Imagine you had so much money that you

could give $2500 to every American, not

just adults, everyone, including kids,

including infants, all $330 million of

them. $2,500. Here you go. That means

10K for a family of four. That is how

much money Open AAI is now valued at.

Despite the fact that $70 billion just

vanished from this deal, the stock still

goes up. If you missed it or if you're

new to the channel, here's a quick recap

of events that bring us up to current

day. September 2025, OpenAI and Nvidia

announced a $100 billion quote landmark

partnership deploying 10 gawatts of

Nvidia systems. January 2026, we

reported on this channel that the deal

was on ice due to internal doubts at

Nvidia. And there's tons of juicy gossip

there, including Jensen saying some

things in private circles about the

stability and long-term prognosis of

OpenAI not looking so hot. Just this

month, Financial Times confirms a new

deal. Nvidia is still getting involved

with OpenAI, but not at that hundred

billion amount. They're actually just

getting involved only with $30 billion.

And the structure of the deal is now

completely different. But remember,

Jensen Hong was very careful to say

before the deal fell through, which is

what it did, that the original deal was

quote unquote nonbinding. That's per

CNBC. The implications of this are vast.

They ripple outwards. They affect

everything, even outside of AI. Let's

talk about where the shock wave goes.

First, the elephant in the room. This

was never a real deal. Nvidia was never

going to give a hund00 billion to

OpenAI. This was a press hype move. This

is a classical PR nut job. Let's pump up

the stock price so we can sell it high.

We can get more investors on board.

That's exactly what this is. This was

never going to be real. And if we sort

through the paperwork, you go back to

September 2025.

They do declare that they're doing this

partnership, but it's a letter of

intent, which seems solid. I mean, you

write a letter of intent. you you should

intend on doing something. But that's

our mistake for taking these companies

at their word. They're just manipulating

the market with a letter of intent. If

you follow the numbers, things get even

shadier. So, it's contingent on 10

gawatt of power, but that power supply

is not built yet, nor are there plans to

build it. But when it was announced, it

pushed Nvidia past a $5 trillion market

cap. And to just recap, OpenAI's total

announced deals, $1.4 $4 trillion total

in potential commitments, i.e. farts in

the wind. That's 250 billion from

Microsoft Azure, 38 billion from Amazon

AWS, Oracle kicking in 300 billion.

Coreweave 22 billion, AMD up to 300

billion. None of these are binding

purchase orders. They are letters of

intent, which we just saw in the case of

Nvidia don't mean anything to these

people. They've ratcheted it up the PR

game. Now, a letter of intent is just an

ad buy. The whole AI economy is built on

announced deals. Maybe, should we, in

two more weeks, we might, but it's not

based on signed contracts. And so,

here's what's likely happening, and I'm

reading the tea leaves, but this be

where my money is, is we do know, we do,

we do know is a fact that OpenAI was

dissatisfied with the inference speed of

Nvidia's chip. What inference is is if

you've interacted with chatgpt it's the

machine generating a response. So

inference is just the class of computing

you could bucket it under where uh an

LLM generates a response and actually

works with you. This is opposed to

training which is the other major use of

these chips. So as per Reuters, we know

that OpenAI was dissatisfied with the

performance of Nvidia GPUs uh

specifically on inference tasks. They're

serving chatbt to a record number of

people. The scalability concerns are off

the chart. And in a certain level, you

want to optimize a couple of things. And

so my thoughts are, and this is, you

know, coming from someone who's worked

with software and hardware in the

defense tech space, is it's always a

collaboration between software and

hardware. If you write very efficient

software, it doesn't have to use too

many system resources. And in fact, one

of the most tried and trueue interview

rounds for software engineers is to

write algorithms that maximize the

efficiency of the software. In essence,

that's why software engineers get paid

so much, especially at these big

corporations like Meta or Netflix, is if

they can shave 50 milliseconds of

compute time off of a function. That

sound that sounds like nothing. You

can't even 50 milliseconds is under

almost under a gap of what we can

perceive. But if you reduce 50

milliseconds across a bajillion

different instances of that function

call, then you've saved some serious

serious money. But at a certain point,

you might get constrained by your

hardware. This is a bit lofty. So let's

do an example to keep it concrete. Think

about a race car driver, a Formula 1

driver, let's say. And there's two

components of that equation in our

simplified system. There's the car and

there's the driver. And so if we

consider the car to be the hardware and

the driver to be the software, if we

have really bad software, like if you're

someone who's never driven a Formula 1

car, I've never driven a Formula 1 car.

Let's say that's a safe assumption. And

you just get put in a Formula 1 car.

