This AI Stock Crash Is a GIFT — Here’s What I’m Buying — Transcript
Full transcript
- 0:00the semiconductor sell-off. It suddenly
- 0:02created some of the biggest valuation
- 0:04gaps that I've seen all year, but it's
- 0:07also created some very obvious traps.
- 0:09Broadcom that fell around 5% on Friday.
- 0:12AMD dropped almost 9% and Nvidia, Micro,
- 0:16Marvel, and the equipment names, they
- 0:18were all hit together. But if we zoom
- 0:21out to the full week, the picture is
- 0:23completely different. Most of the
- 0:25sectors are green, while Nvidia, that's
- 0:27down almost 13%. I mean, in less than
- 0:30one week, Nvidia alone has lost roughly
- 0:33660 billion in market value. That's not
- 0:37a normal move in a normal market. And at
- 0:40the same time, we've got one AI stock in
- 0:42our ranking today. It appears to be
- 0:44trading at almost half of my calculated
- 0:47intrinsic price with a margin of safety
- 0:49close to 50%. Yet another worldclass
- 0:53semiconductor company, it still looks
- 0:55about 16% overvalued even after the
- 0:58recent weakness. Remember, falling
- 1:00prices, they don't automatically mean
- 1:02cheap shares. So today, I've rebuilt the
- 1:05valuations using the same 9% discount
- 1:08rate throughout. And I've ranked 13
- 1:10semiconductor stocks from Dtier all the
- 1:13way to S tier. We've also got two
- 1:16extremely sickle names where the
- 1:18spreadsheet prints enormous upside. I'm
- 1:20going to show you why those numbers
- 1:22cannot simply be accepted at face value.
- 1:24And then we've got the number one stock
- 1:26is not the one with the biggest headline
- 1:28upside. It's the one where the downside
- 1:30case, the base case, and the business
- 1:32quality, they fit together best. So
- 1:35today's ranking is about where I put
- 1:37fresh money today. A brilliant company,
- 1:40it can still rank poorly when the price
- 1:42already assumes brilliant execution. And
- 1:45the wider market is important today.
- 1:47information technology. It traded around
- 1:5018 times Ford earnings below both the
- 1:525-year average of 25 and the 10year of
- 1:5523. Where the S&P 500 forward multiple
- 1:58is also fallen sharply from the recent
- 2:01peak and it now sits below the
- 2:03longerterm average lines. Valuation
- 2:05pressures they've already arrived. And
- 2:07before we get into the individual
- 2:08valuations, this is the bigger question
- 2:10hanging over the entire ranking. how
- 2:13much of Nvidia's value belongs to the
- 2:15company and how much belongs to Jensen
- 2:18himself. We're gonna listen to Bill
- 2:20Aman's thoughts.
- 2:21>> You allay fears of the AI trade and I
- 2:23know you also tweeted about um Jensen.
- 2:26>> Well, I think Jensen's amazing CEO. Sad
- 2:29to see uh him leave, but you know,
- 2:32Nvidia will survive. Uh and you know,
- 2:35I'm sure the company will do great.
- 2:36>> What do you think of Nvidia?
- 2:37>> And I think I think Jensen will do
- 2:38something else great. and Nvidia being
- 2:41the central bank of AI. What do you make
- 2:42of the key manner of risk and how one
- 2:45company is at the heart of all the AI
- 2:47trade on
- 2:48>> there a lot of iconic CEOs and and you
- 2:50know Apple had an iconic CEO and company
- 2:52did pretty well after
- 2:53>> that's the optimistic version the
- 2:55founder matters enormously but the
- 2:57platform can outlive him. I'm going to
- 2:59test that view against the cash flows
- 3:01rather than simply choosing a side. It's
- 3:04also important to note that earnings
- 3:06expectations, they've not collapsed.
- 3:08Bottom up S&P 500 earnings, they're
- 3:10still projected to rise from roughly 275
- 3:14to 365, then 425.
