YT YouTube2Text

YouTube2Text → Transcripts → This is What “Always” Happens Before a Market Crash

This is What “Always” Happens Before a Market Crash — Transcript

by Mark Tilbury Economics · 3,486 words · 519 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00Market crashes don't come out of
0:01nowhere. There's always a warning and it
0:04looks strangely similar every time if
0:06you know what you're looking for. In
0:08June, nearly two trillion dollars was
0:10wiped out of the markets in a single
0:12day. An entire national stock exchange
0:15had to halt trading and then within
0:17weeks stocks were pushing record highs
0:19like it never even happened. 2008 felt
0:22exactly like that before one of the
0:24biggest crashes in history and I know
0:27because I lived through it with millions
0:29of dollars on the line and came out
0:31stronger than before. Look, I've been
0:33investing for 40 years and I'll be the
0:35first to tell you that these crashes
0:37aren't random. They actually follow a
0:39five-stage pattern and we just reached
0:42stage five. So, let's unpack each of
0:44them, discuss exactly what's triggered
0:46each stage this time around and most
0:48importantly, what [music] you can do to
0:50protect yourself.
0:55I want you [music] to think about the
0:56last time you talked yourself into
0:58something you already knew was a bad
1:00idea. Maybe it was a car you couldn't
1:02quite afford or maybe it was something
1:04simple like eating a second slice of
1:07cake you definitely didn't need. It's
1:09hard to explain but I think we could all
1:12agree there's quite a specific feeling
1:14in that moment. A little voice in your
1:16head that knows the truth and a much
1:18louder voice that says, "Ah, screw it.
1:20It'll be fine just this once." Now,
1:22imagine that feeling but shared by the
1:24entire stock market at the same time.
1:27We've seen this time and time before
1:28especially with dot com stocks back in
1:302000 but I won't bore you with the
1:33history lesson because I know you've
1:35heard it over a hundred times before.
1:37But in all of those past moments, smart
1:40people convinced themselves that the old
1:42rules about money and value didn't apply
1:44anymore and this time it was different.
1:46There's a measure Warren Buffett once
1:48called the best single gauge of where
1:50valuation stand at any given moment.
1:52People call it the Buffett indicator and
1:55all it does is compare the total value
1:57of the US stock market to the size of
2:00the entire US economy. Right now, as I'm
2:03recording this video, that number hit an
2:05all-time record high of around 238%.
2:10But I know that number on its own
2:11doesn't really mean anything. So, let's
2:14put it this way. Warren Buffett himself,
2:16who is universally considered the
2:18greatest investor of all time, said that
2:20when that number gets near 200%
2:23you're playing with fire. We're now
2:25sitting well above that, higher than the
2:27dot-com bubble, and higher than ever in
2:30recorded history. Wall Street's own
2:32analysts are forecasting long-term
2:34earnings growth for America's biggest
2:35companies of around 25% a year, which
2:39sounds great, because
2:40who doesn't like a bit of profit? But
2:42when you zoom out a little, that's a
2:44higher level of confidence than
2:45investors were pricing in right at the
2:47very top of the dot-com bubble, just
2:50before everything came crashing down.
2:52Companies like Nvidia and Broadcom have
2:54been trading at well over 20 times their
2:57sales, not their profits, their sales.
3:00And when you think about it, that kind
3:02of valuation only makes sense if you
3:04believe that the growth never slows
3:06down, which is far from guaranteed. None
3:09of this is a secret. Every person I
3:11quote in this video knows these exact
3:13numbers, and Buffett himself is sitting
3:15on a mountain of cash, which
3:18I think already tells you what he
3:19thinks. I mean, right now there's loads
3:22of videos being put out there talking
3:24about the AI bubble. I've a couple
3:26myself, and everyone in the comments
3:29seems to agree that we're in trouble.
3:31Yet, at the same time, people still
3:33decide to invest in these risky
3:34companies. But that's the worrying part
3:36about the delusion stage, because it's
3:38clearly not a lack of information.
