock,
58:48but we're told uh it's not closing till
58:51the first half of 2027. That is a long
58:55way to go.
58:59My suspicion is that this will probably
59:03be called off before the end of the year
59:06because both stock prices puked on this
59:10by about almost 50%. You can't be happy
59:14if you're a Marll shareholder. You can't
59:16be happy if you're a Lmentum
59:18shareholder. There is no cash component
59:21in here. It would have been better if
59:22there was, but there isn't. It's all
59:24stock, so it's how low can it get. As
59:27Marll continues to sell off, which I
59:31mean given its valuation, it could. We
59:33don't know what it's going to say on
59:35December 1st. Uh if it sells off, then
59:38the value of momentum just keeps going
59:40down.
59:43I uh
59:45I put a low probability on the deal
59:47happening. So, I think that this
59:48arbitrage trade isn't going to be there.
59:50If you think it is, if you have some
59:51information that it is, obviously
59:53there's money to be made because there's
59:55about a $100 difference
59:58uh between the price on this. The
1:00:01acquired is selling at a $100 discount
1:00:04to the acquiry. If you don't think it's
1:00:06going to happen, I think Lum has a
1:00:08better chance on the upside uh than
1:00:10Marll does at this point. We're not
1:00:12going to hear anything till December
1:00:131st. Okay, we had uh Fabrinet FN. I
1:00:18should put their ticker in the top
1:00:20corner here. FN had this early uh in the
1:00:23week. Let's look at their valuation
1:00:25first. So, we put everything in
1:00:27perspective. Cash of 755 million, debt
1:00:30of 74 million. That was a debt that was
1:00:33just raised this quarter. Market cap of
1:00:3612.6 billion. Enterprise values just shy
1:00:38of 12 billion. the Ebida Q2 this is next
1:00:42month's run rate um 248.5
1:00:47uh 248.5 billion I took their revenue
1:00:51and I gave them their margin they're
1:00:53running about 9.9 last year was 10.2 2.
1:00:56So I said okay well let's give him 10%.
1:00:58Uh let's annualize it add back
1:01:00depreciation 248.5
1:01:03enterprise value to Ebida 48 times
1:01:07uh with very thin net income margins
1:01:09you're talking around 10% to have an
1:01:12enterprise value EBA multiple that high
1:01:14when Fabernet was $5 and $600 it was an
1:01:18easy short it was there's no way with
1:01:20margins that low that you'll live into
1:01:22that multiple. So on a valuation
1:01:24perspective alone, even if they did
1:01:26well, they still couldn't justify that
1:01:29multiple 48,
1:01:31that's still not interesting to me. Uh
1:01:34EPS Q2 run rate, uh 22.36. This is a
1:01:38little bit more reasonable uh if you
1:01:40value them on an EPS basis. Value them
1:01:43on an enterprise value basis, little a
1:01:46little bit pricey.
1:01:48Guiding revenue uh for the next quarter,
1:01:511.625 billion. That's 43%
1:01:54year-over-year, 10% quarter over
1:01:56quarter. I'm always nervous when I
1:01:58calculate the growth rate and I get to a
1:02:00number that is so round and I think well
1:02:03there wasn't a lot of bottom up analysis
1:02:05here. It was like well we we can grow by
1:02:0710%. And it comes to just 10%. There's a
1:02:10couple other companies I did where the
1:02:12numbers their forecast were so round you
1:02:16know like you know.1300
1:02:19is like okay that's 13%. It's rare that
1:02:22you get it right on the button. It's
1:02:24almost as if that that's the number you
1:02:26started with to get your forecast. Like
1:02:29how much can you grow revenue? I don't
1:02:30know. 10%. Well, there's our forecast
1:02:3210%. GAP EPS 3.885 the midpoint 36.2
1:02:37million shares. They are adding more
1:02:40capacity. So, their capex is uh
1:02:42increasing. Free cash flow for the
1:02:44quarter negative 148 million. uh this
1:02:48new site in California that was a big
1:02:50jump in their uh in their capex but once
1:02:53you took out you start with free c the
1:02:55cash from operations take out sharebased
1:02:57compensation that was 11.25 25 you got
1:03:00about 10 million uh in free cash flow
1:03:03really from an enterprise value
1:03:04perspective minus your capex you are uh
1:03:07you have negative cash flow
1:03:10uh share repurchase uh increased
1:03:13uh $103 million to 637.3
1:03:18million which is 5% of its market cap
1:03:21that it is authorized to repurchase.
1:03:23They did repurchase $30 million in the
1:03:26quarter.
1:03:28Um they are talking about the CPO, NPO
1:03:30and OES markets. Uh without giving any
1:03:32numbers just sort of saying each of them
1:03:35uh could be multihundred uh multiundred
1:03:38million dollar businesses
1:03:40without giving much more guidance. It's
1:03:42this combination of how close can you
1:03:44get to the 10.1000.
1:03:47It's like okay well clearly you're
1:03:49guiding to 10% growth at the midpoint.
1:03:52You just took this number and said well
1:03:54let's put 10% on.
1:03:56Uh, so I would have liked to have more
1:03:58detail here. I don't really care too
1:04:00much for this valuation. Let's have a
1:04:03look at some of their numbers. Uh,
1:04:04revenue year-over-year 50.74%.
1:04:08Highest growth in 10 years they claim
1:04:1012% quarter over quarter guiding to 10%.
1:04:14They did 12, they're guiding to 10.
1:04:16That's a step down. Uh, gross profit
1:04:1811.78.
1:04:20Uh, last quarter was uh, sorry, last
1:04:22year was 11.9. So, you lost 12 basis
1:04:25points year-over-year. Last quarter was
1:04:2712.01, so you lost 23 basis points
1:04:30quarter over-arter.
1:04:32Operating income 9.9% of revenue. Last
1:04:36year was 9.62, so you did have a pickup
1:04:38year-over-year, but you went from 10.2
1:04:41last quarter to 9.9, lost 12 basis
1:04:44points. Operating income is up 55%
1:04:47year-over-year. So you do have some some
1:04:50operating leverage here. If your uh
1:04:52revenue is up 50% but your operating
1:04:54income's up 55, you got a little bit of
1:04:56operating leverage here. 1.08
1:04:59quarter over quarter 8.76.
1:05:03You know, if your revenue is up 12%
1:05:05quarter over quarter and your operating
1:05:07income's up 8.76, that's negative
1:05:09leverage.
1:05:11Um sitting in here, this complicates
1:05:14things. You have an unrealized loss and
1:05:16a minority interest. Um so for uh June
1:05:20uh uh sorry the last quarter you uh had
1:05:23a gain that kind of distorts numbers. So
1:05:26really the only thing that has value
1:05:28here are uh these if we go into net
1:05:31income
1:05:33um for this year 8.78%
1:05:36of revenue. Last year was 9.8%. You lost
1:05:40112 basis points year-over-year. But you
1:05:42can see where it is here. It's just
1:05:44unrealized loss. uh last quarter 10.6%.
1:05:48So you're down quarter over quarter by
1:05:50quite a bit. Your net income grew uh a
1:05:54little under 8% year-over-year, but is
1:05:57down uh 20 almost 26% quarter over
1:06:01quarter. Last quarter you did 139
1:06:04million, you're down to 103 million.
1:06:06Even giving this back,
1:06:08uh you're still down quarter over
1:06:10quarter on your net income. Diluted uh
1:06:142.86. They love to give you non-GAAP.
1:06:17Every company wants you to pay attention
1:06:19to non-GAAP. Now, I just pay attention
1:06:21to GAP. Share count increased
1:06:23year-over-year, but it's down quarter
1:06:25over quarter. They are buying back uh
1:06:28they are buying back stock. Overall, I
1:06:30would say I get to this point, I go,
1:06:35I'm un, you know, I mean, okay, fine.
1:06:37It's it's profitable. It's doing its
1:06:40thing. Negative cash flow, very thin
1:06:42margins overall. uh to be to have a
1:06:45multiple enterprise value at a Eban
1:06:47multiple at 48. Your margins are way too
1:06:50thin. This wouldn't be on my buy list.