Even if it's the best car in the world,

it won the world championship last year.

Like you're probably going to crash it

if you can even get it off the line and

not stall it out, right? because the

software is garbage even though the

hardware is great. Conversely, if you

take someone like Max Versappen, Formula

1 world champion, and you put him in a

Kia Soul, let's say, he's not going to

set a good lap time because he's in a

Kia Soul. He's constrained by the

hardware despite the fact that he's

maximizing all the juice out of that

hardware that there is to be gotten. So,

OpenAI gets to the point where they're

already paying AI researchers and

software engineers

millions of dollars. Millions of the

comp packages are insane. Insane.

They're work here for a year and cash

out after we go public and you don't

have to work again in your life type of

money. So, they have the best people on

it. They have maxed out the software on

the inference side. They cannot do

inference much more efficiently. They're

hardware limited. They need a better

car. They have the best. They have the

world champion racing drivers, but they

don't have the cars to allow them to go

as fast as they want to do. So, what

does OpenAI do? We're hardware

constrained. Nvidia is what we use for

inference. Is anybody else making

inference chips that are faster? And

confirmed by Reuters, OpenAI starts, you

know, looking around seeing there's some

competitors out here. We got AMD,

Cerebrris, and Grock. We covered that

acquisition by Nvidia all the way back

last uh last fall. And they start

thinking, you know, hey, maybe we need,

you know, to start on boarding some of

these other providers for inference. The

most damning thing, and I I'll translate

the slimy PR speak for you, is Altman

says on X, he tweets. He says, quote,

"We love working with Nvidia." The

translation of that into regular human

is, "Oh man, we really got into bed with

Nvidia. We're not sure about it, and now

it's really kneecapping us." Then you

have more slimy speak coming from Jensen

when he's in Taiwan and interviewed by

some reporters. And he calls those

dissatisfaction reports quote unquote

nonsense. He calls them nonsense while

slashing the deal by 70% though. So I

guess the actions speak a little louder

than words in that situation. But here's

where it gets really interesting and

maybe it's interesting to you, maybe

it's not. If you live through 1999,

maybe you're just like, "Ah, again, here

it goes." And it's the circular funding.

We've talked about this many times on

the channel, but a quick refresher.

Nvidia says, "Here's $30 billion to

OpenAI." The contingency being with

those $30 billion, you need to buy $30

billion worth of our chip. So that money

comes back to us. Nvidia's revenue goes

up, line go up, investor happy, investor

toss more money at Nvidia, and Nvidia

can invest more in AI companies. It's

the vehicle for this bubble, the nexus

of funding. There's shell games going on

everywhere. There's proxies. You also

have Cororeweave. This one's a little

tricky to get your head around. Nvidia

invests in Corewave. They're a big

backer. Coreweave buys Nvidia GPUs and

then they sell those Nvidia GPUs to

OpenAI. Microsoft though is triple

dipping. They they have this thing

figured out. They don't even like need

to make Windows anymore. Windows is just

like a a passion project for them at

this point because where they're making

the big bucks is they invest in OpenAI.

They then serve through Azure as

OpenAI's cloud compute provider. And

OpenAI's products also run on Microsoft

platforms. So that's licensing as well.

And if you live through 1999, this is

going to sound real familiar to what

Cisco did where Cisco backed telecom

providers. Telecom providers bought

Cisco hardware. Cisco said, "Hey, that's

not circular funding. That's actually

revenue." So, if they had a loan deal,

they provided vendor financing to

someone, they bought $20 billion worth

of Cisco equipment and then paid them

$20, even though Cisco gave them the $20

to buy the equipment in the first place

on some kind of a loan term. They

reported that as income, we use our

lemonade stand as the example. If I'm

running a lemonade stand on the corner

and you come by and you're like, "Yeah,

I mean, maybe I could go for some

lemonade." And I slip you five bucks.

And I'm like, "Pretend that I didn't do

this. Pretend that I didn't do this.

that I didn't just give you $5 and then

come to the stand and then buy $5 worth

of lemonade and then you did that and

then I told the IRS I made $5 this year.

That's exactly what these companies are

doing. And so when the music stopped for

Cisco in 2000, people found out, hey,

that's not really revenue. The company

lost 86% of its value. So multiple good

sources are saying that the ink on this

deal is already signed. It's just yet to

dry. Bunch of news outlets are reporting

on it over the past few days, but the

latest funding round would put OpenAI at

an $830 billion valuation for a a

company. A company, remember from the

beginning of our video, that's $2,500 to

every American. Not just adults, but

kids, infants. You're tossing a brick of

K. You probably can't toss like a brick

of $2,500.

