Top-Analyst warnt: "Der KI-Trade ist vorbei!" — Transcript
Full transcript
- 0:01The big AI trend, the AI boom, just as
- 0:03we've experienced it on the stock market
- 0:05over the last three years, is over.
- 0:10That is the opinion of, in my view, the
- 0:13most important and best fundamental
- 0:15analyst we have on Wall Street, and thus
- 0:17in the US.
- 0:22But the fact that this is no reason to
- 0:24simply sell all your stocks now is in
- 0:26the nature of things, and I wouldn't
- 0:28value this analyst so highly if he
- 0:30didn't delve deeply and thoroughly into
- 0:32this subject.
- 0:35How he views the current situation,
- 0:37which stocks one might want to move away
- 0:39from, and where the great opportunities
- 0:41still lie, that is what I would like to
- 0:43discuss in this video. Let's get started
- 0:46.
- 0:53I welcome you all and I hope you are
- 0:55doing well.
- 0:57Do you know what, from my humble
- 0:59perspective, is a real problem?
- 1:03It's that we are currently in a world
- 1:05where shades of gray are no longer
- 1:07particularly popular—no more nuance, no
- 1:09more compromise.
- 1:13Either for it or against it. For
- 1:15instance, do you think Elon Musk is good
- 1:17or bad?
- 1:19Can't I say that I truly admire what he
- 1:21has achieved as a visionary and founder
- 1:23, while still considering him to be a
- 1:25very questionable character? No, no, no
- 1:28, no, no—either for or against. What
- 1:31about the stock market here? Will it
- 1:33keep rising or will it keep falling?
- 1:35Should I buy AI stocks now? Yes or no?
- 1:37And as stock market investors, we still
- 1:39have the advantage that it's only about
- 1:41the stock market. In politics, it is
- 1:43much more difficult. For, against, black
- 1:45or white. So, that is a—well, if you
- 1:48call it a mindset, that's perhaps a bit
- 1:50of a big word. Call it, just call it a
- 1:53societal phenomenon.
- 1:55You can see this very clearly and
- 1:57especially pronounced when you spend
- 1:59time on Twitter or X. Yes, I know, who
- 2:02still says Twitter these days?
- 2:04There really isn't anything in between
- 2:06anymore.
- 2:08But it is also increasing more and more
- 2:10among the media that, in the past,
- 2:12reported on the stock market in a more
- 2:14balanced way.
- 2:16And I'll take the blame if someone says
- 2:18, "Yeah, but your titles contribute to
- 2:20that too."
- 2:22Regarding the titles, yes, I gladly
- 2:25accept that feedback.
- 2:27But regarding the content of my videos—
- 2:29and I hope that no one forms their
- 2:31opinion based solely on the headlines—
- 2:33it is definitely not the case.
- 2:37And that is why I want to talk to you
- 2:39today about someone I consider brilliant
- 2:41, who doesn't just give a thumbs up or a
- 2:43thumbs down, but actually goes into a
- 2:45bit more depth.
- 2:50And I want to talk to you in a webinar
- 2:53about how buying opportunities exist
- 2:55beyond just the AI sector, and I want to
- 2:57make this webinar free for you.
- 3:02I want to look at three very concrete
- 3:04buying opportunities that are now
- 3:05arising because the market is correcting
- 3:07.
- 3:09As long-term investors, that's what we
- 3:11want; we need these corrections to be
- 3:13able to add to positions at a discount,
- 3:15but it's also an opportunity for active
- 3:18investors.
- 3:20And because there isn't just the big AI
- 3:22trade, the AI trend, but more, even
- 3:24within the tech sector, I thought I'd do
- 3:27this for you.
- 3:31The webinar is absolutely free for
- 3:33everyone who is in the app on Thursday.
- 3:37I will keep it short and concise, three
- 3:39very specific buying opportunities as
- 3:41they appear to me.
- 3:43Yes, I will of course include the
- 3:45disclaimer that you have to make your
- 3:47own decisions.
- 3:49So, make use of that so far.
- 3:52And you can almost verify it, because
- 3:54you can see how many people are signing
- 3:56up.
- 3:58A lot of people are signing up for the
- 4:00app because it's free, because you get
- 4:02value there, because you get content
- 4:04there that you don't get here, because
- 4:06it just makes more sense for me not to
- 4:08repeat myself everywhere, but to create
- 4:10exclusive content.