1:06:53Uh given it's uh uh uh the the PE I said
1:06:57EPS here. This really should be PE your
1:06:59PE of 22.36. I don't know that I would
1:07:03say I I'm definitely shorting this
1:07:05thing, but I wouldn't be buying this
1:07:07thing. This one here, I would just I
1:07:09would just avoid.
1:07:12Um, but for the overall picture of where
1:07:14we are in the AI trade, data center
1:07:17products up 78% year-over-year, 16%
1:07:20quarter over quarters is 53% of the
1:07:22revenue. If you're looking for uh the
1:07:24growth story in Fabinet, it's right
1:07:26there. Even communications came in
1:07:29strong, 35% year-over-year, which is
1:07:32pretty good. Not quite the 50% in
1:07:35revenues underperform the revenue
1:07:36number, but 8% quarter over quarters,
1:07:3930% of revenue. and automotive 20%
1:07:42year-over-year, 8% quarter over-arter,
1:07:4517% of revenue. Um, average performance,
1:07:50maybe slightly above average performance
1:07:51here. But, uh, your data center product,
1:07:54uh, that is the, uh, that's the big
1:07:56growth driver there, 78% year-over-year.
1:07:59Worth noting that both Lum and Marll are
1:08:03customers of Fabernet. Okay, AMD,
1:08:08let's start with their valuation before
1:08:10we uh dive into any numbers. Uh 15.7
1:08:14billion in cash, 7 billion in debt,
1:08:17market cap of 856 billion gives them an
1:08:21enterprise value of 847 billion. Of the
1:08:257 billion in debt, they added uh 4.75
1:08:28billion in August. Uh so that was the
1:08:31jump up from last quarter.
1:08:33Abida Q4 uh run rate
1:08:37uh because that is their next quarter
1:08:38they're guiding to 19 uh 19.04 billion
1:08:41on Abida enterprise value to EBIDA 44.51
1:08:46uh and the PEQ4 run rate 48.881 81.
1:08:49We'll look at how we got these numbers
1:08:51down here. That's based on a closing
1:08:53price of 51552. When I did this, they're
1:08:56well below that now. So that means these
1:08:59multiples are lower at $600. Their
1:09:02enterprise value at EBIDA was 52 based
1:09:04on these numbers would would have been
1:09:0652. Their PE56 is still would have been
1:09:08a short candidate even with this good
1:09:10quarter. Uh that still would have been a
1:09:12short candidate. These are still uh
1:09:15rather elevated. However, that was at
1:09:17515.
1:09:19I think it had traded down to 440 450
1:09:22somewhere in that range. Uh I am I am
1:09:26long AMD now. I did add AMD because it
1:09:29is it is compelling. Although they do
1:09:31have uh it's it's it I wouldn't say it's
1:09:35the best bet I have out there. I think
1:09:37Nvidia and Broadcom are the best bets
1:09:39along with Micron, but it's no longer a
1:09:42short candidate based on the price of
1:09:44where it's at.
1:09:46Uh it is guiding revenue to $17 billion
1:09:49which is a 27% quarterover growth. Um
1:09:54all of this by the way is non-GAAP.
1:09:55Gross margins of 55 uh 55%.
1:10:00Um that's 110 basis points quarter over
1:10:03quarter. OPEX of 3.75 other income of
1:10:06about 4 million. A tax rate of 13%
1:10:08diluted share count of 1.67 billion. So,
1:10:12because this is non-GAAP, to get to GAP,
1:10:14I'm going to need some amount for
1:10:16sharebased compensation. This is for a
1:10:18quarter half a half a billion dollars. A
1:10:21quarter two billion a year. Uh EPS $2.64
1:10:25a share. So, let's get Q4 EBIT. We'll
1:10:29take our revenue times our margin 55 and
1:10:33uh subtract out uh the amortization
1:10:36um that goes into uh cost of goods sold.
1:10:40And then let's deal with our operating
1:10:42expense. The uh 260 that's in COGS.
1:10:46Uh you also have amortization that is in
1:10:48opex. So we have 3.75 billion. We're
1:10:51going to subtract out the 3.75 plus the
1:10:54um sharebased compensation. I rounded up
1:10:57to 540 figuring there'll be growth. And
1:10:59then the amortization of purchased
1:11:01intangibles. Uh add back your
1:11:03depreciation 4.76 billion. Does that
1:11:06track? Uh it is a 68% quarterover
1:11:11increase. Uh you are guiding to a 27%
1:11:14revenue increase with a 68%
1:11:16quarterover-arter increase uh in Ebida.
1:11:20Giving you some uh operating leverage
1:11:222.51. Um keep in mind though I am adding
1:11:25depreciation back to get that there.
1:11:27Operating leverage quarter over quarter
1:11:29uh was 2.65. This is 2.51. I'm in the
1:11:33neighborhood. If the number were
1:11:35radically different, I'd say, "Okay, I
1:11:37think I missed something somewhere."
1:11:39But, uh, I'm I'm I'm in around two and a
1:11:41half times on the leverage. This is what
1:11:43I'm getting to on just a quick back of
1:11:45the envelope calculation. So, I think,
1:11:47okay, that's a sanity check. I'm in the
1:11:49I'm in the right neighborhood. Our free
1:11:51cash flow for the quarter, 31 million.
1:11:54Uh we haven't accounted for any warrants
1:11:57in any of this because if there are
1:11:59warrants that are um that are
1:12:02exercisable, I would have to use my free
1:12:04cash flow to buy back the shares that
1:12:06are created because I am taking an
1:12:07enterprise value perspective, right? I
1:12:09must own the capital stack.
1:12:12Let's get into their numbers now. Now
1:12:14that we have sort of this is a
1:12:16background, we have an idea of what
1:12:17their valuation is as we're going
1:12:19through this. If we're seeing big huge
1:12:21fat margins and we end up with 50 60 70%
1:12:24net income margins uh and incredible
1:12:27growth, we might say, you know what,
1:12:29it's warranted. Uh but the thinner the
1:12:31margin we get to on net income, the less
1:12:33warranted, higher multiples are. Revenue
1:12:37uh up 16% quarter over quarter and
1:12:40they're guiding to a 27% quarter over
1:12:43quarter showing accelerating revenue
1:12:45growth. That's a positive. that will
1:12:47that will push a multiple up
1:12:49year-over-year 44.7%.
1:12:52Uh if we look at their balance sheet to
1:12:54see if we have any cracks there, uh with
1:12:56revenue up 40 almost 45% year-over-year,
1:12:59accounts receivable are only up 26%,
1:13:02inventory is only up 35%. So it's not
1:13:05tracking the growth in revenue. So we
1:13:07have good management on the balance
1:13:09sheet accounts. AMD called this uh or
1:13:12referred to this as their largest most
1:13:14profitable quarter in AMD history.
1:13:16Largest and most profitable quarter. Um
1:13:20backing out uh their revenue um from if
1:13:24you look at their segments backing
1:13:25everything out and trying to figure out
1:13:27what they sell per gigawatt roughly
1:13:29about 25 billion a gawatt versus Nvidia
1:13:32somewhere around $40 billion a gawatt.
1:13:36uh and I think u Marll
1:13:38uh or was it Broadcom? One of those two
1:13:40are somewhere in around 2020 20 billion
1:13:43dollars for their XPUs uh per per
1:13:46gigawatt but they're around 25 billion.
1:13:49So from a price perspective it it
1:13:51appears to be compelling.
1:13:53Looking at total cost of sales uh 46.1%
1:13:59uh down from 46.2. You say that's
1:14:01probably flat on cost of goods sold.
1:14:04They're probably as efficient as they're
1:14:05going to get, but uh uh certainly uh
1:14:08down from the 483 of last year down to
1:14:11461, I don't know that there's a lot of
1:14:13room in cost of goods sold. If you have
1:14:16memory prices coming down and and
1:14:19component prices coming down, cost of
1:14:21goods sold could really increase. It
1:14:22would benefit an AMD in that sense.
1:14:25Gross profit 539,
1:14:2853.9 on a gap basis. They're guiding to
1:14:3055 on a non-GAAP basis. they really only
1:14:33have uh two small categories uh in
1:14:36there.
1:14:38Um operating income uh 21.17
1:14:43uh significant increase from last year
1:14:46at 13.73 and quarter over-arter uh 17.25
1:14:50expanding to 21.17.
