Hang tight while we fetch the video data and transcripts. This only takes a moment.
Connecting to YouTube player…
Fetching transcript data…
We’ll display the transcript, summary, and all view options as soon as everything loads.
Next steps
Loading transcript tools…
Big Tech Is Faking Revenue | Sasha Yanshin | YouTubeToText
YouTube Transcript: Big Tech Is Faking Revenue
Skip watching entire videos - get the full transcript, search for keywords, and copy with one click.
Share:
Video Transcript
Hey guys, it's Sasha. This week, OpenAI
announced that they will pay AMD about
$100 billion for 5 million graphics
cards. At the same time, AMD announced
that they are giving $100 billion back
to Open AI in the form of stock. A few
days before that, OpenAI also announced
that they will be buying $200 billion
worth of graphics cards from Nvidia. And
at the same time, Nvidia is going to
invest $und00 billion back into OpenAI.
The problem is Nvidia doesn't have a
hundred billion to give to OpenAI. But
that's where the third deal comes in
because OpenAI is going to give Oracle
$300 billion to use their data centers.
Oracle is going to take the $300 billion
and walk it across to Nvidia to buy
approximately three tons of
graphics cards to shove into their data
centers so that those data centers can
actually be used by OpenAI. And then
Nvidia is going to take this money that
they just received from Oracle and use
that money to invest in Open AI. Did you
get all that? No? Well, don't worry
because in this video I am going to
explain a completely unrelated concept
called roundtpping. Roundtpping is a
common type of fraud where companies
just pass each other the same money in a
circle to mutually inflate their
valuations by pretending that they are
doing more business than they actually
are. It's like if there were three
people standing in the room, right? And
each one of them had a business and all
of these businesses were brand new.
They've never made any revenue. Then
business A goes and gives business B
$100 and says, "Thank you. It's a
pleasure doing business with you." Then
business B takes that same $100 and goes
and pays it to business C for, I don't
know, some services. And then business C
takes that same $100, goes back to
business A and pays that $100 back. So
now each business has $100 worth of
revenue. OpenAI's current valuation is
50 times their annual revenue. So that
would make each of these three companies
that just passed around a $100 bill
worth $5,000. And you can raise money at
that $5,000 valuation and collect a
whole load more than that $100 from
investors. That means that when they go
to investors to collect the latest
funding round, their valuation will be $5,000
$5,000
and they can then use all the money
raised to pay themselves really fat
salaries, have a pretty good time. Now,
roundtpping is a practice that is
generally considered illegal in the
United States and many other countries
around the world because it artificially
manipulates the market. It artificially
pumps the stock. It's a form of
securities fraud. So clearly, just so
there is no misunderstanding, I could
not possibly insinuate or suggest that
the tech bros at all of these companies
would be doing something that is illegal
to enrich themselves. That is 100%
definitely not what's happening here.
But while we are looking at these
companies, I thought I would walk
through exactly what has happened in the
last few weeks and draw it on a handy
whiteboard so that you can see for
yourself how this AI bubble is in no way
whatsoever a really obvious example of
roundtpping that is happening right in
front of us in the open because tech
bros get to pick and choose which laws
apply to them. you know, laws like
securities, fraud, copyright protection,
those only apply to schmucks like you
and me, not the cool kids wearing
hoodies in Silicon Valley. So, let's
start at the beginning. The story really
starts when a bunch of investors along
with Microsoft find 58 billion down the
back of the sofa and give it to OpenAI
in various funding rounds. Immediately
after all of that happens, OpenAI takes
those tens of billions of dollars and
gives them back to Microsoft to use the
Microsoft Azure data centers. Microsoft
takes the 10 billions of dollars that
they've just received from OpenAI after
they gave it to OpenAI in the first
place and goes and gives them to Nvidia
to buy graphics cards to put in their
Azou data center so that OpenAI can then
train their models using those graphics
cards in the Microsoft data centers. But
Microsoft are not the only ones buying
graphics cards from Nvidia. Google,
Meta, XAI, and Amazon are all standing
in an orderly queue buying the same
graphics cards that are now being sold
for four to six times the price that
they used to be sold at before ChatGpt
arrived in November 2022. All of these
guys, all of these companies also have
their own AI work. None of it actually
makes any money, but it sounds really
cool. It's pretty futuristic, right?
When you tell girls that you work in AI
on a Saturday night out. Just kidding.
Of course, if you are a tech bro working
in AI, you don't have time to go out on
a Saturday. You're too busy coding to
conquer the world. Anyway, then along
comes Anthropic, another AI startup. And
they also collect a load of money from
investors. And the biggest one for them
is Amazon, who give them billions of
dollars to Anthropic. And then Enthropic
takes those billions of dollars and
gives it right back to Amazon to use the
Amazon Web Services to train their
models just like what Microsoft did.