I don't even know how much that is in

hundreds. I mean, it's at least like one

of those drug bricks, right? I mean, you

just think about an infant. You just

toss them $2,500. Like, it's probably

going to hurt them, but that's that's

10K for the family. That's 10K for the

family. That's a lot of money. That's

everyone. That's the amount of money

that OpenAI is valued at now. And some

crazy stats on that. That's a 65x

revenue rate. Remember the math we've

done on this before. They are making 4

cents on the dollar. OpenAI is not a

company that is profitable. And now it

is valued at $830 billion. A company

that is not profitable. That is does not

make money. They're not in the black,

they're in the red every year. If the

best positioned AI company needs a

massive cash influx like this, what does

that tell you about the industry at

large? That it's a bubble. And if we

extrapolate that to the future sentiment

on what this means for the valuation of

OpenAI and even the belief that these

other vendors and other investors have

in the success of the company, you have

their primary backer, Nvidia, cutting

their original amount that they

committed to down by 70%. That does not

signal confidence to me in the valuation

and long-term success of OpenAI or the

bubble at all, which again I've argued

has already popped. We're just waiting

for the financial ramifications at this

point. So, we'll watch the funding

round. Remember, this hasn't officially

closed yet, but is it actually going to

close at 100 billion? We don't know for

sure. We don't know for sure. That's

what the best guidance says right now.

That's what most of the sources are

saying is that we're closing at $100

billion. And the big move to watch,

we're going to watch this in Nvidia's

earnings on Feb 25. Remember, we'll drop

a video right after that as well, is

we're going to be real curious about

OpenAI's chip diversification strategy.

They have a deal with Cerebras. They

already demoed some tech. We talked

about that on the channel last week.

It's not there yet. It's not as good as

Nvidia's tech, but it's fast. It's It's

way way way faster. It's not as

accurate. They can't run as big of a

model on these chips, but it's way

faster. And that's that's not nothing,

especially if they can solve the the

accuracy of the models that run on those

chips, which I I assume they will

probably inside of 6 months. So, we got

to see what is OpenAI's future play

going to be on inference chips. We're

also going to keep a close eye on

terminology around this, especially

coming from the tech sponsored media.

And the term that we're starting to see

pop up is a quote unquote sober phase

for the market. Okay? And a quote

unquote sober phase. I'm sure these tech

sponsored outlets would define it

differently, but 24/7 Wall Street called

it quote moving from exuberant promises

into a more sober long-term growth

phase, which is optimists speak for the

bubble already popped and we're just

trying to slow drip the news so we don't

tank the economy. And the argument and

positioning of this channel, of myself,

is that AI is is transformative. It is a

landmark technological achievement that

is already made a lot of things easier.

It's going to change to a degree how we

work, but it is not the replacement drop

in replacement for a human technology,

the godlike intelligence that these

sages of Silicon Valley have led us to

believe. It's a better automation. And

for what it's worth, that's very

helpful. I mean, it's helps us do stuff

quicker, easier. Hopefully we have to do

less menial tasks because of this as it

gets more embedded into everything. But

it's not a drop-in replacement for a

human. That's the human. That's

ludicrous. Absolutely ludicrous. But

it's what these salesmen are pedalling

in the valley. But the prediction and

the stance of this channel is that the

thing that is unstable is the finance

and infrastructure behind this. It is

unmaintainable. The numbers are off.

It's unprecedented in every way. And the

closest thing that we have to this is

the 1999.com bubble where the

proportions are way worse than they are

for what it's worth than the 1999

bubble. But it's the closest analog that

we have to understanding what the market

is going through right now. So phase is

really just code word for we're going to

need to really PR this bubble pop

financially or else every other sector

is going to panic and we're going to be

in real trouble in the market. They

don't want to say they don't want to say

that. They don't want to spook people.

And remember, just like in the.com

bubble, the internet wasn't fake. You're

watching this video because of the

internet, because of tech that was

developed and started in the dot era.

But the finances were off. The finances

were incorrect, but the tech was

transformative. If you're not already

subscribed, let's fix that. Click the

subscribe button below the bell to be

notified. If you'd like to support the

channel, I've had a lot of people

reaching out in the comments about that.

Become a member of the channel. You can

sponsor me. I would appreciate that. All

of the money goes back into producing

this content. And if you really want to

dig into the facts and figures on this,

sign up for the newsletter down there in

the description. I send out a quarterly

analysis of the numbers behind the AI

finance bubble. No PR spin or messing

about. It's just the straight numbers.

If you want to dig into that, subscribe

to the newsletter. Thank you for

watching.