1:14:53Uh 744 basis point improvement in
1:14:56operating income year-over-year 392
1:14:58basis points quarter over-arter.
1:15:01uh operating income's up 42.3%
1:15:04quarter over-arter. So we can calculate
1:15:06some operating leverage. 16% uh growth
1:15:09in revenue for a 42% uh growth in
1:15:12operating income gives you 2.65 quarter
1:15:15over quarter. That is that is a robust
1:15:18operating leverage amount. 123%
1:15:20year-over-year versus 44 year-over-year
1:15:22gives them 2.75 yearover-year. And
1:15:25implied in here was 2.51. we're in
1:15:28around the two and a half times on
1:15:29operating leverage. The higher the
1:15:32operating leverage, the higher the
1:15:34multiple you can justify for a given uh
1:15:37a given uh forecast or revenue. So if
1:15:39you say revenue will increase 20%
1:15:41quarter over quarter. Two companies
1:15:44identical in all respects except one
1:15:46company has higher operating leverage,
1:15:48they would justify a higher multiple in
1:15:50an efficient market. So you can start to
1:15:53look at the 445 and say okay well 2 and
1:15:55a half times operating leverage. Let's
1:15:58give them the benefit of the doubt with
1:15:59that. It appears fairly stable year over
1:16:01year quarter over quarter and on the
1:16:04forecast on implied for the forecast
1:16:07things uh fall apart once uh once you
1:16:09get past operating income because all
1:16:12these companies have all these interests
1:16:15all these warrants or all these other
1:16:17interests in other companies. So that
1:16:18shows below the line uh in this
1:16:21negative92 is a 1.1 billion warrant
1:16:24cost. As the price of AMD goes up, uh
1:16:27this uh loss in here also increases so
1:16:30that your uh um
1:16:34uh your uh uh income
1:16:36um uh below the line begins to drop.
1:16:39Look at it's 13.9% of revenue. Last
1:16:43quarter was 22.11 because it didn't have
1:16:46any warrant expense in there. Um, last
1:16:49year was 14.22.
1:16:52So, you're slightly below last year, but
1:16:53a big sequential drop quarter over
1:16:56quarter. The more successful the stock
1:16:58price of AMD is, the uh more it gets hit
1:17:01uh down here because it shows up here
1:17:03and it will show up in a higher uh uh
1:17:07share count down here.
1:17:09uh net income uh 18.15%
1:17:13certainly better than Fabernet which we
1:17:15saw earlier sitting under 10% it's it's
1:17:18double that but compare that with Nvidia
1:17:20that runs around 60% on that line item
1:17:24three times more net income per dollar
1:17:27of revenue out of Nvidia than you're
1:17:30getting out of AMD at a valuation much
1:17:33much lower their enterprise value to
1:17:35EBITA is below 15 below 15 This is 44.5
1:17:42uh 18.15
1:17:44uh sequentially down from 19.91
1:17:47again uh the warrants right u and up
1:17:51from 13.44 a year ago. Diluted earnings
1:17:55per share up 6% quarter over quarter
1:17:57944% year-over-year. Your share count is
1:18:01growing and will continue to grow. Uh
1:18:04you have roughly 20% dilution. So this
1:18:06can grow to uh two billion up from 1.661
1:18:10billion. So you have to have significant
1:18:14increases in net income to overcome that
1:18:17increase in share price. But the higher
1:18:19the share count goes, the higher this
1:18:21warrant expense goes, the lower uh the
1:18:24uh the income number here is before you
1:18:27take taxes and calculate your uh your
1:18:30earnings per share. It's going to be
1:18:31tough to see how they get a uh a
1:18:35multiple a PE multiple that uh comes
1:18:37down unless that stock price comes
1:18:40really down. But maintaining the stock
1:18:42price are going higher, those big
1:18:44purchases have to come in.
1:18:47Um let's look at their segments. Data
1:18:49center revenue for the quarter was 90
1:18:5296.86%
1:18:54year-over-year growth, 27% quarter over
1:18:57quarter. noted that uh this growth
1:18:59accelerated every quarter this year. Uh
1:19:03last year they had guided uh sorry last
1:19:05quarter they guided to 80%
1:19:07year-over-year growth. It came in at
1:19:0996.86%.
1:19:11And this is how it's going to have to be
1:19:12done is your revenue is going to have to
1:19:15far exceed uh what the waiting of all of
1:19:19those warrants would be uh for the price
1:19:21to go up.
1:19:23Um this was an interesting line uh from
1:19:27uh Lisa Sue. Adoption is just starting
1:19:30when it comes to AI agents. Adoption is
1:19:33just starting. Um and they uh are
1:19:36attributing uh most of the beat to
1:19:39Agentic and Inference which is driving
1:19:42the results. Um Inference's CPU side. Um
1:19:48and they have CPUs. Intel has CPUs. Uh
1:19:50but so does uh so does Nvidia, the Vera
1:19:53uh is their uh their CPU. The Reuben is
1:19:56their GPU.
1:19:57Um client and gaming. Client is your
1:20:00consumer PC. Uh client up gaming down uh
1:20:05blaming on memory costs year-over-year
1:20:07eight that segment's down 8% quarter
1:20:10over quarter down 3% and embedded up 29%
1:20:14year-over-year 13% uh quarter over
1:20:17quarter.
1:20:18uh PC uh unit shipments um are
1:20:21forecasting to be down by double digits
1:20:24uh for the next uh quarter. Um same this
1:20:26is what Intel was saying as well. So
1:20:28it's consistent with what Intel is
1:20:30saying. However, they are outperforming
1:20:32Intel on this one although they are much
1:20:34smaller in that space.
1:20:37Uh 2027 visibility.
1:20:40Server revenue uh expect to be double
1:20:43what it was versus 2026. data uh center
1:20:46revenue uh growth. Uh they're
1:20:49forecasting higher data center growth
1:20:51than they did just 90 days ago. This is
1:20:53brutal, right? More than 80% revenue
1:20:57growth in 2027 for data center.
1:21:00Consistent with everything else we've
1:21:02seen from every company that has data
1:21:05center exposure is the growth is is is
1:21:08everything uh for their overall growth.
1:21:11Uh I haven't seen one company yet
1:21:15say we're lowering our guidance for
1:21:17growth for data centers or we're guiding
1:21:19flat for data centers. I haven't seen
1:21:21one yet. Uh it is all 80 90 100%. Nvidia
1:21:2670% across the board for the next two
1:21:29years. 70% growth.
1:21:31Um who did we look at last week? I think
1:21:34it was uh Qualcomm was forecasting uh uh
1:21:3860% growth year-over-year as well. And
1:21:41there was another company that was
1:21:42forecasting uh um I think it was like
1:21:4680% growth for the next two years. I do
1:21:48so many it's hard to keep them. It's
1:21:51hard to keep track of them. So what's
1:21:52the play on AMD? Um AMD is interesting.
1:21:56Uh, I just don't think I think dollar
1:21:58for dollar if I had to invest somewhere
1:22:01whether it be Nvidia, Broadcom or AMD,
1:22:04AMD would be third on the list. Does AMD
1:22:07outperform the broader market? That's
1:22:09the question because if it doesn't, then
1:22:11you should just buy triple Q as opposed
1:22:13to AMD.
1:22:15Um, I think it's I think, you know, as
1:22:18you start pushing up towards $600, it is
1:22:20it is clearly a short, but as you start
1:22:22pushing down towards 400, it's certainly
1:22:24not on the short list. Uh, I am a a an
1:22:28owner of AMD, but a very small amount. I
1:22:31think it's due for a bounce, but I'm
1:22:33certainly not a long-term holder of AMD.
1:22:37I think there are just um better
1:22:40companies to buy that give you more net
1:22:42income per dollar of revenue uh at the
1:22:46same or higher growth rates than what
1:22:48we're getting from AMD. Okay, let's have
1:22:50a look at Qualcomm. On the last screen I
1:22:52had uh sort of indicated that we had
1:22:55done this last week. I didn't not in the
1:22:58last week's market outlook. I did the
1:23:00screen last week when call reported and
1:23:03as I was narrating the previous screen I
1:23:05can remember doing call I just hadn't
1:23:08presented it. So here it is. Let's start
1:23:10with our valuation to put things in
1:23:12perspective. Cash 7.6 billion debt 15
1:23:15billion. So net debt on the balance
1:23:17sheet, market cap of 172, enterprise
1:23:20value of 180.