3:40Clever people are just looking at the
3:42same warning signs and deciding that it
3:44doesn't really apply this time around. I
3:47mean, I've sat at dinners where someone
3:48lays out in perfect detail exactly why
3:51something is overvalued. And then, in
3:54the next breath, they tell you they just
3:56bought more of it. It's not really
3:58stupidity, and more so just human
4:01nature. Everybody crumbles under the
4:03fear of missing out. And if I'm being
4:05honest, I feel it, too. I'm 58 years
4:09old. I've been doing this for nearly 40
4:11years, and I still get that itch when I
4:13see people making money on something
4:15exciting without me. But every single
4:18crash I've ever witnessed started with
4:20the delusion that value doesn't matter
4:22anymore. A market being expensive isn't
4:25exactly a crime, and it isn't what kills
4:27people, either. I've overpaid for plenty
4:30of things in my life and lived to tell
4:32the tale. The real trouble only starts
4:34when you can't actually trust the
4:36companies you're overpaying for, which
4:38brings me on to
4:42Have you ever tried to read something so
4:44complicated, like a a phone contract or
4:47an insurance policy, that your mind just
4:50gives up and you tick the box anyway? I
4:52think everyone's done that, well, at
4:54least once in their life. But that
4:56little moment of giving up is a blind
4:59spot. And when a financial market gets
5:01so complicated, the people whose job it
5:04is to monitor it just tick the box and
5:06move on. In 1987, it was portfolio
5:09insurance, which was a clever bit of
5:11financial engineering that was supposed
5:13to make crashing impossible, but instead
5:15helped to cause one. In 2008, it was
5:18mortgages, chopped up and repackaged so
5:21many times that the people buying them
5:23genuinely had no idea what was inside.
5:25In both cases, the complexity was the
5:28danger, because nobody even bothered to
5:30check what was going on. For me, 2026's
5:33version of this is something called
5:35circular financing, and it's happening
5:37within the world of AI. I've spoken
5:39about this before, but I think it's too
5:41big and too relevant not to discuss
5:44here. And if If already heard this,
5:46stick with me, because I also have some
5:48new information which I'll get into in a
5:50second. Picture three companies in a
5:52circle, Nvidia which designs computer
5:55chips, OpenAI which runs the LLMs, and
5:58Oracle which provides the cloud
6:00computing to run it all. Now watch the
6:02money move. Nvidia invest billions into
6:05OpenAI. OpenAI then takes the money and
6:08commits hundreds of billions of dollars
6:10to Oracle for computing power. And
6:12Oracle takes that money and spends it by
6:16buying computer chips from Nvidia. The
6:18same dollars have gone all the way round
6:20the circle and landed back right where
6:22they started. And every time they pass
6:24through a company, that company gets to
6:26record it as brand new demand and
6:29growth. By some estimates, around a
6:31trillion dollars in deals is now looping
6:33around the same small handful of
6:35companies like this. Even though no new
6:37money has actually entered. If you've
6:39been watching from the outside, it
6:41looked like separate transactions. And
6:43that's the bit that's worrying me.
6:45Because some of what the market is
6:47calling raw in demand for AI might just
6:49be the same money getting counted three
6:51or four times. But this isn't just a
6:53modern problem. And back in 1999, there
6:56was a company called Lucent Technologies
6:59which at its peak was the most widely
7:01held stock in America. Lucent's
7:03customers were new telecom startups that
7:05couldn't actually afford its equipment.
7:07So Lucent lent them the money to buy it
7:10and recorded the sales as new revenue.