- 3:18Where information technology is also
- 3:20producing some of the strongest earnings
- 3:22growth within the index. That is why I
- 3:25don't want to treat every red
- 3:26semiconductor stock today as being
- 3:28completely broken. I also want to make
- 3:31it clear that for every DCF that we use
- 3:33today, we're using a 9% discount rate,
- 3:353% perpetual growth rate, then stress
- 3:38the operating growth assumptions across
- 3:40a low, medium, and high case. Now, all
- 3:43companies, they're going to start in the
- 3:45D tier and then move them up to the
- 3:46relevant allocation or in fact keep them
- 3:49at D. But make it clear D in fact, it
- 3:51means I would avoid adding fresh money
- 3:53at this price. C tier, it means the risk
- 3:56to reward is still unattractive. B is
- 3:58going to be the watch list or measured
- 4:00buying territory. And A tier means I see
- 4:03an attractive entry with a real margin
- 4:05of safety. S tier today is reserved for
- 4:08the strongest opportunities. But I want
- 4:10to make it clear, some S tier names,
- 4:12they carry serious cycle risk. Most
- 4:15importantly, I'm not ranking the
- 4:17companies by quality alone. I'm ranking
- 4:19the shares by what today's price asks us
- 4:21to believe about the future cash flow is
- 4:24why the reverse DCF is also handy. And
- 4:27if we kick off with Apply Materials,
- 4:29this is a highly profitable business and
- 4:31they remain positive and the forward
- 4:33earnings multiple is sitting just below
- 4:3620. If we look at a trailing non-GAAP
- 4:38earnings while Apply Materials, it tra
- 4:40around 28.7 times. We can see that's
- 4:43around 32% richer than the sector median
- 4:46and also 23% above their own 5-year
- 4:49average. Now their revenues moved from
- 4:51roughly $7 billion in 2005 to an
- 4:54estimated 34 billion in 2026. The
- 4:57longerterm execution for AMAT has been
- 5:00excellent. And then if we look at the
- 5:02estimates they do become much more
- 5:03aggressive rising towards 47 billion 56
- 5:06and then 61. The future growth is
- 5:09exactly what buyers are paying for
- 5:11today. Now my DCF it produces a price of
- 5:14$311
- 5:15compared with today's price. Well,
- 5:17there's actually no margin of safety.
- 5:19you're looking at a premium of 16%. Now
- 5:21the low case in our model it assumes 10%
- 5:24growth the value that falls to $213.
- 5:27Medium case as we said 311 while the
- 5:30higher case that reaches 453. It means
- 5:33the base case it still shows roughly 14%
- 5:36downside and the reverse DCF well that
- 5:39requires around 17% growth merely to
- 5:42justify today's price. Now apply
- 5:44materials I like it as a company. I
- 5:47simply don't like the amount of
- 5:48execution all you required for fresh
- 5:50money today apply materials it's staying
- 5:53right in the D tier we then move on to
- 5:55advanced micro devices the share price
- 5:58has been incredibly volatile and the
- 6:00session we saw on Friday it wiped almost
- 6:029% from the stock trailing non-GAAP
- 6:05earnings well they're valued at more
- 6:07than 80 times we're talking 270%
- 6:10richer than the sector and 63% above
- 6:13their own 5-year average now consensus
- 6:16This EPS we can see here it does rise
- 6:18steeply through the next several years.
- 6:20It gives AMD huge operating leverage if
- 6:23the estimates arrive but it also raises
- 6:26the cost of disappointment. My DCF it
- 6:28arrives at $525
- 6:31against the market value pretty much
- 6:33seeing it at fair price. Therefore
- 6:35there's no margin of safety just around
- 6:371% which is fairly trivial. At 10%
- 6:40growth the value falls to $439
- 6:43around 16% downside. And at the upper
- 6:46trend the high case where we can see 628
- 6:50that indicates 21% upside the reward it
- 6:53does exist but the base case it gives
- 6:55almost no protection today. Now AMD it
- 6:59could absolutely outperform if AI
- 7:01accelerated share in margin surprise
- 7:03upwards. At this exact price though have
- 7:06to be honest the valuation it gives me
- 7:08no margin for error. AMD today it goes
- 7:11in the C tier. We then move on to KLA.
- 7:13We can see it's got an A+ on
- 7:15profitability and a remarkably durable
- 7:17position in process control. But the
- 7:20valuation grade it speaks for itself.
- 7:22It's sitting at an F. I mean the 5-year
- 7:24average forward P it sits at 23.5
- 7:27currently sitting around 25. It means
- 7:30KLA it sits around the normal valuation
- 7:33which is why when we look at the blue
- 7:35tunnel from simply safe dividends which
- 7:36highlights intrinsic fair price is
- 7:38sitting within that but right at the
- 7:40upper end indicating potentially an
- 7:42overvaluation signal but this one last
- 7:45time it showed us severe undervaluation
- 7:47signals. I mean you'd have to go back
- 7:49all the way to around 2019. There were
- 7:52some points where it slightly
- 7:53undervalued. More often than not though
- 7:55this is a premium stock. Now the
- 7:57business is also earned that respect.
- 7:59Revenues compounded at 17% annually over
- 8:02the decade rising from $3 billion to
- 8:05around 14 billion for FY26. And my DCF
- 8:09it comes to $142. Compared to market
- 8:12price, there's actually no margin of
- 8:14safety, a 6% premium. If we look at the
- 8:16lower-end growth, 10%, models around
- 8:19$100, that's $ 33% downside. If we take
- 8:23a look more aggressive case at 20%
- 8:25around $200 33% upside and the reverse
- 8:29DCF it is asking around 15.8% growth.
