HomeVideos

NVIDIA's Biggest Customers Just Lost $1 Trillion in One Week

Now Playing

NVIDIA's Biggest Customers Just Lost $1 Trillion in One Week

Transcript

281 segments

0:00

$1 trillion in market cap gone.

0:05

That's how much Nvidia's four biggest

0:08

investors lost over the past week. Not

0:10

because their products failed, not

0:12

because of a recession, but these four

0:15

companies each announced how much money

0:16

they were planning to light on fire in

0:19

this insane crusade for more AI

0:21

infrastructure and AI chips. Wall Street

0:23

saw that in the earnings report and they

0:25

said, "Doesn't sound like a good play to

0:27

us." Let's break it down. Over the past

0:29

two weeks, these four companies

0:31

announced that they'd be spending about

0:33

690 billion in capex in 2026 for AI

0:36

infrastructure and chips. Let me say

0:38

that number again. 690 billion spend in

0:42

2026 for AI chips in one year. And if

0:46

you've lost track where almost $700

0:48

billion is going, let's break down a

0:50

highle overview for you here. So

0:52

Microsoft announced that they're going

0:54

to be spending 120 billion or more in

0:56

fiscal 2026 on AI infrastructure. They

0:59

dropped $ 37.5 billion in just a single

1:02

quarter as per their last earnings. And

1:04

they still have an $80 billion backlog

1:06

of unfulfilled Azure orders because they

1:09

don't have enough money to turn the

1:11

power on. Alphabet doubled down. I just

1:13

did a video on that a couple of days

1:15

ago. They're announcing guidance for

1:17

$180 billion in spend this year. And

1:20

when I say double down, I mean that

1:22

quite literally. That's about double

1:23

what they spent on AI infrastructure

1:25

last year. So this is quite unusual. It

1:27

is like they are preparing for a war,

1:30

not a product launch. Amazon wins the

1:32

trophy, though. $200 billion in spend

1:35

announced this is robotics, AI,

1:37

infrastructure, data centers, even

1:39

satellites. That's 50% year-over-year

1:42

growth from what they spent last year

1:45

and 200 billion. I mean, it's the it's

1:47

larger than the GDP of a lot of

1:49

countries. Meta announces between 115

1:51

and 135 billion. They're building a one

1:54

gigawatt factory in Ohio somewhere. I

1:57

don't even know what a gigawatt is.

1:59

That's like back to the future numbers.

2:00

Zuck is calling it Meta Super

2:02

Intelligence Labs. We'll throw a fifth

2:04

one on there uh just as an addin. It's

2:06

kind of like your your baker's dozen.

2:08

You get an extra in the bag. Oracle

2:10

chipping in 50 billion, you know, chump

2:12

change. So that's around 700 billion

2:16

from all of these companies, which

2:18

again, this is like an imaginary amount

2:20

of money. Like I have no idea like if

2:22

you had $700 billion in $100 bills, like

2:25

how much space would that even take up?

2:27

It's an incomprehensible amount of

2:28

money. So you think you're thinking like

2:31

maybe this is just pretty typical uh

2:32

capex spend for a business of this size

2:35

involved in the AI war. And it's not. If

2:37

we add them all up and then compare what

2:39

they all did last year, it's about

2:40

double. When you come out with your

2:42

earnings report for the quarter or for

2:44

an entire fiscal year like these last

2:47

two weeks entailed for those companies,

2:48

it's a direct report on how the business

2:51

is doing. But of course, you're going to

2:53

try to skew those numbers a little bit

2:55

because essentially what you're doing is

2:56

you're trying to portray, you know,

2:58

confidence and like this is a good

3:00

investment for the big whales on Wall

3:03

Street to invest in your company or if

3:06

they're already invested to not yank

3:07

their money out of your company. When

3:09

these earnings came out for each of the

3:12

companies, their valuation, their stock

3:14

price dipped. People sold. Wall Street

3:17

was not thrilled. And it's not like over

3:19

the next couple of days. I mean like

3:21

immediately as soon as earnings end, you

3:24

see like a big sell-off and dip of these

3:26

stocks. To get you some specifics about

3:27

that, Alphabet dropped about 9% after

3:30

the announcement. We covered that here

3:31

on the channel just a few hours later.