- 4:14So feel free to use it, the webinar is
- 4:16Thursday at 6 PM, and you will be able
- 4:19to watch the recording immediately
- 4:21afterward.
- 4:24Sign up for this free app; including the
- 4:26download, it shouldn't take a minute.
- 4:33And now I want to talk about Aswath
- 4:34Damodaran, who just published an article
- 4:37.
- 4:40Yes, I won't go through the original
- 4:42with you now, but I'll summarize it.
- 4:45One could more or less say the big AI
- 4:47trade is over.
- 4:49However, it's now getting really
- 4:51interesting.
- 4:54And that's because it's no longer the AI
- 4:56trade, but rather you should pay
- 4:57attention to and think about which
- 4:59stocks you actually have in your
- 5:01portfolio.
- 5:03And I want to show you the difference.
- 5:06Look here, this is just some random Neo
- 5:09Cloud, a Cipher, or look at an uh Iron.
- 5:14That actually still looks somewhat
- 5:16decent on the chart, but still, we were
- 5:18at 80 dollars and then at a low below 20
- 5:20dollars; that's quite a lot of
- 5:22volatility.
- 5:24Quite a bit of the AI fantasy has been
- 5:26taken out of that.
- 5:28In the webinar, I’ll discuss how you can
- 5:31decide from the outset: okay, do I want
- 5:33to open a trade where I aim for a few X
- 5:36, two, three, four hundred percent?
- 5:39Then you're inevitably in a sector that
- 5:41has just been heavily punished.
- 5:43Or would you rather say, yes, I want to
- 5:45stay invested in this AI trade, but more
- 5:47or less conservatively?
- 5:49And take a look at Palantir in
- 5:51comparison.
- 5:53Yes, it has corrected a bit too.
- 5:55Or Amazon—heads up, I am privately
- 5:58invested here.
- 6:00Yes, what exactly has been unwound here
- 6:02regarding the big AI trade?
- 6:05That’s also the problem when people stop
- 6:07making distinctions—and many, "skeptic"
- 6:10is almost too helpful a term, I mean
- 6:12permabears who say this AI trade is
- 6:14being unwound and now there's a crash
- 6:16coming.
- 6:18Don't you have to see the signals for
- 6:19that somewhere?
- 6:22For the hyperscalers, that would likely
- 6:24even be a bullish scenario for reasons
- 6:26we’ll discuss right now.
- 6:30So, Aswath Damodaran distinguishes four
- 6:33groups with four different interests in
- 6:35this AI trade.
- 6:39What he says, by the way, is that the
- 6:41last few weeks were a paradigm shift,
- 6:43and that’s also something you can take
- 6:45away.
- 6:46Once there are such clear corrections,
- 6:48you’ve already experienced that in many
- 6:50, many smaller industries and sectors.
- 6:54Whether we’re talking about quantum
- 6:56technology or hydrogen—where no money is
- 6:58being made yet, yes, just like in the AI
- 7:01sector.
- 7:02Many of the pure AI stocks, including
- 7:04the model manufacturers, aren't making
- 7:06any money yet on the bottom line. So, in
- 7:09that sense, even if it has a different
- 7:11dimension, one can certainly compare the
- 7:13two. And when such a clear sense of
- 7:15disillusionment sets in, it doesn't mean
- 7:18the rally is over, but to simply say it
- 7:20will be like the last three years and
- 7:22the same stocks will rise—that is rarely
- 7:24if ever the case, and that’s why we need
- 7:26this distinction.
- 7:29He distinguishes between the four groups
- 7:32.
- 7:33He says one, the first group are the
- 7:35suppliers.
- 7:38Who is the supplier par excellence right
- 7:40now for reasons—Jensen Huang, you've
- 7:42heard about it, but that's not the point
- 7:45, there are other semiconductor
- 7:46manufacturers as well.
- 7:50Everything we see here in the
- 7:52semiconductor index, among others—they
- 7:54have corrected as well, but if you look
- 7:57at it, that was April '25 and here we
- 7:59are now.
- 8:01So you can’t conclude from that, "well,
- 8:03that's the end of it." Giving back all
- 8:06the gains.
- 8:07For anyone who had that feeling in their
- 8:09portfolio—I talked about that on Monday
- 8:11—maybe you were just invested too
- 8:13heavily in the high-high-high-risk
- 8:15stocks.
- 8:16You don't need to cry about it, look
- 8:18back, or feel regret; you just do better
- 8:20for the future.