- 8:32Now that is achievable for KA but it's
- 8:35also close to the company's excellent
- 8:37decadel long revenue compound rate as
- 8:39well as free cash flow rate. KLA for me
- 8:42this is a quality compounder at a
- 8:44quality price not a distress bargain. I
- 8:47would happily revisit it after a deeper
- 8:49pullback. For now, KLA, it remains in
- 8:52the C tier. Now we move on to ASML. The
- 8:55company's competitive position that's
- 8:57exceptional and the stock is held up
- 8:59better than most of the semiconductor
- 9:00group. The forward P it sits at 23
- 9:03significantly lower than the 5 average
- 9:05of 32. That does indicate a potential
- 9:08severe undervaluation signal which is
- 9:11also confirmed when we look at the blue
- 9:12tunnel. There is quite a disconnect
- 9:14between the point today and the lower
- 9:16end of the fair value. again over the
- 9:18last 5 10 years. This one, not all the
- 9:21time, but there were some points in time
- 9:23where you could get this in a severely
- 9:25undervalued level. Most recent 2025 and
- 9:28their service and field option revenue
- 9:30that reach around 3.3 billion. It shows
- 9:33the value of the installed base. This
- 9:35recurring element, it makes ASML more
- 9:37durable than just a pure equipment cycle
- 9:40story. Now, my headline DCF, it comes
- 9:42around $1,540.
- 9:45There is a margin of safety, although
- 9:47fairly trivial around 4%. If we look at
- 9:50the lower case at 10% growth, it falls
- 9:52down to around $1,100, indicating 25%
- 9:55downside. If you want to go more
- 9:57aggressive at 20, we're talking 2,141.
- 10:01That is, in fact, 45% implied upside
- 10:04from today's price. So, ASML is close to
- 10:07my fair value in my base case, while the
- 10:09downside case still matters. Monopoly
- 10:12like quality, it reduces the risk. It
- 10:14doesn't remove valuation risk. Now,
- 10:17ASML, I would in fact be comfortable
- 10:19accumulating gradually, but I don't see
- 10:21enough of a margin of safety for an A
- 10:24tier. ASML is the first one going in the
- 10:27B tier group. We then move on to
- 10:28Qualcomm. This one is pretty much flat
- 10:30year to date over the last 5 years.
- 10:33Pretty much the same. So, it hasn't had
- 10:35the greatest short to medium-term in
- 10:37terms of where it sits around the lower
- 10:38end of the 52- week range where we only
- 10:41get one buy rating. very weak at that
- 10:43from seeking Alpha. Now interesting as
- 10:45well the yield while it is in line with
- 10:47their 5year average the forward
- 10:49valuation it does sit higher than
- 10:51historical 18 against 14 for which when
- 10:54we look at the blue tunnel we notice
- 10:55here a signal of potential overvaluation
- 10:59this one in fact qualcom you can see the
- 11:01underlying fundamentals they have been
- 11:03very volatile likewise so is the price
- 11:05now we're seeing it overvalued the
- 11:07opposite of what we saw in 2022 and 2023
- 11:10also one thing to point out when we do
- 11:12look at the DCF in quite some detail.
- 11:15Well, historical free cash flow has been
- 11:17uneven, but also substantial, rising
- 11:20from 4.8 billion in 2015 to around 12.8
- 11:24billion in 2025. Now, we've only used 2%
- 11:27growth because when we look at the
- 11:29fundamentals, both historical and
- 11:31expected moving forwards, they're not as
- 11:33fast growing as some of the other
- 11:34companies that we're discussing today.
- 11:36And with that value, we get $187
- 11:39implied 10% upside, margin of safety
- 11:42coming around 9%. Now, if you wanted to
- 11:44look at the lower case today, that hits
- 11:46$137,
- 11:48downside of 19%, more aggressive at 7
- 11:51coming to 64% upside. But some people
- 11:53may look at this and say this is a lot
- 11:55more conservative than what we've used
- 11:58for other companies. Which is why I'd
- 12:00say today for Qualcomm, the base case,
- 12:02while offering around 10% upside, the
- 12:05reverse DCF very important, is only
- 12:08highlighting here less than 1% growth
- 12:10expectations for Qualcomm. They already
- 12:12look fairly modest. So overall
- 12:15attractive. Not enough for an A tier
- 12:17because handset exposure, licensing
- 12:19risk, and the need to prove meaningful
- 12:20diversification that still caps my
- 12:23conviction. For Qualcomm, this goes into
- 12:25the B tier. Is a reasonable value entry
- 12:28rather than a screaming bargain. I'd
- 12:30consider a measured buy rather than
- 12:33going all aggressive. Now we reach
- 12:35Nvidia. The stock will probably create
- 12:37the most disagreement. The shares well
- 12:40they fallen hard yet analyst ratings and
- 12:42profitability they still remain
- 12:44exceptionally strong and as we
- 12:46highlighted earlier the markets remove
- 12:48roughly 660 billion in just under a
- 12:52week. The move it creates opportunity
- 12:54but the size of the fall is not a
- 12:56valuation by itself. First though the
- 12:58bull case Jim Leventhal explains why the
- 13:01collapse in the share price is not
- 13:03changed what he's personally doing with
- 13:05fresh money. The bulls like myself are
- 13:08going to point to the strong order book,
- 13:10the growth that's projected. We'll get
- 13:12an update on that with next week's
- 13:14earnings. I'm not expecting a major
- 13:15change. Um, we're going to talk about
- 13:17things like free cash flow over 300
- 13:20billion, the forward multiple of 10. Uh,
- 13:22and say, "Wait a second, how can we be
- 13:24selling this here?" And I'm not. I'm not
- 13:26selling it here. And new money that
- 13:27comes into account, I am putting into
- 13:29Nvidia.