1:23:22EBIDA, I'm going to use a Q1 run rate
1:23:25because we just finished Q4. We're
1:23:27guiding uh 10.4 to 112. So take the
1:23:30midpoint of 10.8.
1:23:32Uh net operating income margin.2, which
1:23:35is what they did over uh 20 uh 26 fiscal
1:23:39year, down from 28% the year before. Uh
1:23:43the last quarter was 13. I think that's
1:23:46too low. I'm going to give them the 20.
1:23:47Annualize it. Add back depreciation
1:23:4910.29.
1:23:52This was 8.6 uh billion. Uh next over
1:23:56the next 12 months 10.29. That tracks.
1:23:58That's not unrealistic. It's not like
1:24:00it's 60 80 90 100% higher. So we'll use
1:24:04that. We get a multiple 17.52
1:24:08using their uh earnings. And here I'm
1:24:10going to use non-GAAP. uh reconciling it
1:24:13back to gap just looked like a lot of
1:24:15work. So said okay let's just use gap on
1:24:17this one with the caveat that keep in
1:24:19mind this is a non-GAAP multiple 17.79
1:24:23it's not uh that for if you gave me
1:24:26these numbers uh I'd say well I don't
1:24:29think that's a short candidate based on
1:24:32valuation it's not a short candidate
1:24:35free cash flow uh CFO minus sharebased
1:24:37compensation minus capex I get the 5.2 2
1:24:40billion 12% of revenue as a percent of
1:24:43market cap it's 3%
1:24:46you know it's not great but keep in mind
1:24:48I am taking an enterprise value
1:24:49perspective so I add back this $3.4 4
1:24:52billion. Uh if you don't add it back,
1:24:55obviously uh it'll be a higher percent
1:24:57of revenue and a higher uh free cash
1:24:59flow uh to market cap. Let's look at
1:25:02their numbers. Uh they report gap and
1:25:05non-GAAP revenue which may uh be kind of
1:25:08curious. Uh it is how they account for
1:25:10the warrants these warrants they gave
1:25:12for Amazon. Uh it is a reduction of
1:25:15revenue. So there was 120 million
1:25:17reduction in revenue. Uh so they report
1:25:2010.322
1:25:22equipment and services down 8%
1:25:23year-over-year. Licensing down 10% and
1:25:26total revenues uh down 8.4%.
1:25:30That was because of uh losing Apple as a
1:25:33customer. Looking at their cost of
1:25:35revenue
1:25:37uh and that really only uh is reflected
1:25:39with equipment and services because
1:25:41licensing is just margin. Uh 57.2% 2% up
1:25:45from 52%.
1:25:47And uh that's probably going to be under
1:25:50pressure for as long as Amazon is
1:25:52exercising warrants on this one because
1:25:54it will be a reduction to revenue but
1:25:57not a reduction to cost of revenue. So
1:25:59the percentage will will increase.
1:26:02Operating income down 53%
1:26:04year-over-year, 1.3 billion down from
1:26:072.9 billion. This was almost 26% of
1:26:10revenue down to 13% of revenue.
1:26:14Share count is decreasing year-over-year
1:26:16because we do have repurchases down
1:26:172.9%. That is a positive return of uh
1:26:21equity uh capital. Uh we had share
1:26:24repurchases and dividends amounting to
1:26:2611.67
1:26:28billion. Uh that was for all of fiscal
1:26:30year 2027.
1:26:33Let's break down their categories.
1:26:34Handsets came in at 5.14 billion of
1:26:37revenue down 26% year-over-year.
1:26:41That is 1.8 8 billion uh $1.8 $8 billion
1:26:4557% of revenue versus 71% last year they
1:26:49blame two things Apple being the big one
1:26:51but memory uh so they do expect even
1:26:54with the loss of Apple going forward
1:26:56they expect this number to be down low
1:26:59to mid teen percentages because of
1:27:01memory costs there again memory costs
1:27:04not one single company I've listened to
1:27:08is forecasting flat or decreasing memory
1:27:10costs every single one says the same
1:27:13thing memory costs are going to be
1:27:15higher. Uh yet Micron is trading at
1:27:18three and a half times next year's
1:27:20earnings. Three and a half. I get that
1:27:22it's cyclical, but 35% long-term
1:27:25agreements in place and they're trading
1:27:26at three and a half. That is the wrong
1:27:28price.
1:27:30I could see if half the companies are
1:27:32saying, "Yeah, you know what? We see uh
1:27:34memory at flat to lower and another half
1:27:37saying flat to higher." But every single
1:27:39one every single one higher higher
1:27:44automotive big jump year-over-year IOT
1:27:47uh 4%. Uh QCT which is their first line
1:27:52and licensing is QTL 9 billion. This was
1:27:56at the high end of their guidance range,
1:27:58they want you to know. And QTL down 7%
1:28:02year-over-year guiding uh QCT to be 9.3
1:28:06billion, up from 9.02 billion and QTL
1:28:10from 1.3 up to 1.4 for the next quarter.
1:28:135 billion data center target for all of
1:28:162027 with 650 million of that 5 billion
1:28:20showing up in Q1. They expect data
1:28:23center to be a 15 plus billion dollar
1:28:26market for them uh by 2029 annually by
1:28:30by by 2029 and forecasting nonhandset
1:28:34revenue growth of greater than 60%.
1:28:38Um they did talk about uh SpaceX and
1:28:41their 5G. Uh you know given that it's
1:28:44SpaceX,
1:28:46given that it's Musk um I'm not going to
1:28:50call that a contract. I'm going to call
1:28:51that an option that maybe maybe at some
1:28:55point maybe there's something there, but
1:28:56I'm certainly not going to give any
1:28:58value to it. Really, not yet. Uh let's
1:29:01call it a nice little call option on
1:29:04something uh for the communication side.
1:29:07Uh Amazon September 3rd, that is a deal
1:29:10uh for uh inference chips and
1:29:12connectivity products uh at the 1.6 6
1:29:16terabyte uh per second uh level plus a
1:29:1910-year warrant for up to 25 million
1:29:22shares which is about 2.4%
1:29:25of their outstanding float. They
1:29:27decreased it year-over-year by 2.9. This
1:29:30is for 2.4% of the outstanding float
1:29:32over 10 years.
1:29:35So the dilution effect of this can
1:29:37probably be well managed. Um and this is
1:29:40at 16126 a share. So the real cost to
1:29:44Qualcomm shareholders only comes above
1:29:46the 16126
1:29:49uh if Amazon exercises everything and
1:29:52holds it and pays for it. Uh this is a
1:29:55raise of four 4 billion in proceeds to
1:29:58Qualcom. 3.75 million of those shares
1:30:02vested on September 3rd and vesting will
1:30:05be in tranches. The full vesting occurs
1:30:08at $60 billion in revenue over those 10
1:30:11years which is a run rate of about $6
1:30:13billion a year. Uh if we look at uh
1:30:16their revenue last year at 43 billion
1:30:20uh 6% is 17th of that 14 roughly about
1:30:2414% of their revenue per year. And at 6
1:30:27billion a year it makes up for the $1.8
1:30:30billion hit uh from handsets from uh
1:30:33from the loss of Apple. So it uh it it's
1:30:37on net a little bit incremental. It is
1:30:40accounted for as a reduction of revenue.
1:30:43Uh and the reduction of revenue is not
1:30:45based on what they buy. It's based on uh
1:30:47the value of the warrant above the 16126
1:30:51price. So if they buy $6 billion worth
1:30:54of product over the next 12 months and
1:30:57the stock price of Qualcomm stays below
1:31:00161.26, 26 there is no recognition of a
1:31:03hit to revenue at all.
1:31:06Uh but if they run up to $250 $300, well
1:31:09good for the the the shareholder, but
1:31:12you will have you will have a reckoning
1:31:14on uh on reporting day when they uh when
1:31:17revenues are lower by the amount that
1:31:20the stock price exceeds 16126.