3:33

And there's been some noise, but they

3:34

haven't recovered. They're still down

3:36

about 21% from their recent high. I'd be

3:39

calling that bare market territory for a

3:41

$2 trillion company. Microsoft down

3:43

about 25% from its recent high. We

3:45

covered that on the channel right after

3:46

it happened. Azure Growth missed

3:48

expectations on Jan 29th and the stock

3:51

has been in basically a freef fall since

3:53

then. And quick shout, if you're missing

3:55

these videos and you're not subscribed,

3:56

let let's fix that. Just click the

3:58

subscribe button right now. You'll be

3:59

notified when these videos and earnings

4:00

reports come out. Meta is about 2 to 3%

4:03

away from bearer territory. They're

4:05

really skirting the line there. And

4:07

that's a crazy one because they actually

4:09

beat earnings, right? But there was

4:11

still a decrease in their stock price

4:14

because the amount that they said they

4:15

were going to set on fire for AI

4:17

development scared people. Now, Alphabet

4:20

fared the best out of these. They're

4:21

only down about 9% and my hypothesis is

4:25

that is because they are not paying the

4:27

Nvidia tax. They are not in a lifetime

4:30

of slavery with Nvidia. They are making

4:34

a lot of their own chips as you know. So

4:36

their TPUs, their own proprietary chips

4:38

have given them a little bit of

4:39

insulation from the Nvidia black hole

4:42

that's pulling all of these companies

4:43

in. Again, combined that's about a

4:45

trillion dollar in market cap

4:47

evaporating. Remember that's not because

4:50

they announced, hey, we missed our

4:52

earnings and it was bad. They beat

4:54

earnings. There was a good prognosis.

4:55

This is because they said, "We're going

4:57

to spend an insane amount of money on AI

4:59

infrastructure. We're taking a huge bet

5:01

on this. All chips on red." And the

5:03

market is saying, "Where is my money? I

5:06

want to be paid. And I don't believe

5:08

that this AI gambit is going to pay off

5:11

for you." Jensen is laughing all the way

5:14

to the bank, though. A huge chunk of

5:16

every dollar lost for all of these

5:18

companies is flowing straight into

5:20

Nvidia's pockets. They posted 57 billion

5:23

in revenue in a single quarter. Their

5:25

data center division alone did like 50

5:28

billion. And that growth is nutty. I've

5:30

never seen anything like it other than

5:32

in a very small startup where the

5:34

economics worked out. They're up 66%

5:36

year-over-year and 25% quarter

5:38

over-arter. They're guiding 65 billion

5:41

for Q4, which they're going to report on

5:43

February 25. We'll have the analysis and

5:45

coverage of that on this channel. Full

5:47

fiscal year consensus for 2026. Make

5:50

sure I'm getting this number right.

5:51

213.3 billion in revenue. That is crazy.

5:56

At the end of the day, it doesn't matter

5:58

to Nvidia. It doesn't matter to Jensen

6:00

if any of these AI companies ever make a

6:03

dime. He sells chips. They sell GPUs.

6:07

They sell compute infrastructure. As

6:09

long as people keep buying their compute

6:11

infrastructure, it keeps the business

6:13

going. It keeps the business profitable.

6:16

If I sell arteasonal coffee mugs and I'm

6:19

selling you this coffee mug for this is

6:21

$5,000, right? And you're missing rent,

6:24

you're earning minimum wage, you can't

6:25

you don't have $5,000. So, you get it on

6:27

CLA or whatever the hell people are

6:29

using for a buy now pay later thing and

6:31

you split the payments up. That's going

6:33

to cause you a lot of hardship to make

6:35

those payments. But at the end of the

6:36

day, I I don't care. As long as you're

6:38

solvent and you pay me, I get the money.