- 13:30>> That is the cleanest bullish argument.
- 13:32The order book and cash generation may
- 13:34be stronger than the fear. But there's
- 13:36also another side to this trade. The
- 13:38estimates they support the bullc case.
- 13:40Consensus EPS it keeps rising to 2031.
- 13:44Although the investment still depends on
- 13:46sustained demand and exceptional
- 13:48execution. Stephanie Link takes the
- 13:50opposite view. Her question is whether
- 13:52investors can earn a better return
- 13:54elsewhere while the leadership overhang
- 13:57remains. But this stock is not going to
- 13:59trade on fundamentals, I don't think,
- 14:00for a while because you have the
- 14:02overhang of the investigations, of
- 14:04leadership changes, of potential
- 14:06departures. And so I think it's really
- 14:08for me as as an investor, I'm thinking
- 14:11the opportunity cost here because I do
- 14:14think it's going to be kind of in a
- 14:15trading range for the next 6 to 12
- 14:17months in my opinion. Um, I will not
- 14:20have a problem buying the stock back if
- 14:21we get a a really uh amazing leader and
- 14:24we can get past this, but I think we
- 14:26have time.
- 14:27>> An excellent company. It can still
- 14:29become dead money. That opportunity
- 14:31cost, it keeps Nvidia out of my top two
- 14:33tiers. And this final clip, it puts a
- 14:36number on the keyman risk. And even it's
- 14:38without Jensen estimate, it remains
- 14:40above my model price.
- 14:42>> The valuation without Jensen 225 per
- 14:45share. With him, it's 300. So, it's a
- 14:48different story, but that's why I just
- 14:50continue to believe like they will the 3
- 14:53to four trillion of spend. They're going
- 14:54to get a significant piece.
- 14:56>> But the tech is the tech. These chips
- 14:57are the best chips in the world.
- 15:00>> My own DCF, well, it's more conservative
- 15:02than both of these figures. And Wall
- 15:04Street, well, they've continued to hit
- 15:05the company with downgrades. While most
- 15:07of them still reaffirming as a buy, we
- 15:10can see the average target now was above
- 15:12$300. It's come down to around $280, but
- 15:15still implying significant upside around
- 15:1872%. It's useful context, but it's not
- 15:21what my valuation is telling me. Now, my
- 15:24model, and I will say as they're
- 15:25reporting their earnings in the coming
- 15:27days, it will change depending on how
- 15:29they perform. So, before that, we have
- 15:31$179
- 15:33against today's price at $162. That's a
- 15:369 to 10% margin of safety. If we use the
- 15:39lower growth rate at 10%, we get $145
- 15:43indicating downside of 10. If we use the
- 15:45more aggressive 20%, 220 indicating 36%
- 15:49upside with a reverse DCF. We can see
- 15:52here it requires 12.7% growth for
- 15:55Nvidia. That is credible. But the base
- 15:57case is only offering us around 10% MOS.
- 16:00So Nvidia is not a selling the framework
- 16:03and I'm not betting against the
- 16:04business. it simply doesn't yet offer
- 16:06enough valuation protection to go
- 16:08straight into the A tier. So for fresh
- 16:11money, I would buy it in measured
- 16:13amounts rather than chase it
- 16:14aggressively. Nvidia, it goes into the
- 16:16upper half of a B tier. We then move on
- 16:19to Marvel, which on the surface, it
- 16:21looks far cheaper than Nvidia with
- 16:23strong growth and a share price well
- 16:25below our fair value. We've also got
- 16:28Consensus EPS rising sharply all the way
- 16:30to 2031. Now after that point we can in
- 16:33fact see analyst count it does collapse
- 16:35to one. So I'm not going to treat any of
- 16:37these distant estimates as reliable
- 16:39evidence where they're currently trading
- 16:41at a forward P of 28 times slightly
- 16:44below the 5year that sits around 30. It
- 16:46supports the idea for Marvel that the
- 16:48recent selloff is compressed
- 16:50expectations rather than destroyed the
- 16:52story for which when we look on the blue
- 16:54tunnel we do get a slight undervaluation
- 16:56signal sitting right there towards the
- 16:58bottom end. I mean, this one we can see
- 17:00was trading at such a massive premium
- 17:02earlier in the year, and we do tend from
- 17:05time to time release videos considering
- 17:07stocks to trim or sell. Marvel, well,
- 17:09it's always been one when we see such a
- 17:11massive disconnect. And we've also got
- 17:14their revenue chart. It combines Marvel
- 17:16with momentum and shows enormous
- 17:18estimated growth all the way through
- 17:202029 is also the exact reason the
- 17:22current standalone numbers that do
- 17:24require a bit of caution. Now, the
- 17:26headline DCF, it gives us a valuation of
- 17:29$210 compared to today's price of $152.
