Michael Burry Really F**K'd Up | AI Collapse

People are pissed. Michael Bur is back

at it again. And people are super pissed

because now Michael Bur is literally

attacking everybody's favorite play of

the last 3 years and that is artificial

intelligence and chips. Michael Bur

straight up suggesting that Oracle,

Meta, and companies like this such as

Google, Microsoft, Amazon may all be

totally overstating their revenues.

borderline defrauding their investors.

And in this video, I'm going to tell you

exactly how Michael Bur could be 100%

right, but how he's 100% wrong today.

And the best way to make this analogy is

by talking about my pee peep. Yeah, I I

really apologize that we have to do

this, but honestly, we've explained this

on the channel before, and most people

just don't understand this or they

didn't watch this, and that's okay. I'm

going to assume you will understand and

that you didn't watch this explanation

yet, but we're going to use a very easy

explanation because honestly, this is a

real example and it is the best example

you could use to explain what the hell

Michael Bur is saying and it's very

simple. Okay, so here's how it works.

Let's say you buy a plane for $13

million

and you assume that that plane is going

to last you for 13 years. The plane in

this example is totally

indistinguishable from a chip. So think

imagine this is a blackwell chip. It's

just easier to imagine it as a plane. So

it's kind of like a car, right? Like

imagine you buy a car and it's kind of

like I'm going to assume the wheels are

going to start falling off by like year

9 or 10 or maybe 15 or whatever. I mean

hopefully 15 if we get those 15-year car

loans. [laughter]

You will owe nothing and be happy. But

anyway, so depreciate this plane over 13

years. How does that work? If a plane is

going to last 13 years or we think it's

going to last 13 years and that's

reasonable, then accountants will agree

with you. It's not fraud. Accountants

will agree with you and say, "Okay,

we're going to write it off over $1

million per year." That way it doesn't

look like you're like massively negative

one year. It just sort of butters out

the expense. That's all it does. That's

why a company like Meta isn't going to

100% write off their chips right now

because their earnings would absolutely

tank and people be like, "Oh, what the

hell happened here?" It'd be crazy.

You're not going to see that.

The problem is that these chips have not

actually gone down in value and neither

have current chips. They usually do,

right? Like here you have Google tensor

pro tensor processing unit demand is

outstripping supply claims 8-year

hardware iterations and have 100%

utilization. So in other words, runtime

of eight-year-old chips is at 100%.

Okay, so if runtime for 8-year-old chips

is 100%, that means they probably

haven't lost value. They certainly

haven't gone to zero, right? So, they've

had more value than we thought they

would. It's kind of like you go in

thinking, I think this car is going to

last me 8 years, but then you end up

driving it for 13 and you're like, wow,

I can't believe that thing keeps

kicking. This is fantastic. Right? Okay.

So, now I want you to think about the

plane example because this is huge.

This aircraft, and this is a real

example, has not actually gone down in

value at all. In fact, when I sold the

plane after 3 years of flying it and

putting on over 800 hours onto the

airframe, I sold the plane itself for a

profit. Obviously, that's different from

operating costs, fuel, pilots,

insurance, just the actual plane. The

chip, I sold the chip for a profit. This

would be like buying an Nvidia chip 3

years ago and selling it for a profit

today. Okay, like ordinarily we think

that's crazy. Like planes go down in

value, cars go down in value, chips go

down in value over time because we make

new ones, we make better ones. So it

should go down in value. Why did it go

up in value? Ah, okay. So this is

exactly why Michael Bur is currently

entirely wrong. We have way too much

demand for chips and these particular

aircraft and we have too little supply.

See, a company called Ember, whom I

started investing in back when they were

worth like 15 bucks. We talked about

them in our course member live streams.

We talked about them on the channel.

Heck, I invested in them with different

vehicles.