- 8:22So, these are the semiconductor
- 8:24manufacturers, companies in electrical
- 8:26engineering, energy supply, and data
- 8:28center construction.
- 8:32These companies all, every single one of
- 8:34them, have a huge and often overlooked
- 8:36advantage today.
- 8:39They make money.
- 8:44They actually make a lot of money, and
- 8:46that is obviously a big difference in
- 8:47market phases like this—whether you make
- 8:50money from others 'investments or if
- 8:52you're the one making the investments.
- 8:58These companies will foreseeably
- 9:00continue to profit from this trend in
- 9:02the coming years.
- 9:04The second group is the infrastructure
- 9:06builders—take Amazon, for example—yes,
- 9:09who are the big cloud providers?
- 9:12Amazon, Microsoft—often everyone just
- 9:15copies everyone else.
- 9:18Maybe that’s also a bit of a
- 9:19disadvantage of AI. Yes, Microsoft held
- 9:22up because Microsoft still has a
- 9:24positive cash flow. No reasonable
- 9:26analyst—pardon me, I take back that
- 9:28remark—no reasonable analyst would look
- 9:30at it that way, because every other
- 9:32hyperscaler could also immediately say,
- 9:34"Fine, if you want 100 billion in cash
- 9:36flow next quarter, you'll get it, we'll
- 9:38just scale back investments."
- 9:42So, it made sense for Microsoft right
- 9:44now, but to say that’s why they will
- 9:46outperform everyone else is just not a
- 9:48very clever approach.
- 9:55We have one—these are also
- 9:56infrastructure builders—Alphabet came
- 9:58under pressure more because they are
- 10:00likely putting their own search engine
- 10:02business under pressure, perhaps even
- 10:04disrupting it, with AI.
- 10:11Technically, that would also be an
- 10:13infrastructure builder.
- 10:17But why is a company like CoreWeave also
- 10:19being punished?
- 10:21Simply because they aren't making money
- 10:23, right?
- 10:24Those are negative figures; although,
- 10:26either the numbers are already out
- 10:28tonight—I'm recording this video a day
- 10:30early—or they are coming today, results
- 10:32have to be reported now.
- 10:35They are investing—Amazon, Microsoft,
- 10:37and so on—investing hundreds of billions
- 10:40into data centers and infrastructure,
- 10:42and that initially weighs on their free
- 10:44cash flows, of course.
- 10:47The crucial question is whether they
- 10:49will later earn an appropriate return on
- 10:51these investments.
- 10:54That is somewhat clearer and more
- 10:56obvious for the cloud providers, because
- 10:58they will recoup a large portion of what
- 11:00they’ve invested within the next two
- 11:02years.
- 11:03But of course, the market is watching
- 11:05that very closely. It doesn't just want
- 11:07the investment back; it also wants a
- 11:09very decent return on top of it.
- 11:14And that is the statement you cannot
- 11:16reliably make today, because today, no
- 11:18one should really dare to say where AI
- 11:20will be in 2030.
- 11:25Then we have the third group, which are
- 11:27the actual AI product companies, like,
- 11:30uh, Anthropic or OpenAI.
- 11:36To be fair, you would probably also have
- 11:38to include a Palantir, and these
- 11:39companies simply have to prove that they
- 11:42can develop profitable business models
- 11:44out of these technologically undoubtedly
- 11:46impressive products.
- 11:51Personally—and this is not a
- 11:53recommendation regarding the
- 11:54subscription or whether there will be
- 11:56subscription gains or not—we will
- 11:58certainly take another look once we know
- 12:00more about the Anthropic IPO.
- 12:04So, uh, the numbers and the growth are
- 12:07undoubtedly impressive.
- 12:11I personally would not invest in OpenAI
- 12:13or Anthropic at the current valuation
- 12:15because it is too uncertain for me.
- 12:18These could be very profitable business
- 12:20models.
- 12:22But I don't belong to the group of
- 12:24people who are convinced today; I
- 12:26believe that AI will not be rolled back
- 12:28, but whether the companies or the
- 12:30infrastructure providers will actually
- 12:32make that much money with their models—
- 12:34it could be, but it's too uncertain for
- 12:36me, so I choose other companies.
- 12:39And then there is a fourth group.
- 12:41Damodaran names them quite explicitly
- 12:44here: the customers.