- 17:33That's a 28% margin of safety. If we
- 17:36look at the cases, well, on the low end
- 17:38at 15% growth, $150, pretty much saying
- 17:42fair value, more aggressive 293, nearly
- 17:44a double with 93% upside. But bear in
- 17:47mind, this model uses $877 million
- 17:51shares. the planned all stock momentum
- 17:53transaction. It changes the future
- 17:56earnings base cash flows and the share
- 17:58count materially. The deal it could also
- 18:00create substantial data center scale and
- 18:02synergy value but it also means the old
- 18:04DCF it can't be treated as a clean 38%
- 18:07upside signal. Now I like the strategic
- 18:10logic and Marvel could eventually move
- 18:12higher into the ranking until we have a
- 18:14proper proformer model. This one it
- 18:16enters into the B tier is exactly why I
- 18:19don't rank stocks by the largest number
- 18:21in just one spreadsheet cell. Model
- 18:23quality matters just as much as model
- 18:25output. We then move on to Taiwan
- 18:28Semiconductor, the manufacturing
- 18:29backbone behind much of the AI ecosystem
- 18:32and the operating momentum where it
- 18:34remains exceptional. their recently
- 18:36released October detail. Well, revenue,
- 18:39it reached 587 billion Taiwan dollars,
- 18:43up 60% year-over-year. January through
- 18:45October revenue that was up 46%. And
- 18:49consensus EPS, well, it continues
- 18:51climbing, supported by leading edge
- 18:53demand. The forward P, it also sits
- 18:55below the 5year average 18 versus 21, an
- 18:59undervaluation signal for which also
- 19:01confirmed when we look at the blue
- 19:02tunnel. Again, most of these companies
- 19:04that we're talking about very rarely
- 19:06stay for long in the severely
- 19:08undervalued level. We saw it once in
- 19:102025 and then the end of 2022. If you
- 19:13want to go back even further, well,
- 19:15you'd have to look at data pre206.
- 19:18Now, my DCF, it comes to $492.
- 19:21That indicates a 20% margin of safety,
- 19:24implying upside from today's price of
- 19:26around 25%. Low growth case, that one at
- 19:3015% does imply downside. We're talking
- 19:32around 9% more aggressive 675 71%
- 19:37upside. This is what I like when we look
- 19:39at TSM. The low case of 15% whilst it is
- 19:42lower in fact than today's price is not
- 19:45materially away is the downside
- 19:47protection that I want from a company
- 19:48that I'm considering putting into the A
- 19:50tier. Now the obvious risk is
- 19:52geopolitical and no valuation model can
- 19:54reduce that to a neat percentage.
- 19:56Position size matters more here than it
- 19:58does for a normal industrial company. on
- 20:01business quality, growth, and valuation.
- 20:04While TSM it earns a tier, I would buy
- 20:07it, but I would still cap the position
- 20:09because the tail risk is real. We then
- 20:11move on to Lamb Research. The market,
- 20:13well, it raised profitability at an A+
- 20:15while the share price is pulled back
- 20:17enough to create a meaningful valuation
- 20:19gap. All-time highs not long ago sitting
- 20:22around $440.
- 20:24It's now sitting near half of that
- 20:26point. We've also got consensus EPS
- 20:28rising from $965 in 2027 to more than
- 20:32$19 by 2031. Now look, the first three
- 20:35forecasts always important to see. They
- 20:37do carry more weight than the final two
- 20:39where their forward valuation is only
- 20:41marginally lower than history. 22 versus
- 20:4423 is not distressed but it is
- 20:46reasonable for this quality of a company
- 20:48for which when we look at the blue
- 20:50tunnel we can see it's heading towards
- 20:52the lower end of that reasonable entry
- 20:54in terms of the last 5 years or this one
- 20:56we can see was trading at a premium from
- 20:58mid 2025 very rarely again important to
- 21:01highlight this isn't a company that
- 21:03typically trades undervalued and share
- 21:06repurchase they've reduced diluted
- 21:08shares by roughly 14% since fiscal 21
- 21:11that's supported by $17 billion of
- 21:14buybacks over the years. And obviously
- 21:16capital return, it matters here because
- 21:18each remaining share owns a larger
- 21:20portion of the company. I still
- 21:21obviously want repurchase made at
- 21:23sensible prices, not simply at just any
- 21:26price. Now my DCF for LRC, it comes to
- 21:29$317 indicating a 30% margin of safety.
- 21:33This implies upside of around 42%. The
- 21:36low case 10% pretty much coming to fair
- 21:38value today's price more aggressive 20%
- 21:42pretty much a double at $441.