7:12They called it vendor financing. And for
7:15two years or so, it looked absolutely
7:17amazing until the startups ran out of
7:19money, couldn't make the payments, and
7:21Lucent's stock fell 99% as they
7:24discovered they never actually made the
7:26money their sales numbers suggested. So
7:29with that in mind, you'd like to think
7:30someone official would be keeping an eye
7:32on things. However, the US Treasury only
7:35produced its first proper draft warning
7:37about the AI bubble burst just this
7:39month. And global regulators have only
7:42just started warning that the AI
7:44infrastructure spending could become a
7:46real threat to the wider financial
7:48system, but we're already years into the
7:50build-out, and that's the annoying thing
7:52about regulators. They're almost never
7:55early. By the time the official warning
7:57arrives, the risk it's outlining is
7:59usually already old news. They act like
8:01they're shining a light in the dark, but
8:03in reality, they're warning you about
8:05pothole that just burst your tire. A
8:07rule that's never let me down is not to
8:10wait for a regulator to tell you
8:11something's risky, because they don't
8:13get paid to be early. They get paid to
8:16be right, which usually comes
8:18afterwards, and you shouldn't need some
8:19guy wearing a suit to confirm what your
8:21eyes are already seeing. But the thing
8:23is, a blind spot doesn't automatically
8:26mean disaster, and this AI problem could
8:28potentially go on for years before
8:30anyone gets hurt. What accelerates it is
8:34adding fuel to the fire.
8:38>> [music]
8:39>> The stock market dropping 10% is
8:41completely normal, and I'd even go as
8:43far to say it's healthy. Most
8:46corrections come and go, and no one even
8:48remembers them a year later. So, what
8:50turns a normal market correction into
8:522008 all over again? Well, the answer,
8:55almost every time, is debt. If we go
8:58back to 2008 for a moment, house prices
9:01falling on their own would have been
9:02painful, no doubt about it, but
9:04definitely survivable. What turned it
9:07into a global catastrophe fact that the
9:09whole thing was built on borrowed money.
9:11When you buy an asset with your own
9:13money and it drops, you might feel
9:14poorer, but when you buy it with
9:16borrowed money, you can get wiped out
9:18completely, and so can whoever lent you
9:21the money. Basically, debt is the fuel,
9:24and in 2026, it's hiding in a corner of
9:27the market that most everyday people
9:29have never even heard of, which is
9:31partially what makes it such a big
9:33problem. It's called private credit,
9:35which is essentially when lending
9:36happens away from traditional banks like
9:38big funds lending money directly to
9:40companies out of sight and off the
9:43public books. And over the last couple
9:45of years, a huge chunk of it has been
9:46pulled directly into AI, which to put it
9:49lightly is a very speculative and
9:51somewhat controversial investment. I
9:53mean, some people believe it's the
9:55future. Others think it's dangerous, and
9:58some people say it's just a fad that
10:00will fizzle out in a few years.
10:02AI-related deals made up around a third
10:04of all private credit issued in 2025,
10:07which is pretty crazy. But it gets even
10:10crazier when you consider that over the
10:12previous 5 years, that figure averaged
10:14just 17%. So in a very short space of
10:17time, the world of private credit has
10:19become completely drenched in AI risk.
10:22And now, the people who lend this money
10:24for a living are starting to get a
10:26little bit shaky.
10:27>> I don't get nervous,
10:30but I'm starting to get a bit shaky, you
10:31know? I've been on a little bit of it.
10:33>> Big banks have started modeling what
10:35happens if the AI spending doesn't turn
10:37into real revenue fast enough. Morgan
10:40Stanley reckons default rates in private
10:42lending could surge to 8%, and UBS says
10:45if the AI disruption is rapid and
10:47severe, they could hit 15%. And this
10:50isn't just hypothetical. One closely
10:52watched measure of private credit
10:54defaults has already climbed to a record
10:56of 6%. To put that into perspective,
10:59those numbers start approaching the kind
11:00of stress we saw during the pandemic.
11:02And because so much of this debt is
11:04private, everyday people like you and me
11:06can't easily see the cracks forming. We
11:09only really find out that it was
11:10cracking after it's already broken. The
11:13main problem here, at least in my
11:15opinion, is that a lot of this money was
11:17lent against the exact companies AI is
11:20supposed to replace, like software firms
11:22for example. They sell you a
11:24subscription to a tool to complete a
11:26task that if the AI dreamers are right,
11:28an AI will soon do instead. If AI fails
11:31or even falls short, all our spending
11:33was for nothing and the loans go bad.
11:36But, if they succeed, it kills the
11:38companies that the money was lent
11:40against and the loans go bad anyway.