A great company. There's so much

freaking demand for these aircrafts that

Ember Air stock has like quadrupled. And

it makes sense because this aircraft is

the bestselling business jet in the

country. It became that I believe in

2023 when NetJets honestly just they buy

like every single one of them. They're

waitless for like years to get your

hands on a brand new one of these and if

you buy a resale one they sell for a

premium. The exact same thing is true

for chips. There is so much demand right

now and supply is so constrained because

we just don't have the manufacturing

ability. We got plenty of designers.

Amazon's a designer. Google's a

designer. Nvidia is a designer. AMD is a

designer. They're all designers. The

problem is we don't actually have the

manufacturing lines available to make

all these chips. Same thing with planes.

Okay? There are plenty of plane

designers. We just don't have that many

manufacturing lines to actually produce

as many planes as are required. So in

that case, if there is so much demand

for an old chip and you're actually

depreciating it, like in this example,

$1 million per year, let's say, but then

after 3 years, the value has actually

gone down zero. What's actually

happening is exactly the opposite of

Mike Kilbur's argument. Meta and Oracle

and Microsoft are actually overstating

their expenses, which means they're

actually understating their revenue

because these chips have not lost value.

The plane didn't lose value. So, if

you're writing it off, any dollar you're

writing it off means you're overstating

your expenses. So you could actually

make the argument today that Michael Bur

is entirely wrong, that the opposite is

true, that these chips or that plane

example have actually gone up in value

or at least stayed stable and therefore

they shouldn't be writing off anything.

Okay, how does this change? And this is

what I want you to know. So, if you've

made it 6 minutes into this video, 3

minutes on 2x, this is why you want to

subscribe to the channel because I want

you to understand that you could be a

bull on that fact right now. Okay, the

bull is, yo, right now they actually

might be overstating their expenses. It

literally takes Bur's argument, turns it

on its head. Bur is literally going,

>> "Bitch me."

>> And he's like, "Bro, how is that

possible? Why is that possible?" Because

that's why

[laughter]

>> right now that is a fact. Companies are

probably overstating their depreciation

because we are still utilizing these

little chips because supply is so low

and demand is so high. So what does this

rely on? Ah right. Let's go back to the

plane example for a moment. What happens

if, let's say, there's a recession and

all of a sudden demand slows for private

aviation or, I don't know, the FAA comes

in and says you're no longer allowed to

fly private jets or I don't know, Greta

Thornberg finally wins and people aren't

allowed to have private jets. Well, all

of a sudden, demand for these aircraft

is going to plummet. And as soon as that

demand falls below the level of supply

that is able to be produced, the value

of these aircraft will plummet. You'll

actually see this massive normalization

where all of a sudden when somebody buys

an aircraft, wow, it actually goes down

in value every year. It's odd for a car

or a plane like during COVID cars went

up in value, right? Weird, right? Used

cars were skyrocketing. Weird. not

normal. Demand really high, supply

really low. Okay, everybody wanted a new

car.

When that demand falls off a cliff

because of whatever some legal risk or

recession or whatever, then the value of

these aircraft will plummet very, very

quickly because why would you want an

old one when you can get a new one? The

same is true of chips. And then

companies might actually have to go, "Oh

man, you know, we we were able to use a

6 or 8year depreciation schedule on

these chips because old chips were still

being utilized at 100%. So it was okay

to change our accounting method." This

is Bur's argument here. You know,

instead of depreciating these over 3

years or four years, we'll go to six or

seven years or whatever. Six, seven.

that will change if demand falls off a

cliff for these and supply or and or it

could be a combination supply

skyrockets.

Okay, then companies will look and go

crap. If let's say for some reason this

plane or chip depreciates much more

rapidly, then we're going to have to go

back and we'll have to restate our

financials and we'll actually have to

write off more expenses more rapidly and

it'll actually look like in that case we

have overstated our earnings.

Aha. That is where Michael Bur's

argument comes in.

Michael Bur says by 2028,

Oracle will overstate its revenues by

26.9%, Meta by 20%, etc.