6:41

Your financial situation doesn't matter

6:43

to me. That's exactly the situation that

6:45

Nvidia is in is OpenAI anthropic. How

6:48

are they going to profit from these? Are

6:49

they going to do ads? Are they not going

6:50

to do ads? That's important stuff. But

6:53

to Nvidia doesn't matter because they're

6:55

going to keep buying chips. That's the

6:56

only thing that matters. They sell

6:58

chips. They're in the chip sales

6:59

business. They're not in the AI business

7:01

and stock is doing good. So, let that

7:03

sit in your head that the market

7:04

actually thinks that Nvidia is doing

7:06

well and that this is sustainable. I

7:08

guess we'll find out if there's a big

7:09

upset after those Feb 25 earnings, but

7:11

right now it seems like smooth sailing

7:14

for Nvidia. And you might say, Dr. J,

7:15

but wait, I'm worried. I'm worried for

7:17

Sam Alman, the humble, noble, visionary

7:21

tech leader who's never had a real job

7:22

in his life and is not a software

7:24

engineer. What about him, Dr. J? And I

7:28

got good news. I mean, Sam Sammy boy is

7:31

going to be making a little money off of

7:33

this as well. And after some brutal

7:36

difficult math was able to come up with

7:38

a figure of about four cents of revenue

7:40

on every dollar spent on AI. That's

7:42

about how much profit is actually being

7:45

generated. It's four four cents on the

7:47

dollar. Okay. Even dinosaur

7:49

organizations like Fidelity are finally

7:52

starting to catch on to, hey, this might

7:55

be a bubble. We may have gotten ahead of

7:57

ourselves a little bit here. So, let's

7:59

take their five indicators of, hey, this

8:01

is a bubble, and see if they apply to

8:03

everything we've just said in this

8:05

video. Elevated valuations, check. Rapid

8:08

spending increases, check. Remember,

8:11

that's double growth and AI capex this

8:13

year. Circular financing. Open AAI uses

8:16

Nvidia chips. Nvidia gives OpenAI money,

8:19

but it's not money. It's money to buy

8:21

the chips. Check. Debt reliance, check.

8:23

If you saw my video on SoftBank the

8:25

other day, it looks like they're going

8:26

to be leading the next funding round for

8:28

OpenAI. Check. In lagging monetization,

8:31

check. They don't even know by their own

8:33

admission in the Super Bowl commercials

8:34

how the hell they're going to make money

8:37

off of this. So, that's five for five.

8:39

And that's Fidelity talking, not some

8:41

random bearded YouTuber. I mean, that's

8:43

circular funding. It It sounds like a

8:45

fun merrygoround to be on. I wish I was

8:47

on it. I'm not on a fun carnival ride

8:50

right now, but to be on that

8:51

merrygoround, you know, they're tossing

8:53

money at you. They got Mr. Alman over

8:55

here. You know, you're a celebrity. You

8:57

got tons of money. If this thing goes

9:00

belly up, you're you're going to be

9:02

fine. You're going to be fine. You got

9:03

plenty of money. You were on the

9:04

merrygoround. The trick is hopping off

9:06

before the music stops. And when is that

9:07

going to happen? I don't know. That's

9:08

the million-dollar question. If I could

9:10

answer that, I probably wouldn't be

9:11

making YouTube videos. I'd probably be

9:13

retired on some island somewhere. But

9:15

nobody knows. Uh, all I know is that the

9:17

fundamentals are off. The market has

9:19

become emotional and speculative. It

9:21

meets all criteria of any bubble that's

9:24

ever happened. And it's not a matter of

9:26

if the collapse is going to happen

9:27

around this. It's a matter of when.

9:29

Michael Green of Simplify said it best.

9:31

Quote, "Artational market would price

9:33

these risks." Unfortunately, we're not

9:35

in a rational market. Again, we're going

9:37

to be hitting a full analysis of

9:39

Nvidia's Feb 25 earnings. That should be

9:41

the best information we've gotten on

9:43

this in terms of raw numbers in quite a

9:46

while. So, I'm really stoked for that

9:48

coverage. I hope you're here for it. And

9:49

if you haven't signed up already, sign

9:51

up for the newsletter. You'll get my

9:52

free AI spending tracker that's released

9:54

quarterly. It has all of the numbers and

9:56

metrics about what we're talking about

9:58

in here in paper so you can sit down and

10:00

form your own conclusions. Thank you so

10:02

much for watching.

Interactive Summary

Four major tech investors of Nvidia recently lost $1 trillion in market cap, not due to product failure or recession, but because they announced colossal spending on AI infrastructure and chips. These companies plan to invest around $690 billion in capex for AI in 2026, with Microsoft earmarking $120 billion+, Alphabet $180 billion (double last year), Amazon $200 billion (50% YoY growth), and Meta $115-135 billion. Wall Street reacted negatively, causing significant stock dips for these companies, despite some beating earnings, indicating skepticism about the return on investment for this 'AI gambit'. Conversely, Nvidia is thriving, posting $57 billion in revenue in a single quarter, as its business model relies on selling chips and compute infrastructure, independent of its customers' eventual AI monetization. Concerns are mounting about an AI bubble, with Fidelity's five indicators—elevated valuations, rapid spending increases, circular financing, debt reliance, and lagging monetization—all present in the current AI market. The speaker concludes that the market is emotional and speculative, fulfilling all bubble criteria, suggesting a collapse is a matter of 'when', not 'if'.

Suggested questions

6 ready-made prompts