1:31:24Uh so there'll be some some correction
1:31:26in place uh at that point but small
1:31:29enough at six billion a year all of it
1:31:32represents 2.4 billion so you're talking
1:31:34about maybe 0.24%
1:31:37per year if it is if if if if it does
1:31:40take the full 10 years to get done. Uh
1:31:44it is small enough that it might escape
1:31:47a lot of notice.
1:31:50Uh overall, given their multiple uh I
1:31:53think they're worth owning some uh uh
1:31:56exposure uh to it and I have a SpaceX
1:32:00call option on 5G.
1:32:03So if you are long uh AT&T and you're
1:32:07long call, you have a good hedge against
1:32:09being long AT&T. If in fact SpaceX is
1:32:12going to do something, they're going to
1:32:14spend a significant amount on 5G rollout
1:32:16which would benefit Qualcomm. So there
1:32:18is your hedge against a long uh along
1:32:21AT&T perhaps. All right, last one here.
1:32:24Coherent. We're going to see the same
1:32:27story in data center. Start with their
1:32:31valuation cash of 1.75
1:32:34billion, debt of 4 billion, market cap
1:32:36of 38.9
1:32:38gives them an enterprise value of 41.253
1:32:42billion. their debt increased by
1:32:44threequarters of a billion dollars in uh
1:32:47Q1.
1:32:49Enterprise value to EBIDA, we're going
1:32:52to use a Q2 run rate. I get to 607
1:32:55billion, and I'll uh do that down here.
1:32:58Uh 16.97
1:33:01uh for the EV to EBITA uh and EPS uh Q2
1:33:06run rate. Uh they guide to a non-GAAP
1:33:09EPS of 238.
1:33:12Uh this quarter's non-GAAP EPS was 209.
1:33:15So I'll just take the percentage
1:33:17increase in this and I'll multiply by Q1
1:33:20uh uh gap EPS. I get to 216. Annualize
1:33:24it uh 22.3 times. Free cash flow is not
1:33:28a pretty picture already. Uh negative
1:33:31operating uh operating cash flow right
1:33:33off the start. Take out sharebased
1:33:35compensation, take out capex, negative
1:33:38almost a billion dollars in in Q1.
1:33:41That's uh you know if you're pointing
1:33:43out something negative here. There you
1:33:45go. Let's see how we got to our six
1:33:47billion. Uh they love everyone likes to
1:33:49guide as non-GAAP. Uh so you do have to
1:33:54um put some things in here before um
1:33:56coherent. Um look what they uh do over
1:33:59here. Interest expense. So they give you
1:34:02revenue and then they give you all of
1:34:04your costs and then they just give you
1:34:06EBT, earnings before uh income tax EBT.
1:34:11So you have to make the adjustment for
1:34:12the interest expense to get to EBIT a
1:34:15bit and then add back D. So we have to
1:34:17add back interest and then add back
1:34:19depreciation amortization. So they're
1:34:21guiding midpoint 2.65 which is a growth
1:34:25of 11.8%.
1:34:28They are telling us that fiscal year
1:34:302027 will be uh revenue will be 60%
1:34:34greater than 60% versus fiscal year uh
1:34:372026.
1:34:39Um, how do I get to our EBT? I'm taking
1:34:42this quarter's EBT, 358 million. There's
1:34:46the growth rate in revenue, and I
1:34:48calculated a total leverage because it
1:34:50is the growth rate uh of our EBT
1:34:55uh divided by the growth rate of our
1:34:57revenue. This is total leverage now
1:34:59because we have uh our interest expense
1:35:02in here. So, I took the growth rate in
1:35:04revenue times the total leverage to get
1:35:06to uh what EBT earnings before tax would
1:35:09be. Let's put the interest expense back
1:35:11in. I'm just going to keep it at the
1:35:13same 464
1:35:15to get to uh 460. Add back depreciation.
1:35:18And this is just a run rate again. Uh
1:35:20607.
1:35:22There's our EBIT. Annualize it by four.
1:35:25I get to about 17 times enterprise value
1:35:28to EBIDA. Not unrealistic. Now, Coherent
1:35:31is another one of these companies has
1:35:33really come down in price. They were on
1:35:35my short list. They're not on my short
1:35:36list anymore. Uh especially being down
1:35:3950 60% uh from the last time I
1:35:42calculated it. So, they are not on a
1:35:44short list. They're more likely to be on
1:35:46a buy list.
1:35:48Um just to give you an idea of how I get
1:35:50to these numbers, if if we're going to
1:35:52get to GAP, uh we start with non-GAAP uh
1:35:55adback sharebased compensation and the
1:35:58amortization of acquired intangibles.
1:36:00That's all that coherent is doing.
1:36:02That's the difference between GAAP and
1:36:03non-GAAPAP are those two things. I and
1:36:05I've made the case before that I can
1:36:07understand acquire amortization of
1:36:09acquired intangibles, but I I I am firm
1:36:12on sharebased compensation.
1:36:15Let's look at their numbers. Revenue uh
1:36:1850% year-over-year uh almost 16% quarter
1:36:22over quarter. Cost of goods sold uh
1:36:2560.86%
1:36:27last year was 63.5. So, that was a nice
1:36:30pickup of 251 basis points in your cost
1:36:32of goods sold. If we're looking at our
1:36:35incremental gross margin
1:36:39um or or sorry, our incremental cost of
1:36:41goods sold, you would take the
1:36:43difference in revenue divided by the
1:36:46difference in cost of goods sold, you're
1:36:47at 55.82.
1:36:50Uh which means that for every extra
1:36:52dollar of revenue, the cost of goods
1:36:54sold, the incremental cost of goods sold
1:36:57is 5582. So we should expect as revenues
1:37:01increase uh more improvement
1:37:04uh uh in the cost of goods sold. Uh we
1:37:06would expect to see that move down
1:37:08towards the incremental cost of goods
1:37:10sold.
1:37:12Uh getting down to earnings before
1:37:14income taxes, 15.13% of revenue versus
1:37:1713.71. It is an improvement of 142 basis
1:37:21points uh and an increase of 65%
1:37:25year-over-year
1:37:26and 13.61 61% uh quarter over quarter.
1:37:31Quarter over quarter isn't that
1:37:32impressive. 13.61% but revenue was up
1:37:3616% the year-over-year is the bigger the
1:37:38bigger improvement 65% versus 49. That
1:37:42gives you a 1.31
1:37:44total leverage
1:37:46because again you have your uh interest
1:37:48expense uh your interest expense in
1:37:51there.
1:37:53um basic and diluted count diluted 202
1:37:57million up from 190 uh million basic 196
1:38:00up from 156.
1:38:03Here's a breakdown of their uh segments.
1:38:05Data centers uh quarter over quarter up
1:38:0817% uh 85% year-over-year.
1:38:11Communications did better 34% quarter
1:38:14over quarter uh 84% year-over-year. They
1:38:18see these two together as a $90 billion
1:38:20market opportunity uh by 2030 uh based
1:38:24on uh where they're guiding 2.65.
1:38:27They've got a run rate of about 10 10
1:38:29billion. They're seeing uh just data
1:38:33centers and communications being $90
1:38:35billion market opportunity. Industrial
1:38:38flat. Nothing much to talk about there.
1:38:41Um they claim to have visibility into
1:38:432028 uh with their orders. They have
1:38:46just this quarter alone, Q1, they signed
1:38:49multiple long-term agreements that last
1:38:51to the end of the decade. Listen to this
1:38:54quote. We continue to see no attenuation
1:38:57in demand.