11:42It's like flipping a coin, but losing
11:45whether you pick heads or tails, which
11:47makes no sense whatsoever. But, that's
11:49where the financial system is heading.
11:51But, if that's true, it raises the
11:53obvious question,
11:55why is everyone still buying?
11:59When I first started investing, [music]
12:01the risk felt real and that fear kept
12:03people honest and made them ask hard
12:05questions before they invested. But,
12:07that mindset is basically dead now. I
12:09mean, if you think back over the last 15
12:11years or so, every time the market has
12:13had a serious decline, something has
12:16been there to catch its fall. Whether it
12:18was a rate cut, emergency lending, a
12:20bailout, or a policy reversal, it
12:23doesn't matter. The point is that every
12:25time investors have looked down and seen
12:27the ground rushing up towards them, a
12:29safety net has appeared out of nowhere.
12:31And as humans, we're pretty good at
12:33learning patterns. Do something enough
12:35times and we internalize it completely.
12:37So, a whole generation of investors has
12:39been essentially trained to believe that
12:42someone always steps in to save the day.
12:44It's kind of like the one friend
12:46everyone has whose parents always give
12:48them money when they make a mistake. It
12:50seems great, but you'll notice that that
12:52person never learns. And one day,
12:55they'll mess something up beyond repair
12:57and have one big lesson that could have
12:59easily been avoided by making a few
13:01smaller mistakes first. On Wall Street,
13:03they've got a nickname for a version of
13:05this called the Fed put. It's basically
13:07the idea that central bank will always
13:09come to the rescue if things get bad
13:11enough. But, honestly, I think it's
13:14grown into something much bigger than
13:15the Fed. It's become a kind of blind
13:18faith in the entire system, a belief
13:20that no matter how reckless things get,
13:22someone with a big enough checkbook is
13:24watching and they won't let it fall.
13:26This is a psychological engine that
13:27drives everything we're discussing in
13:29this video. It's why stage one, two, and
13:32three can all be true at the same time,
13:34but the market still climbs. Because if
13:36you genuinely believe you can't lose,
13:38then why would you ever bother pricing
13:40in risk or worrying about that? It just
13:42seems like unnecessary stress. A lot of
13:44people assume that market crashes happen
13:46because no one's really paying
13:48attention, but it's actually quite the
13:50opposite. Everyone's paying attention.
13:52Everyone can see the warning signs and
13:54chooses to ignore them. Not out of
13:56stupidity, although it is stupid, but on
14:00purpose, because they've decided that a
14:02rescue is guaranteed and nothing else
14:04matters. This is a market that in a
14:06single month saw major tech stocks
14:08crash, panic spread throughout global
14:10markets, and even warnings emerge about
14:13the AI bubble, then still ended at
14:15record highs anyway. Now, some people
14:18will look at that and call it the market
14:20being resilient, and maybe they're
14:22right. But, I've been around long enough
14:24to know that sometimes the most
14:26dangerous markets are the ones that stop
14:28reacting to bad news. Because when
14:30investors become convinced that every
14:32warning can be ignored, that's usually
14:34when complacency [music]
14:35starts creeping in. Complacency is what
14:38leads us on to
14:43Markets don't usually go from calm to
14:45chaos in a [music] single day. More
14:47often than not, there's a tremor first.
14:49And in June, we had one of those
14:51moments. It all started with memory
14:53chips, not because prices fell, but
14:56because investors suddenly doubted the
14:58AI spending propping them
14:59[clears throat] up. The Nasdaq had one
15:00of its worst stretches in a long time.
15:02Micron fell around 13%. Samsung and SK
15:06Hynix dropped around 12%, and then it
15:09went global. South Korea's Kospi index
15:11fell 10% in a single day, so fast that
15:15the exchange had to halt trading
15:16altogether. And by the time the dust had
15:18settled, close to a trillion dollars had
15:21been completely wiped out. For a few
15:23days, it felt like the floor was giving
15:25way, and then
15:27it just stopped. And within 3 weeks, the
15:29S&P 500 climbed all the way back to the
15:32very edge of its record high. Now, I've
15:35watched the market pull this move
15:36before, but it's one time in particular
15:39that I've never forgotten. March 2008.