- 12:46That is the big question of the time,
- 12:48because in the end, someone has to pay
- 12:50for all this infrastructure. Companies
- 12:52and private customers must be willing to
- 12:54spend significant sums on AI services
- 12:56over the long term, and that is the
- 12:58crucial point. Yes, I agree with him on
- 13:00that. He estimates the current annual
- 13:02revenue from AI products and services at
- 13:05around 250 billion; that is a large
- 13:07market, but measured against the
- 13:09investments, measured against what is
- 13:11currently flowing in, and measured
- 13:13against the stock market valuations,
- 13:15that is still very, very little.
- 13:19And the entire AI boom is based on the
- 13:21assumption that these hundreds of
- 13:23billions will eventually have to become
- 13:25several trillion, otherwise the
- 13:27investments being made today would be
- 13:29too high.
- 13:32Why are interests diverging? Until now,
- 13:34from Damodaran's perspective, everyone
- 13:37had more or less the same interest. We
- 13:39remember: the more capex, the better;
- 13:41the neo-clouds, everyone somehow
- 13:43benefited.
- 13:46But he says the time of investing as
- 13:48much as possible in AI is over.
- 13:51Because they simply do not all have the
- 13:53same interests. Let's take Nvidia as an
- 13:56example. If Microsoft, Amazon, and
- 13:58Alphabet keep building new data centers
- 14:00, Nvidia undoubtedly earns excellent
- 14:02money from it. We will see on the 17th.
- 14:04Of November, just how excellent. Even if
- 14:07a portion of these data centers should
- 14:09end up underutilized later, Nvidia has
- 14:11already sold its chips regardless. For
- 14:13the operators, the calculation naturally
- 14:15looks different. They have to
- 14:16pre-finance, operate, and depreciate the
- 14:19infrastructure. And if demand falls
- 14:21short of expectations, then the return
- 14:23on capital declines.
- 14:25So what is excellent for Nvidia might,
- 14:28in retrospect, turn out to be
- 14:30economically problematic for Microsoft
- 14:32or, uh, Amazon.
- 14:34And now let’s flip the example, or
- 14:36rather, that’s what Damodaran does in
- 14:39his text.
- 14:40Assuming the hyperscalers come to the
- 14:41conclusion that they need to be more
- 14:43cautious with their expansion.
- 14:46They reduce their investment plans and
- 14:48focus more on utilizing their existing
- 14:50infrastructure.
- 14:52Then that could be very positive for the
- 14:54hyperscalers themselves. They would
- 14:56likely become cash-flow positive again
- 14:58immediately, capital expenditures would
- 15:00drop, but
- 15:04the return on capital improves for those
- 15:06from whom the chips are ordered.
- 15:09So for an Nvidia or other semiconductor
- 15:11companies, that is logically not
- 15:13positive at that moment, because growth
- 15:16is still very good, but significantly
- 15:18decreasing.
- 15:19A slowdown in the AI infrastructure boom
- 15:22does not necessarily have to be negative
- 15:24for all AI stocks.
- 15:28You remember the initial reaction after
- 15:30the blue wave last Wednesday, or rather
- 15:33Thursday, Wednesday, Thursday, like that
- 15:35.
- 15:38AI stocks came under pressure, but the
- 15:40hyperscalers gained.
- 15:42And that, in my opinion, is one of the
- 15:44most important statements of the entire
- 15:46analysis. The market is now beginning to
- 15:48differentiate.
- 15:50And Damodaran also recognizes this shift
- 15:52in the price performance. Between 2022
- 15:54and 2025, the total market
- 15:56capitalization of the examined AI
- 15:59infrastructure rose by around 180%.
- 16:03And at that time, the companies were
- 16:05only investing; the economic success was
- 16:07still pending.
- 16:10By the end of 2026, these examined AI
- 16:12companies together reached a share of
- 16:15roughly 37%of the total US stock market
- 16:18capitalization.
- 16:20And that shows how strongly the
- 16:22expectations were already priced in here
- 16:24.
- 16:2537%of the total stock market
- 16:27capitalization earning 250 billion.
- 16:31There is, of course, a discrepancy there
- 16:33. We are simply talking about a
- 16:35valuation that was ambitious.
- 16:38And since then, the development has
- 16:40become significantly more uneven.
- 16:42Damodaran then also addresses the
- 16:44elections. He says that on November 5th
- 16:46. November, according to calculations,
- 16:49within a single trading day, almost 1.8
- 16:52billion, million—no, wait, one more—
- 16:55trillions of dollars in market value
- 16:57were pulverized.