1:38:59Uh uh NPO, CPO and chip to chip uh near
1:39:03package co-ackage optics and chip to
1:39:06chip they see as a $30 billion market uh
1:39:09by 2030. they call it a market
1:39:11opportunity uh by 2030 and that is uh
1:39:15within uh within this uh 90 billion. Um
1:39:20they were on my short list. Um
1:39:24but given where they are uh and and how
1:39:26how much they have come down uh
1:39:29multiples of 16.97 and 22.3 and again
1:39:33these are run rates not growth rates
1:39:35right these are not on a valuation basis
1:39:37these wouldn't make it onto a short list
1:39:39to make it onto a short list at those
1:39:41multiples there would have to be
1:39:43something in their product line that you
1:39:44say okay this is this is dead in the
1:39:47water this is this is obsolete for me
1:39:50that is something like circle or any of
1:39:52the of the Neoclouds just the very basic
1:39:55structure of what they do asset heavy
1:39:58with software with claiming claiming to
1:40:00have software margins doesn't exist
1:40:02doesn't exist um where's coherent no I
1:40:06uh I can't with the multiples where they
1:40:09are and what it does and what it's
1:40:10guiding to uh and their margins their
1:40:13margins are okay earnings before income
1:40:15tax uh around uh you know 15% not really
1:40:20great they're Okay, given the multiple
1:40:23that they have at 17 times enterprise
1:40:25value to Ebida, given the growth rates
1:40:27that they're guiding to at 60%, I would
1:40:30say I can't I can't have this on my
1:40:32short list anymore. I am long coherent
1:40:35and I think I think it ended the week
1:40:38for me green as well. I am uh red on
1:40:42quite a few names. Uh I did buy Western
1:40:45Digital. I did buy Seagate. I did buy
1:40:47more uh Nvidia. I bought more bought
1:40:50more Broadcom. Made a big riskreversal
1:40:53bet on Broadcom as well for January. Uh
1:40:57risk reversal means you sell the puts,
1:40:59you buy the calls, and you hope to get
1:41:00them for free, which I did. I paid zero.
1:41:02I got them for free. My risk is that
1:41:06they get assigned to me, which I don't
1:41:07mind on on Broadcom. Um I think I bought
1:41:11more SanDisk. I bought more Micron. Uh I
1:41:14didn't load up on Coherent. I bought
1:41:16some. I would seem to think that as we
1:41:19continue to go down, I would probably
1:41:21say, "Well, I like coherent now. I don't
1:41:25not like it, right? Uh, it's not on my
1:41:28short list anymore." But if it ran right
1:41:30back up, it would be right back there
1:41:31again because its margins are too small
1:41:33to justify anything much over 22 to 25
1:41:38times enterprise value to EBIT.
1:41:41Uh, and even Lum
1:41:43say, wow. I mean, you've corrected by
1:41:46over 50% on your price. I mean, at at
1:41:49what point do at what point do you
1:41:52change your mind and say, "Okay, that's
1:41:53pretty good." You know, I I use an
1:41:56analogy that if I sell my house for 5
1:41:58million in my neighborhood, you might
1:42:00say, "That's a little too much." If the
1:42:03price comes down uh to less than half of
1:42:05that, you'd have to look at it again,
1:42:07say, "Well, I mean, with this new price,
1:42:10maybe it is something I want, right?"
1:42:13Uh, so a lot of things went on sale last
1:42:16week. Before I continue on with the S&P,
1:42:19let's get an assessment of where we are.
1:42:23We have Cororeweave and Nebius reporting
1:42:25next week.
1:42:27What would be interesting is if they
1:42:30could show profitability and I mean not
1:42:32just accounting, not just by moving
1:42:34numbers around on the screen and
1:42:36massaging this assumption and that
1:42:38assumption for useful lives, but really
1:42:40show uh profitability that would mean
1:42:43something.
1:42:44I don't uh I don't anticipate either one
1:42:47of them will. Uh but there's going to be
1:42:49a lot of questions asked. Those are must
1:42:52listen to especially cores. It is the
1:42:54poster child of of management. uh
1:42:56management greed. Um Corweave and Nebius
1:43:00Tuesday night and Thursday night, next
1:43:02Sunday, I will review uh their their
1:43:06results and their and their earnings.
1:43:09I seem to think the selloff we had last
1:43:11week is an overreaction to the
1:43:13regulatory.
1:43:15Um, I believe the governor, the incoming
1:43:18governor of Texas said, "Look, uh,
1:43:20whatever data centers are in operation,
1:43:23whatever permits are out there, they're
1:43:26clean. We're not touching any of that
1:43:28stuff. That's all good to go. The new
1:43:30ones, we're going to have a look at, but
1:43:31did not come right out and say, "That's
1:43:34it. It's over. It's done. Forget it.
1:43:36Pack your bags and go home. We're going
1:43:38to have a look at it." Uh, I think it I
1:43:42think it was was overdone. I think it
1:43:44was just you know on top of everything
1:43:46else uh some valuations had to come down
1:43:50obviously uh and this is coming on the
1:43:53backs of of uh the Neo clouds which my
1:43:56god every day we're learning how many
1:43:58there are like 300 and some odd neo
1:44:00clouds out there that that's just way
1:44:02too many uh and and they're not very
1:44:05effective at what they do and then
1:44:07Friday we had Microsoft saying that well
1:44:11we're going to cancel some uh some of
1:44:14this NeoCloud stuff. So, let me just
1:44:16just give you an idea of of what they
1:44:19did uh because I think I think it uh it
1:44:23it got taken out of context for what
1:44:25they uh what they announced.
1:44:28It was really one company uh Nscale
1:44:32uh and in October they had cancelled a
1:44:35deal worth about $3 billion to be an
1:44:38anchor uh tenant at a site that they
1:44:41were building uh in the UK.
1:44:45Um the grid operator in the UK said it
1:44:50might not be able to power the site
1:44:51until uh uh early to mid 2030s.
1:44:56So yeah, you cancel it. You look at them
1:44:58and you say early to mid 2030s we can
1:45:00build it ourselves. So of course they
1:45:03canled it. Um they also chose same with
1:45:06NScale not to exercise an option and it
1:45:10was just an option for about 700 megawws
1:45:13uh starting in late 2027.
1:45:16Uh it would have doubled the size of the
1:45:18lease they had there. They didn't cancel
1:45:19what they had. They just chose not to
1:45:21exercise another option.
1:45:24um
1:45:27the capacity that they chose not to uh
1:45:30went somewhere else. Figure AI uh took
1:45:32it. So it's not as if Nscale said, "Oh,
1:45:35we're done." Somebody else was in line
1:45:36and took it. Um
1:45:40and in West Virginia, this is still with
1:45:42the same company. They dropped a letter
1:45:44of intent uh to lease 1.35 gawatt.
1:45:49uh and uh there's some reporting about
1:45:53uh not renewing its capacity deal with
1:45:56Oracle.
1:45:57Um and it's
1:46:00they there was some um insider comments
1:46:04u that they're looking at their projects
1:46:06with Nebius and IN
1:46:09um mostly because of delays, cost
1:46:12overruns, permitting problems. You know,
1:46:15if I negotiate with you, let's say I
1:46:18have 10 contractors, you're going to do
1:46:19my painting, someone's doing my
1:46:21drywalling, someone's doing my plumbing,
1:46:23and I'm just getting delay after delay
1:46:24after delay. I'm going to start
1:46:26cancelling some of you guys and looking
1:46:28elsewhere.
1:46:30That seems to make sense to me. So, what
1:46:32has really changed is, well, if the deal
1:46:36was delayed because of permitting
1:46:38reasons, it wasn't going to happen
1:46:39anyway. So, cancelling it, what does
1:46:41that mean? It means you're free to go
1:46:44somewhere else and do it. Doesn't mean
1:46:46that you're you're Microsoft is saying
1:46:48we're done. We're cancelling it. It's
1:46:51over. That capacity is gone. Our demand
1:46:53is gone. It's just look these guys are
1:46:56unreliable. We're going to go somewhere
1:46:58else. So the demand is not leaving the
1:47:01market. It is leaving those people,
1:47:04those players. NScale, perhaps Nebus,
1:47:08perhaps I ran uh Oracle.
1:47:12Um not only that, Microsoft uh has uh uh
1:47:15stated that it does want to reduce the
1:47:18number of suppliers it does have for
1:47:21this uh this thing. It it is not ruling
1:47:24out the use of Neoclouds. It is just
1:47:26saying let's just use fewer of them. Uh
1:47:30and if we can concentrate our purchases
1:47:32with fewer, we make them stronger. And
1:47:35then for the other ones that have no
1:47:37advantage to power over what we have,
1:47:39let's just do it ourselves.