15:42One of the biggest, most respected
15:43investment banks on Wall Street, Bear
15:45Stearns, basically fell apart over a
15:48single weekend. It was so bad that it
15:50had to be rescued in a rush takeover
15:53with J.P. Morgan buying the company for
15:55a fraction of what it had been worth
15:56before. For a few days, just like this
15:59June, it felt like the end of the world.
16:01And do you know how the market
16:02responded? It went up. There was a clear
16:05warning, and for the next couple of
16:06months, stocks rallied anyway because
16:09people decided that Bear Stearns was a
16:11one-off, and the scare was over. Six
16:13months later, Lehman Brothers collapsed
16:16and took the entire global economy down
16:18with it.
16:18>> Going to be one of the watershed days in
16:20financial markets history.
16:21>> It was a manic Monday in the financial
16:23markets.
16:25The Dow tumbled more than 500 points
16:27after two pillars of the street tumbled
16:30over the weekend.
16:31>> Bear Stearns was never the crash. Bear
16:33Stearns was the tremor. And that's kind
16:35of what June feels like to me. But I
16:37want to be careful with my words here,
16:39because I'm not saying June is
16:41definitely our version of this. I don't
16:43know that for sure. No one does. And
16:46anyone pretending to is lying. I'm just
16:48making a comparison of flagging that it
16:50has similarities. A serious crack, a
16:53brief moment when everybody stops and
16:55notices, then a very quick decision to
16:58move on like nothing ever happened. But
17:00not knowing isn't a reason to switch
17:02off, and in my opinion at least, it's
17:04more of a reason to stay awake, because
17:07we've had our warning, and now it's time
17:09to find out what it was warning us
17:10about.
17:14The exact sequence [music] of everything
17:15we've just discussed has played out for
17:17400 years. Whether it was tulips in
17:20Holland, railways, the roaring 20s, dot
17:22com, or this, the assets change every
17:25time, but the psychology underneath is
17:27always the same. I've watched this
17:29five-stage pattern play out multiple
17:31times throughout my life while having
17:33millions invested, and I've made it out
17:35the other side richer every time. Not by
17:37being clever, but by refusing to believe
17:40that this time it'll be different, while
17:42also not panicking and selling
17:44everything. That balance of staying
17:46invested, but not becoming delusional,
17:48is basically the entire game of a
17:50successful investor. It's funny, really,
17:53because when you think about it, it's
17:55actually very easy, but not many people
17:57managed to do it. So, as for what I'm
18:00doing, well, I'm still investing every
18:02single month, the same as always. I'm
18:05well aware that a crash could be right
18:06around the corner, but it could also be
18:08years away, and trying to time the top
18:11is one of the most expensive mistakes I
18:13see people make, and I've made it myself
18:15in the past. However, I'm making sure
18:17that I'm not fixated on the same AI
18:20companies that everyone else has piled
18:22into, because when everyone is standing
18:24on the same side of the boat, I start to
18:26get nervous. And as things stand, the
18:28market is leaning hard on about seven
18:30companies, so I've deliberately spread
18:32myself wider than that. A lot of people
18:35like Warren Buffett, for example, are
18:37sitting on record levels of cash at the
18:39moment, and I'll admit that I'm keeping
18:41a meaningful cash position on the side.
18:43Now, you might think this is just trying
18:45to time the market, but like I said, I'm
18:47still investing consistently, because if
18:50that June tremor turns into any sort of
18:52crash, I want to be the person calmly
18:54buying while everyone else is panicking.
18:56If you want to understand the riskiest
18:58moment of the AI bubble, then I'm going
19:00to leave that video right up there, but
19:02don't click on it just yet. Make sure to
19:04subscribe if you want to stay ahead of
19:05everyone else, okay? I'll see you over
19:07there.

About this transcript

This page contains the full transcript of This is What “Always” Happens Before a Market Crash by Mark Tilbury Economics, generated from the public captions YouTube serves with the video. The transcript has 3,486 words across 519 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 translate the transcript into 100+ languages.

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, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.