- 21:45This is an excellent setup very limited
- 21:48downside in the low case 42% upside in
- 21:50the base case and reverse DCF requiring
- 21:53only 9.7% growth. I mean look at the
- 21:565year keer much higher at 18. Even the
- 21:5810-year sitting a few percentage points
- 22:00above at 11%. Lamb does however remain
- 22:04cyclical and the latest annual free cash
- 22:06flow it did dip 10%. For the current
- 22:08valuation it doesn't demand a perfect
- 22:10cycle for fresh money. Lamb today is one
- 22:13of my preferred equipment names. It
- 22:15earns a strong A tier and it sits very
- 22:17close to the S tier cut. Then we move to
- 22:20Micron where the opportunity cost
- 22:22argument it becomes actionable.
- 22:24Stephanie Link sold Nvidia and moved the
- 22:26money into a cleaner cheaper
- 22:28semiconductor story. the overhang from
- 22:30this whole saga is just going to
- 22:32dominate the headlines for the next many
- 22:35months. And so I sold it. I got a lot of
- 22:38push back from it. I know I hate to do
- 22:40it, but I was down a little bit and I
- 22:42thought, you know what, this is dead
- 22:44money because you have, you know,
- 22:46leadership changes and questions. You
- 22:47have distractions. You may have people
- 22:49that actually leave. So there's so many
- 22:51different things to consider. And so I
- 22:54took the money and I put more into
- 22:55Micron because by the way they're both
- 22:57down about the same amount from their
- 22:59highs in about the 30 35%. Yet Micron is
- 23:04just a cleaner story. It has good
- 23:06fundamentals and yet they don't have
- 23:08this overhang. And oh by the way it
- 23:09trades at four times earnings and
- 23:11they've got a lot of free cash.
- 23:12>> Four times earnings. It sounds
- 23:14irresistible. The cash is deciding how
- 23:17much of Micron's cash flow survives the
- 23:19next memory down cycle. Now micron we
- 23:22can see it has strong growth and
- 23:24profitability grades and this is an
- 23:26opportunity but remember it has an
- 23:28unusually high amount of cycle risk. The
- 23:31forward P while it's sitting 4.5 times
- 23:34compared with the 5-year average at 11
- 23:36on earnings. The market is priced in a
- 23:38major normalization for which we get a
- 23:41severe undervaluation signal like a few
- 23:44companies we've seen in recent days.
- 23:46It's closer to zero than it is to the
- 23:48fair value over the last 5 years. Well,
- 23:51this one pretty much from the beginning
- 23:53of the year. It's only expanded
- 23:55incredibly fast. Share price is yet to
- 23:57reach up with the fundamentals today.
- 23:59We've also got quarterly net income that
- 24:02surged to levels far above the previous
- 24:04cycle. The strength is real, but
- 24:06extrapolating a memory peak indefinitely
- 24:09is one of the easiest valuation mistakes
- 24:11to make. Now, consensus, they've got EPS
- 24:14rising dramatically through 2029 before
- 24:16the growth rate starts to slow. The
- 24:18near-term forecast, they've got broad
- 24:20analyst coverage. We can see the distant
- 24:23figures, well, they're far less
- 24:24dependable. And the DCF is coming to
- 24:27around $1,300. Compared with the market
- 24:29price, we're sitting at a 41% margin of
- 24:32safety, implying upside of 71%. And we
- 24:36have used the bare case rather than the
- 24:38base case today to be more conservative.
- 24:40Now, the DCF, it does begin with 75
- 24:43billion on analyst expectations. In
- 24:45fact, we've been more conservative. They
- 24:47did have a higher figure, but bear in
- 24:49mind, we've continued with 5% over the
- 24:51longer term. The starting cash flow
- 24:53number here is a very critical variable.
- 24:56Now, if you want to use 10% and you
- 24:58believe the starting base will only
- 25:00continue to increase, well, we have
- 25:01$1,800 indicating 137% upside. Wall
- 25:05Street themselves, they see it around
- 25:07$15,600 in the short term. That's more
- 25:10than double 105% upside. The risk for
- 25:13Micron straightforward. If memory
- 25:15pricing, high bandwidth memory demand or
- 25:17capital discipline deteriorates, peak
- 25:19earnings and cash flow, they can fall
- 25:21far faster than a smooth model assumes.
- 25:24The opportunity though equally clear.