It's exactly the same but just with a
different company. XAI is a bit of a
weird one. They're building their own
data centers and XAI was originally
founded by Elon Musk. And to finance
XAI, Elon Musk sold a load of Tesla
shares which at the time didn't do
wonders for Tesla stock. He used a chunk
of that money from sending the Tesla
shares to buy Twitter and then a
separate chunk to start his AI company.
Of course, XAI should have never existed
in the first place because Tesla is a
publicly traded company that also
operates in AI. In fact, according to
Tesla, 100% of their valuation should be
based on AI, not the cars that they're
making. And there is a clear conflict of
interest because Elon Musk is the
executive director of both companies,
the essential person in charge of both
companies and he has basically siphoned
out money, talent and IP to build a
startup that directly competes with
Tesla. Now at the time when this
happened, the official line that was
pared by all the fanboys of course was
that this is completely not the case.
XAI is all about large language models
and Tesla is all about full self-driving
and robotics and these are not the same
thing at all. They attract completely
different people blah blah blah. So the
work that Tesla does is apparently
completely separate in no way related
would not benefit from having this
basically the same line of work using
the same GPUs for the same sort of
thing. It would not be beneficial to
have that within Tesla, which is
precisely why now that XAI has merged
with Twitter and pretty much run out of
money because they've been burning so
much cash. They are trying to get Tesla
shareholders to invest in XAI. It is
proposal 7 at the Tesla annual
shareholder meeting regarding board
authorization of an investment in XAI
and a direct investment in XAI would
strengthen Tesla's access to cuttingedge
AI technologies, talent, and
intellectual property. you know, the
things that Tesla would have access to
anyway if El Fela Musk did not siphon
that entire business line out of Tesla.
But the Tesla fan club on Twitter is
wildly celebrating this opportunity,
this proposal to hand over loads of
money to XAI because they, I think,
suffer from Stockholm syndrome. But I
digress, a bit of a tangent. XAI also
just collected some funding from, drum
roll, Nvidia. And that lets XAI take the
money that they just received from
Nvidia to buy more Nvidia graphics cards
for their Colossus 2 data center that
they are building in Memphis. So the
result of all this activity is that
every company on this whiteboard has
seen an absolute explosion in money
moving in and out of their bank accounts
and cash flow because you see money
moving between them in a sort of round
trip kind of way. And the biggest
benefactor of this has been Nvidia whose
share price has gone up over 1,600%
since Chad GPT arrived less than 3 years
ago. So now that Nvidia is rich, they
are printing money. We have got this new
wave of announcements coming. These ones
are bigger. Nvidia is going to invest
$100 billion into Open AI. They don't
have even remotely close to hundred
billion they could invest, but that does
not matter because OpenAI is going to
apparently take that hundred billion and
use it to buy $200 billion worth of
Nvidia's chips. Open AAI recently had to
raise $6.6 billion because the company
is currently burning money. They're
burning over a billion dollars per month
before accounting for capex. With capex,
I don't know, but presumably they're
probably spending a few billion dollars
a month. I don't know. So, the company
that is currently absolutely
incinerating money, losing a ton of
money every month and keeps doing
funding rounds for relatively small
amounts to keep them going, is going to
buy $200 billion worth of chips from
Nvidia with the $100 billion worth of
investment that they're getting from
Nvidia. You following this? This makes
complete sense. And not all of the money
that they are planning to spend is
coming from Nvidia. There is more.
Remember, they also said that they will
pay $300 billion to Oracle to use Oracle
data centers and Oracle is going to pay
those hundreds of billion dollars to
Nvidia. And Nvidia is going to collect
those hundreds of billion dollars from
Oracle and use that money to invest more
into OpenAI. And then remember that we
also have the AMD deal that just came
out. Open AAI will buy $100 billion
worth. It doesn't seem that he can go
less than hundred billion now. That's
the minimum. That's the minimum of a
good AI announcement. They're going to
buy a hundred billion worth of AMD
graphics cards alongside the Nvidia
graphics cards. But because they don't
have the money to do it, AMD is giving
OpenAI 10% of the total amount of AMD
stock because then OpenAI can leverage
that stock and benefit from the stock
going to the moon and borrow money
against it to pay for the graphics
cards. You know, you're in a very
special kind of bubble when the biggest
companies in the space are going heavy
leverage not just on their own stock,
which is pretty crazy in itself, but
also on the stock of other companies
that they are cooperating with in the
same space. And it's not just OpenAI.