But it gets worse. More detail coming

November 25th. Now,

this is where you have to see to

understand this Bur argument. You have

to understand that it is entirely

possible for Bur to be wrong right now.

And it actually looks like he's wrong

right now. Just like if you said a plane

should go down in value, you would be

wrong right now.

The same is true though in the future

that we should normalize. Old trips,

chips should be worthless. Planes should

go down in value over time. So it is

possible that Michael Bur will be

exactly correct. The only way that he

will be exactly correct though is that

for some reason company like let's say

there's a market crash, companies start

getting punished for massive capex and

all of a sudden companies say you know

what we're just not going to buy

anymore. I mean remember what the next

phase is that we're seeing in artificial

intelligence. We're literally seeing

Microsoft start advertising co-pilot on

X or on Facebook. They're using

influencers to promote Copilot, which I

think Copilot's the worst, but whatever.

Everybody has their own opinion on

these. I'm not trying to offend anyone.

Like, I I've been a I was a GPT disciple

and then all of a sudden I discovered

Gemini. I usually use multiple at the

same time, but I discovered the latest

progress of Gemini has made and I'm like

I personally think they're leading right

now. But who knows, like tomorrow there

could be a co-pilot update and maybe

it's better. And I recognize they use a

lot of GPT uh as their backbone. But

point is like things change. And so

Michael Bur is stating something that

right now makes him look like a fool.

And this is why he's getting a lot of

hate. The reason for the call today,

John, is depreciation. And you get your

tunnel vision on the depreciation bomb.

Bur getting margin called on the short.

I'm sure the big four accounting firms

totally overlook that one. Classic

hammer syndrome. When you got a hammer,

everything looks like a nail. This this

guy right here 100% use GPT. I mean, who

says this? I get it. I once hit big on

an arbitrage and started seeing ARB

everywhere, but you don't step in the

same river twice. What? [laughter]

Dude, the guy's totally using GPT to

write his tweets. Whatever. And then you

got somebody here who's clearly not

using GPT. Just say you're bagged.

[laughter]

And it's like, dude, the internet is

just freaking hilarious. I can't reply

to any of these tweets here because if

you're blocked by Michael Bur,

[laughter] which I am, you can't reply

to any of the underlying, which I think

is so funny that Michael Bur has me

blocked, but it's it's because in my

opinion, uh I you know, I provide a

reasonable counterargument to his

arguments. That's not actually why I got

blocked. Uh to be fair, I got blocked

because I left an eggplant emoji as a

comment to one of his posts.

Been blocked ever since. [snorts] That

said, um, [laughter]

look, [gasps] I think the perspective

here is really, really important and I

think the plane analogy is really great

because like at some point within the

next decade, it seems reasonable to me,

especially if there's a recession, that

plane values will plummet. It doesn't

make sense that you could put 800 hours

on a plane and sell it for more. Like,

that sounds crazy. And it's exactly what

happened.

But again, that's because supply is so

constrained and demand is so high for

these aircraft. That will change whether

it's a recession, supply catching up, a

combination of those, some form of

normalization, that will change and

you're going to have to do a lot of

catch-up depreciating on some of these

assets. So, this is why Bur could be

right in the future, but he's dead wrong

right now. Uh and so I think knowing

that is useful because now you can

identify that inflection point when you

start seeing that change happening.

Supply coming up, demand starting to

taper down, that's your early warning

sign. And again, that's why you want to

make it to the end of my videos. So if

you found that helpful, subscribe and

we'll see you in the next one. I

appreciate y'all.

>> Why not advertise these things that you

told us here? I feel like nobody else

knows about this.

>> We'll we'll try a little advertising and

see how it goes.

>> Congratulations, man. And you [music]

have done so much. People love you.

People look up to you.

>> Kevin Pra there, financial analyst and

YouTuber. Meet Kevin. Always great to

get your take.