- 25:26The forward multiple already reflects a
- 25:28sharp downturn while AI memory demand
- 25:30could keep the cycle stronger for longer
- 25:32than investors expect. I would
- 25:34personally not size Micron like a
- 25:36defensive compounder and I would not
- 25:38anchor to the most optimistic output
- 25:40here, but the risk ward is too
- 25:42compelling for a tier. Micron today it
- 25:45enters into the S tier, but remember
- 25:47this is cycle sensitive. It means
- 25:49buyable at this valuation but only with
- 25:52a position size that can survive a
- 25:54brutal memory downturn. We then move on
- 25:56to SK highix which is down around 14% in
- 26:00just the last 30 days trading around 52-
- 26:02week lows where alltime highs sitting
- 26:04around $200. Wall Street will they give
- 26:06this a strong buy for which on a forward
- 26:09gap basis it's trading around five very
- 26:11similar in the discussions that we just
- 26:13had about Micron. Now we can see in 2022
- 26:16they reported 3.3 billion of negative
- 26:19free cash flow. It's now sitting around
- 26:2218.2 billion in 2025. That was a massive
- 26:25jump. Analysts they're expecting even
- 26:27more in 2026. Like Micron we've been
- 26:30conservative. They see more than 80
- 26:32billion. That's the starting point we've
- 26:34used for 26. The single input here, it
- 26:36drives the majority of the result. So
- 26:38bear in mind, it's got to be treated as
- 26:40a scenario, not a guaranteed run rate.
- 26:42Using that again, light micron, we use
- 26:45the lower rate of 5%. Intrinsic value
- 26:47comes to $221
- 26:49indicating 70% upside, margin of safety
- 26:52of 41%. If you believe that that's too
- 26:55conservative, though, they'll just
- 26:57continue growing. While the midpoint at
- 26:5810% comes to 306
- 27:02135% upside. Now all of these figures
- 27:05they do look spectacular but the model's
- 27:07low case is not truly low if the $80
- 27:09billion starting cash flow itself
- 27:12actually represents a peak cycle
- 27:14outcome. SK Highix well for me today it
- 27:16enters S tier purely an opportunity but
- 27:19it carries the highest model risk in the
- 27:22entire ranking. I would size it below
- 27:24the more dependable S tier names. We
- 27:26then move on to SanDisk, a stock that's
- 27:28fallen from $2,354.
- 27:32Yet the operating momentum is
- 27:34extraordinary. Double strong buy from
- 27:36Wall Street and Quan. And you can see
- 27:38the factor ratings, A+ on valuation, A+
- 27:41on growth, A+ on profitability, where
- 27:44the forward P is sitting at 3.9, an
- 27:47unusually low number for this earnings
- 27:49profile. And we've got the detailed
- 27:51valuation table confirming the discount.
- 27:53Forward non-GAAP earnings valued at 3.89
- 27:57is roughly 8081%
- 27:59below the sector median. And we've got
- 28:01data center revenue that's exploded from
- 28:04$40 million in late 23 to 4.8 billion by
- 28:08October 26. This is the engine behind
- 28:10the earnings surge. And the consensus
- 28:13EPS line, it pretty much peaks in 2029
- 28:16and then it starts to drop sharply. It
- 28:18tells us plainly that analysts expect
- 28:20the current profit cycle to eventually
- 28:22normalize. Now my DCF comes to $1,700
- 28:26indicating a 50% margin of safety and
- 28:29that's upside of 98% pretty much a
- 28:31double. Now again bear in mind we have
- 28:33used $18 billion again lower than what
- 28:36analysts are predicting whilst also
- 28:37using 0%. So that's flat yet it may be
- 28:40extremely conservative but the important
- 28:42question is is the 18 billion actually
- 28:44sustainable. Now, if data center storage
- 28:47demand stays elevated, the market is far
- 28:49too pessimistic, if cash flow falls back
- 28:52sharply, the headline margin of safety
- 28:54that can shrink very quickly. Now, I'd
- 28:57never call SanDisk a low-risk S tier
- 28:59stock. I would call it a deeply
- 29:01discounted high volatility opportunity
- 29:04where position sizing matters just as
- 29:06much as stock selection. So, SanDisk it
- 29:09goes, I would say, in second spot in the
- 29:11entire ranking because the valuation is
- 29:13exceptionally cheap. But I still prefer
- 29:15the risk adjusted setup of the final
- 29:18company. Now before I discuss number
- 29:20one, listen to how the competitive map
- 29:22is changing. This is not only an Nvidia
- 29:24versus AMD contest. Broadcom is helping
- 29:27the hyperscalers build an alternative
- 29:29route.
- 29:30>> I I wouldn't quite say we're at that
- 29:31point yet. What I would say is
- 29:33competition has been mounting for quite
- 29:36quite some time now. AMD is uh has
- 29:38formidable products now. They're very
- 29:40competitive with what Nvidia has to
- 29:41offer. Broadcom who helps the
- 29:43hyperscalers build their own chips is is
- 29:46very very capable and I think you know
- 29:49they were already
- 29:51gaining share they've already got more
- 29:53than a tow hold they've cracked it open
- 29:56and I think the latest troubles with
- 29:57Nvidia the uncertainty around their
- 29:59their future leadership uh accelerates
- 30:03that process
- 30:04>> that is the strategic reason brocom
- 30:06finishes above every stock here it can
- 30:08benefit from AI spending even when
- 30:11customers want to reduce their
- 30:12dependence on a single GPU supplier.