This is how these deals are getting
done. XAI's current funding round
includes $7.5 billion in equity and up
to 12.5 billion in debt through a
specialurpose vehicle. The chips the
company purchases will be leased to XAI
for 5 years allowing Wall Street
financeers to recoup their investment.
So instead of buying the graphics cards
using debt, they are letting investors
buy the graphics cards because that's
off the books and then lease those cards
to the company for 5 years from the
investors, which is probably roughly the
useful life of these graphics cards
anyway, running at 100% capacity for the
entire duration. This is by no means a
complete picture of what's been
happening. There are other players like
Coreweave, a data center company that is
now doing deals with both Nvidia and
OpenAI for tens of billions of dollars,
money moving back and forth. And there
are dozens and dozens of other players
in this space. You can keep writing
stuff on this whiteboard forever. And
the weird thing is that OpenAI's current
revenue is running at around $10 billion
a year. Anthropic is at about half that
on about $5 billion a year. So that's
$15 billion per year from the two big
ones. And the rate of growth is now
slowing to below 100%. It's on course to
be below 100% relatively soon. The
biggest growth at the moment is coming
from businesses, from companies who are
paying for those commercial licenses for
their developers, for staff. They're not
really checking exactly how many
licenses. They just like dish them out
to everybody and staff are all using AI
to do whatever work they're doing. Those
are just the revenues though because
remember there are no profits because
all of these companies are losing money
on all of this revenue instead of making
money. The more revenue that they take,
the more money that they're going to
lose. But at the same time, we are
seeing these endless announcements every
week about hundreds of billions of
dollars moving between these companies.
Dollars that don't actually exist, will
not actually exist, will never exist.
All of this movement is just on paper.
It's fugazi. Fugazi. You know that I
give you $100 billion. You give it to
that guy over there. That guy is going
to give it to that guy over there. That
guy's going to come back and give it
back to me. We've all made $und00
billion. The valuations of these
companies are now astronomically high
because there is an expectation that in
the long run they will figure out how to
make money. Maybe it will be through
ads. I know you you're looking forward
to that. I bet the issue is that the
moment that ads enter these chat bots,
you're immediately going to lose a giant
chunk of the appeal, a giant chunk of
trust for using LLMs in the first place.
Imagine the ethical dilemma of
recommending products or services or
whatever for which OpenAI gets paid
from the ad revenue from the company's
advertising instead of maybe
recommending other better products that
do not happen to advertise and chat GBT.
At the moment, there is no business
model in this space that can actually
feasibly run at a profit in any short
medium term because the physical cost of
producing the answers is higher than
what people are currently prepared to
pay for those answers. And I just don't
see a world where people are prepared to
pay 10 times or 100 times what they're
currently paying. Over time, this cost
is not decreasing. It is increasing
because of LLM's becoming more expensive
to run and because users are increasing
the usage per dollar that they spend and
the expectation of how much usage they
get keeps going up. So tech companies
have come up with this new solution.
Just keep passing the money around in
everinccreasing amounts. First, you
know, invest 5 billion, you give me back
the 5 billion. Now it's up to a hundred
or 200 or even $300 billion per
announcement. I guess right now we're
just seeing what happens. If valuations
keep going up and the SEC keeps keep
sitting around just holding their
private parts in their hands doing
nothing, maybe these companies will
start literally announcing trillions of
dollars of investment into each other. I
don't know why not. We're only three
eggs from there. Today the world has
never in history seen an example of
roundtpping even to a fraction of the
degree that's happening right now
between these companies. There is almost
no precedent for things like companies
donating 10% of their own publicly
traded shares worth about $30 billion
because they know that the moment that
that happens their share price is going
to go up so much that they will
immediately get tens of billions of
dollars worth of more value in the
business and in the long run it'll
probably be 100 billion or more because
of the hype that is being generated.
There are only two ways that this can
unravel. Either these companies figure
out a way to quickly earn about 20 times
what they currently earn in revenue and
at the same time figure out how to turn
a minus 130% margin into a plus 20 or
30% margin or the house of cards will
eventually come crashing down when the
infinite money glitch finally runs out
of infinite money. Think I am going to
send out a newsletter with some more
thoughts on this. If you haven't signed
up, the link in the description, it's
free. go and sign up and I'll send you a newsletter.
Click on any text or timestamp to jump to that moment in the video
Share:
Most transcripts ready in under 5 seconds
One-Click Copy125+ LanguagesSearch ContentJump to Timestamps
Paste YouTube URL
Enter any YouTube video link to get the full transcript
Transcript Extraction Form
Most transcripts ready in under 5 seconds
Get Our Chrome Extension
Get transcripts instantly without leaving YouTube. Install our Chrome extension for one-click access to any video's transcript directly on the watch page.
Chrome Extension