- 16:59And that primarily affected the
- 17:01semiconductor and electrical engineering
- 17:03companies, while the hyperscalers, as I
- 17:06indicated, were actually able to
- 17:07slightly increase their market value on
- 17:10balance.
- 17:12He suspects a plausible connection there
- 17:14.
- 17:16That means the political changes could
- 17:18make building new data centers more
- 17:20difficult. Especially in the US,
- 17:22resistance to large-scale projects is
- 17:24currently increasing. Electricity prices
- 17:27, energy supply, water supply, water
- 17:29consumption, strain on local
- 17:30infrastructure, and that would be a
- 17:32problem for the suppliers. For the major
- 17:34operators, however, this slowed pace of
- 17:37investment could even represent a
- 17:39certain relief.
- 17:40That is the reason why the stocks of
- 17:42these companies look somewhat bullish.
- 17:46And then I would like to add one final
- 17:47term for you here.
- 17:51By the way, one of the three stocks I'm
- 17:53discussing in Thursday's free webinar
- 17:55has nothing to do with AI at all.
- 17:59You can really take a deep breath there
- 18:01and ask, wait, but AI? No, nothing to do
- 18:04with AI. It exists, too. Especially
- 18:06after last week, you would have noticed
- 18:08that. There is more out there than just
- 18:11AI stocks.
- 18:12So sign up, get the app.
- 18:14If you have the app, you'll
- 18:16automatically be included this Thursday
- 18:18the 8th at 6 PM.
- 18:21He calls it here, I don't know, I didn't
- 18:23know how to translate that.
- 18:27"The great market of illusion" doesn't
- 18:29sound very clever either, does it? It
- 18:31refers to the collective overestimation
- 18:34of huge future markets. This is, by the
- 18:36way, something we know from previous
- 18:38technology cycles.
- 18:41This new technology promises everything
- 18:43at first but must then pass the reality
- 18:45check.
- 18:49Especially because it usually doesn't
- 18:51happen as fast as many think.
- 18:54A huge market by no means guarantees
- 18:56that all companies in that market will
- 18:58generate attractive returns.
- 19:04Damodaran assumes that we are moving out
- 19:06of a phase in which we have already seen
- 19:08overinvestment.
- 19:11He talks about infrastructure and the
- 19:13long-term commitments of AI product
- 19:15companies to purchase computing power.
- 19:19These are capacities that have been
- 19:21secured for years, even though nobody
- 19:23knows exactly how much they will earn
- 19:25from them yet.
- 19:26That is something you simply cannot
- 19:28brush aside.
- 19:31And because these companies are
- 19:33economically intertwined in a certain
- 19:35way, these difficulties can naturally
- 19:37spread.
- 19:39However, from his perspective, it is not
- 19:42an argument that artificial intelligence
- 19:44is fundamentally in a bubble.
- 19:46I like this balance.
- 19:49Yes, he expressly points out that
- 19:51overinvestment is often a component of
- 19:53major technological upheavals.
- 19:57It is what makes the progress that
- 19:59brings such investment possible in the
- 20:01first place.
- 20:03And he says, Anthropic will be the big
- 20:05test case.
- 20:07Originally, a valuation of 1.5 to 2
- 20:09trillion dollars was being discussed.
- 20:13Um, he thinks it might be less by now.
- 20:18The equity value from Damodaran's
- 20:21perspective is 1.25 trillion dollars.
- 20:25That is not a forecast for the issue
- 20:26price, but it might be a benchmark.
- 20:30If we come in significantly below that,
- 20:33assuming Anthropic even goes through
- 20:35with the IPO, it would likely have a
- 20:37negative effect on this sector.
- 20:39We have seen that the overall market
- 20:41doesn't have to collapse because of that
- 20:42.
- 20:43Should we come in above that, then it is
- 20:45a somewhat more bullish sign.
- 20:51We will keep an eye on this, here and
- 20:53everywhere else you see me, and I'll
- 20:54wrap things up at this point.
- 20:58Following the moderator makes sense to
- 21:00me, because he really puts things into
- 21:02perspective quite well.
- 21:04So, that's it for today.
- 21:06I'll see you tomorrow in the webinar and
- 21:08I'm looking forward to it.
- 21:10It's going to be really good. I wish you
- 21:12all a pleasant Wednesday evening. See
- 21:13you then, and ciao.
About this transcript
This page contains the full transcript of Top-Analyst warnt: "Der KI-Trade ist vorbei!" by Erichsen, generated from the public captions YouTube serves with the video. The transcript has 3,117 words across 542 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.