- 30:14Now, my number one semiconductor stock
- 30:17from the entire group is Brocom. It
- 30:19doesn't have the largest upside, but
- 30:20it's got the strongest balance between
- 30:22quality and valuation. Now, the forward
- 30:25P, it currently sits around 20 below the
- 30:27five average of 24, which also leads to
- 30:30an undervaluation signal. When we look
- 30:32at the blue tunnel, go back the last 10
- 30:34years. Very hard to find the time where
- 30:36we see such a severely undervalued
- 30:38signal that we're seeing in 2026. And
- 30:41the semiconductor operating income
- 30:43growth is accelerated sharply reaching
- 30:45nearly 145%. It gives the valuation
- 30:48genuine earning support rather than hope
- 30:50alone. And consent to EBS, it rises
- 30:53steadily through 2030. And my valuation
- 30:56comes to $516,
- 30:5834% margin of safety with implied upside
- 31:01around 52%. Even the low case at 15 that
- 31:05comes to 368 indicating upside fairly
- 31:08trivial but still showing the protection
- 31:10that we're seeing. The more aggressive
- 31:1225% that comes to 720 more than a
- 31:15double. Wall Street well they sit around
- 31:18533 not too dissimilar from the DCF
- 31:21today. And the reverse DCF that comes
- 31:23around 14% demanding maybe but it is
- 31:26below the 20% that we've used as the
- 31:28base case. It also has exposure to AI
- 31:31semiconductors, networking and
- 31:32infrastructure software. The
- 31:34diversification makes the cash flow
- 31:36stream more resilient than a single
- 31:38product cycle story. The main risks well
- 31:40integration, leverage and the
- 31:42possibility that AI infrastructure
- 31:44spending calls, but unlike several
- 31:46peers, the valuation already leaves room
- 31:48for slower growth. So this combination
- 31:51it earns Broadcom not only S tier but
- 31:53the top spot number one and the stock I
- 31:56believe offers the best risk adjusted
- 31:58place for fresh money today. So as a
- 32:01quick wrap-up here apply materials it
- 32:03completes the board that sits in D tier
- 32:06AMD and KLA they sit in C because the
- 32:08base valuations they offer too little
- 32:10protection. In B tier we've got ASML,
- 32:13Qualcomm, Nvidia and Marvel. I could buy
- 32:16each of these in measured amounts, but
- 32:18none gives me the clean combination of
- 32:20upside and certainty that's required for
- 32:23a tier, which in that we have TSM and
- 32:26lamb research. TSM has got the
- 32:28geopolitical warning label. Lamb is my
- 32:30preferred equipment name and sits
- 32:32closest to an upgrade. Where Micron,
- 32:34SKH, and SanDisk, they make S tier on
- 32:38opportunity, but not on safety. They
- 32:40require smaller positions because the
- 32:42starting cash flow they can collapse in
- 32:44a memory downturn. Brocom as I said that
- 32:48ranks first because it's lower growth
- 32:49case still clears the market price while
- 32:52the base case offers more than 50%
- 32:54upside from a diversified cash flow
- 32:56machine. So if I were deploying fresh
- 32:59capital today I would start with brocom
- 33:01and lamb use a position for TSM and the
- 33:03cickle S tier names. I would not put
- 33:06equal money into every S tier stock.
- 33:08Broco means high conviction. The memory
- 33:11names mean high opportunity with high
- 33:13volatility. And Nvidia, it remains a
- 33:16measured accumulation, not my largest
- 33:18new buy. I would wait for a wider margin
- 33:20of safety before adding apply materials
- 33:23AMD or KLA. The biggest lesson here from
- 33:26the sell-off is that one red heat map,
- 33:28it can hide completely different
- 33:30valuation stories. Some stocks are
- 33:32merely cheaper, others may actually be
- 33:34cheap. and DCF models. They're not price
- 33:37targets carved in stone. They're a way
- 33:39to expose the assumptions today's price
- 33:41requires and to see which assumptions
- 33:43can actually break the result. For
- 33:45memory, the question is whether today's
- 33:47cash flow survives the next down cycle.
- 33:50For Marvel, it's what the combined
- 33:52momentum business looks like after new
- 33:54shares and integration costs. It's why
- 33:56Brocom finishes first. It gives me the
- 33:58cleanest combination of downside
- 34:00protection, credible growth, cash flow
- 34:02quality, and meaningful upside. Let me
- 34:04know in the comments which semiconductor
- 34:06stock you would rank number one, which
- 34:08placement you think I've got completely
- 34:10wrong. And if you enjoy today's episode,
- 34:12smash the like button, hit the subscribe
- 34:13and bell so you're aware of future
- 34:15episodes, and don't forget to click on
- 34:17the pin comment below. We release one
- 34:19weekly article covering severely
- 34:21undervalued stocks as well as an update
- 34:23on the overall market. More importantly
- 34:25though, have a great day. I'll see you
- 34:27all on the next one.
About this transcript
This page contains the full transcript of This AI Stock Crash Is a GIFT — Here’s What I’m Buying by Dividend Talks, generated from the public captions YouTube serves with the video. The transcript has 5,892 words across 882 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.