1:47:42Um,
1:47:44just to give you an idea of where
1:47:46Microsoft is on some of these with
1:47:48Nscale,
1:47:49uh, it has, uh, through uh, a
1:47:52disclosure, it has 43.8 billion of
1:47:56commitments to 2033
1:47:58with Nscale with Nebius
1:48:0117.4 4 billion
1:48:04uh and they could increase that by
1:48:05another two billion up to 2031. Iran uh
1:48:109.7 billion uh through 2031. Cororeweave
1:48:15more than 10 billion.
1:48:17Um and it's just looking at at these and
1:48:20saying, "Look, we're not going to put up
1:48:21with these delays anymore. Uh if you
1:48:24can't get us to market sooner than we
1:48:26can get ourselves to market on this,
1:48:28we'll do it ourselves." And that's uh
1:48:31that's all that had happened uh over
1:48:33that period of time. Here's an
1:48:36interesting thing I I did. I asked um
1:48:39Perplexity uh for the supply and demand
1:48:42of tokens. I said get search around and
1:48:45see if you can figure out where the
1:48:47demand for token is and where the supply
1:48:50is and geographically
1:48:52where the supply is. That would be
1:48:54interesting interesting to see.
1:48:57Um, and it didn't really have any one
1:49:00really good source. It used open router
1:49:02traffic data uh to to get most of uh
1:49:05what it has. This is rather shocking. Uh
1:49:08the numbers here. I'm going to just read
1:49:11off some numbers here. Uh Google all
1:49:14products all products 3.2 quadrillion
1:49:18per month. And this was as of May 2026
1:49:21from open router data. 3.2 two
1:49:23quadrillion
1:49:26per month.
1:49:28China, their official public figure
1:49:32uh was 140 trillion per day March 2026,
1:49:37175 trillion per day by June of that
1:49:41year.
1:49:44uh including private cloud
1:49:48China including private cloud about 500
1:49:51trillion
1:49:52per day
1:49:55whereas Google is at 3.2 2 quadrillion
1:49:58per month.
1:50:01Way just Google alone beats all of
1:50:04China. Bite dance about 120 trillion per
1:50:07day. They they announced in April of
1:50:102026. The global estimate
1:50:14uh that it could come up with from open
1:50:16router is about 5 to seven quadrillion
1:50:20per month.
1:50:23Um, and you'll get you'll get that
1:50:25number in that uh paper that I
1:50:27highlighted earlier. How did we get it?
1:50:30Because they're telling you how many
1:50:32tokens they use, which represent about
1:50:342%
1:50:36of uh global uh global token uh uh uh
1:50:40consumption 2%. So you just divide that
1:50:43by 02, you get to about 5 to seven
1:50:46quadrillion per month. That's a huge
1:50:49number.
1:50:51So I really don't think this is over.
1:50:53Some of the quotes we've seen uh no
1:50:57attenuation in demand. Uh we're at the
1:51:00very beginning of of uh of this AMD for
1:51:05agentic AI. We're at the very beginning
1:51:08uh of this thing. Uh the scare we had
1:51:13last week I think is unwarranted. I
1:51:16don't think anything is going to change.
1:51:19What you have going on right now is a
1:51:21massive market share grab. Whoever
1:51:25whoever can supply the tokens will meet
1:51:28the demand for those tokens. And that's
1:51:30where we are right now. Who is going to
1:51:33supply the tokens?
1:51:36The demand is already there. Every
1:51:38company's telling you the demand is
1:51:39there. We can't meet the demand. So,
1:51:41it's a matter of the supply of tokens.
1:51:43No one single company is gonna say, you
1:51:46know, now's the time for us to sit back
1:51:48and let's just wait a little. That's not
1:51:50going to happen. And from a game
1:51:52theoretic perspective, um you your best
1:51:57course of action is to move forward.
1:52:01Uh if everyone else is moving forward,
1:52:03you must move forward. If everyone else
1:52:06is holding back, your best course of
1:52:09action is to move forward no matter
1:52:12what. No matter what conclusion you get
1:52:14when you're in a market share grab, your
1:52:16best course of action is to move forward
1:52:20because I do think the production of
1:52:22tokens in the future will be commodity
1:52:25like and whoever has the biggest share
1:52:28of the market will win. That being said,
1:52:31let's go back to that NBR paper that the
1:52:34higher returns are happening to those
1:52:36that are using these frontier models and
1:52:40not the open-source models. So, I don't
1:52:43think the Frontier Lab, if there's truth
1:52:45in that, if that is true, they should be
1:52:47seeing their returns. And if that is
1:52:50true, they're not going to stop.
1:52:53Okay, the carnage. Look at that big red
1:52:56bar down here. one day uh bounced off
1:52:59the 200 day moving average and uh got
1:53:02some support from there. Uh but we did
1:53:05not have conviction on Wednesday. We did
1:53:09not have conviction on Friday. I talked
1:53:12about this in the uh pre-market lives
1:53:14that I do. And somebody had said, "Well,
1:53:16what do you mean no conviction? We were
1:53:18green on on Wednesday."
1:53:21Conviction is uh let's do a little chart
1:53:23over here. And uh this is 9:30 and this
1:53:27is uh 400 p.m. over here is that uh
1:53:30you're doing this and maybe you do this
1:53:32in the afternoon, but at the end of the
1:53:34day, you're hitting the highs into the
1:53:36end of the day. This is money coming
1:53:38into the market late in the day. Even
1:53:41though you're green, what is not
1:53:43conviction is you're up and then
1:53:45sometime midafter afternoon, money
1:53:47starts leaving the market. And even
1:53:49though you went green, why is money
1:53:51leaving the market in the middle of the
1:53:52day? That money doesn't show conviction.
1:53:56A market that hits the highs into the
1:53:58close, that is conviction. We did not.
1:54:00Even though you had a green day on
1:54:02Wednesday, a green day on Friday, it
1:54:04wasn't conviction. That bothers me.
1:54:08S&P is down 5.72%
1:54:12from its all-time high on October 6th.
1:54:14You have IWM in correction territory.
1:54:17Triple Q firmly in correction territory.
1:54:20Bare market is 20%. I don't really see a
1:54:24reason to get there. But that doesn't
1:54:25mean psychologically the market won't
1:54:27talk itself into it. That's the wild
1:54:30card when it comes to the market, right?
1:54:31Is that is that what will what will the
1:54:34the group talk itself into?
1:54:39Um
1:54:40Thursday was heavy on the regulatory,
1:54:43but there was more stuff going on
1:54:45because the the pace of selling out of
1:54:47the gates was fast. We were down a,000
1:54:49points in 21 minutes.
1:54:52Uh by 10:30 I had done some nibbling in
1:54:54the morning. By 10:30 I thought I'm
1:54:56that's it. I'm done for the day because
1:54:58this is this is just going to keep
1:54:59getting ugly and uh when I got back I I
1:55:03left. Got back later and they looked
1:55:04down 1600 points like oh ouch. Now I
1:55:09have a long short which is nice. I have
1:55:12a long short long beta short uh is where
1:55:15I get my alpha and the short the alpha
1:55:17side has done fairly well over that
1:55:20period of time but I've exhausted my
1:55:21alpha.
1:55:23I mean how much how much more do I want
1:55:25to be short? Some of the names that I
1:55:27was short that benefited me on the way
1:55:28down. I look at I think well I would
1:55:30want to be long now. So this is the
1:55:33problem when you have this kind of
1:55:34selloff is you still have a beta
1:55:36portfolio but you're searching for your
1:55:38alpha. It's like well where's my alpha
1:55:39now? Um, and if I'm the one and if I'm
1:55:43still looking for it, the professionals
1:55:45must be doing the same thing, saying,
1:55:47"Look, uh, it's hard to short these
1:55:49names at this point given given
1:55:51everything we've heard on the calls,
1:55:54given the economic background, the the
1:55:57or the backdrop, GDP growth in the third
1:56:00quarter over 3%. Productivity positive,
1:56:03unit labor costs negative, job growth,
1:56:05revisions for the last couple months,
1:56:08ISMs in expansion territory and consumer
1:56:11sentiment increasing into the end of the
1:56:13week. Markets love a divided government
1:56:16and that's what we have. It's hard at
1:56:19this point to say sell, sell, sell.
1:56:22especially when nothing is falling apart
1:56:24other than just some some belief about
1:56:27where the return is on AI and it is
1:56:30showing up anecdotally here and there. I
1:56:35think it's still a little too early. I
1:56:37think the agentic AI has a lot of
1:56:39promise to show up but
1:56:42fundamentally
1:56:43I would be long this market and say we
1:56:46should we should be moving up into the
1:56:48end of the year. behaviorally beats the
1:56:51hell out of me. I couldn't explain
1:56:54Thursday. I thought this is just stupid.
1:56:56This is just stupid. But here we are.
1:56:59Let's see. Uh let's see what our uh
1:57:02numbers look like here. Forward four
1:57:04quarter operating earnings 405. Closing
1:57:07S&P at 7376 puts that as 18 low 18 times
1:57:12uh earnings. Beginning of the year we
1:57:13were at 23
1:57:1523x. We're down a multiple of five. Last
1:57:18week we're at 18.69. 69 435 companies
1:57:21have reported 5.4%
1:57:24above the expectation revenue at 2.5. If
1:57:28I adjust the forward four quarter
1:57:30operating earnings for the surprise
1:57:32factor, we have an adjusted PE of 17.25
1:57:36with a lot of volatility as well.
1:57:40We do have some earnings this week. Um
1:57:43in light blue is uh S&P earnings.
1:57:46Everything else is non S&P. Uh Monday,
1:57:49uh Oxy, uh Constellation, AS Rocket
1:57:53Labs, and US.
1:57:56Um Tuesday, Liquidia. You have two
1:58:00Quantum in here, D-Wave and Regetti. And
1:58:03Cororeweave.
1:58:04Uh I'd like to say no one really cares
1:58:07about this, but this this has become the
1:58:09poster child of the of the Neoclouds.
1:58:13Um,
1:58:14when it broke, when Coreweave broke, oh,
1:58:18their yields hit 15, 16%, it broke. I
1:58:22kept watching insider sales. And sure
1:58:24enough, they came, they came, they came,
1:58:26they came. Four more insider sales on
1:58:29the way down. Insiders selling. Four
1:58:31different insiders selling on the way
1:58:33down. Not a good look, Corweave. Not a
1:58:35good look.
1:58:37If you really believed in it after that
1:58:40break, you should have been buyers.
1:58:42even even just symbolically be buyers
1:58:45but sellers still. The last one Thursday
1:58:49that big ugly red day insider sale
1:58:53uh for about a million dollars. Um just
1:58:55just rats running from a sinking ship.
1:58:59That's what it feels like.
1:59:01I would not based on what I've seen and
1:59:04I don't know the management. I don't
1:59:06know these people. Uh, but I would not
1:59:10uh uh uh rule out some funny accounting
1:59:14to make to make to make numbers or to
1:59:16try to to to show the market that things
1:59:18are okay. I wouldn't rule it out. So,
1:59:20I'm going to be pouring over those
1:59:22statements for anything that makes me
1:59:23go, hm, this is odd. Uh, and round
1:59:27numbers, especially when you're doing
1:59:28ratios and percentages, round numbers,
1:59:33that that makes me wonder. So, uh,
1:59:36Cororeweave, I don't expect it to show.
1:59:39I I think its revenues will increase. I
1:59:41think it'll it'll it'll be in the game
1:59:44in that, uh, particular respect, but I
1:59:46think it'll become clear that there is
1:59:48no pathway to profitability for this
1:59:51company given the level of debt that it
1:59:53has. And there's probably no pathway uh
1:59:56to raising capital other than selling
1:59:58shares. Uh so I think I think coreweave
2:00:02is probably
2:00:04uh probably not going to help the market
2:00:06at all but this is a neocloud issue.
2:00:10Thursday we get Nebius. Nebus is better
2:00:13run but still unprofitable.
2:00:16Uh we also get applied materials Cisco
2:00:18and Cerebrus a chipmaker in here which
2:00:21has a an interesting approach. They were
2:00:24on my short list for a while but they've
2:00:26fallen by so much I don't know where I
2:00:28sit with these guys. So, I'm going to
2:00:29have to see their earnings, but I will
2:00:32do Cororeeavee
2:00:34uh and I will do Nebius. I don't know
2:00:36that it's worth waiting till next Sunday
2:00:38for Coral Weave because it will matter a
2:00:40lot. So, maybe I'll put out uh one video
2:00:45uh just in reviewing what Corore uh what
2:00:48Cororee has to say. But I don't I don't
2:00:51know that it can save the market unless
2:00:54it can pull a rabbit out of the hat.
2:00:56Here's what I think has to happen with
2:00:57Corore. Number one, a moratorum on
2:01:00insider selling. It's not a good look,
2:01:03guys. A moratorum on insider selling. A
2:01:06moratorum on sharebased compensation.
2:01:09Cuz what are we paying you for to lose?
2:01:11Cuz we're losing. So what's this
2:01:14sharebased compensation for? Enough of
2:01:16it. Uh get rid of that. Um
2:01:21take some of the money that you've that
2:01:23you've earned that that that management
2:01:24has pulled out and buy the debt. Show,
2:01:28show, don't buy the stock, buy the debt.
2:01:30Be the provider of capital if you really
2:01:32believe in it. I don't think we're going
2:01:35to get any announcement from Cororee
2:01:36about Nvidia, although that would be
2:01:38helpful if Nvidia is going to step in
2:01:40and say, "Look, we'll we'll we'll
2:01:42guarantee the debt." I don't think
2:01:44they're going to they're going to do
2:01:45something like that.
2:01:47Um, you'd have to see true accounting
2:01:50operating leverage, not management just
2:01:53telling you they have operating leverage
2:01:55when the numbers clearly show they
2:01:57don't. Last quarter, they talked about
2:02:00how wonderful their operating leverage
2:02:02is, but when you calculated it, it was
2:02:03negative. It It's This is not operating
2:02:06leverage, guys. Now, you have a finance
2:02:08background. I expect that you would know
2:02:10that. So, either they're adjusting some
2:02:13numbers very aggressively to make the
2:02:16point
2:02:18or they're just lying to you. For me to
2:02:20say I don't think they understand
2:02:21operating leverage, that can't be true.
2:02:23They're either just straight out lying,
2:02:26uh, or they're massaging numbers so
2:02:28greatly to make it true, which is a form
2:02:31of lying.
2:02:33Um,
2:02:35yeah. So overall, I you know, even even
2:02:37even saying what I think they need to
2:02:39do, I don't think they're going to do
2:02:40it, and I don't think they will be
2:02:42profitable. And I don't think they can
2:02:43show a pathway to profitability.
2:02:46I don't know that we can uh hope for
2:02:48much on this, but I will I will review
2:02:51it the moment it comes out.
2:02:55Well, maybe not the moment it comes out,
2:02:57but uh because I need the uh I need the
2:03:00replay of the conference call because I
2:03:03do have to go back and forth and back
2:03:04and forth. And on the live, you don't
2:03:06get to do that. You miss something, you
2:03:07got to wait for the recording to be up.
2:03:09So, I'll try to get that done.
2:03:12Overall, I am uh still long. Uh NASDAQ
2:03:1721 contracts of NQ.
2:03:20Uh I'm going to remain long. I'm still
2:03:22hopeful for an endofear rally. The
2:03:25fundamentals tell me that everything is
2:03:27just fine. The economic fundamentals
2:03:31and the uh uh uh uh accounting uh
2:03:34fundamentals from the companies suggest
2:03:37that this is still this is this is going
2:03:39this is going it's not going to stop.
2:03:43It is the sentiment right now on uh if
2:03:47we look at uh poly not poly market. If
2:03:50we look at uh
2:03:52hyperl
2:03:54uh we have as of now we have S&P set uh
2:03:59to open some 25 points lower than the
2:04:02close
2:04:04on uh on Friday.
2:04:08We have oil set to open higher.
2:04:11Um,
2:04:13it said this last week and we opened
2:04:16very differently than what Hyperliquid
2:04:18said. So, let's see what happens.
2:04:21That is it for this week.

About this transcript

This page contains the full transcript of AI over-reaction and earnings for LITE, FN, AMD, QCOM, and COHR by Mark Meldrum, generated from the public captions YouTube serves with the video. The transcript has 18,845 words across 2,768 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.