Tech stock markets crash on the AI bubble collapse. 🌊 Join our private group https://techleadpro.com Your Community for Crypto, Stocks, DeFi & Tech. Get the inside track on crypto and stock trading, DeFi yield strategies, passive income systems, and tech industry deep-dives you won’t find anywhere else.
The AI sector is facing a critical turning point as tech giants shift from high-margin software models to high-capex hardware bets. With negative free cash flow appearing at companies like Google and Tesla, combined with trillions in \"shadow debt\" and aggressive spending to avoid obsolescence, the investment profile of Big Tech has fundamentally changed, signaling a potential bubble.
Well, this could be it, guys.
I have been selling a bunch of my tech stocks.
For the first time, something very strange just happened.
Google just reported negative free cash flow for the first time since their 2004 IPO.
Negative $6 billion.
Tesla followed with another negative $1 billion in cash flow, all while last year these companies
had been printing cash.
This week has seen tech AI stocks collapse, Google down 8%, Tesla negative 19%, Meta negative
7%, Amazon down 6%, Oracle down 8%, Zendesk in the past month down nearly 35%.
And then in just a recent report, even more debt was uncovered.
The Nikkei dug into the actual financials of Google, Microsoft, Amazon, Meta, Oracle.
What they found was, in addition to the $1.35 trillion of debt reported on the books, there
was another $1.65 trillion of unreported debt.
This is known as shadow lending.
The result is this more than doubles the actual debt reported.
Meta alone is sitting on roughly $420 billion of additional debt, that is 3 times its visible
debt.
Now, behind some of this shift is an escalating war between the open-source and closed-source
AI model communities.
While the White House has accused the Chinese lab Moonshot AI of potentially stealing frontier
models that American companies had just happened to spend billions of dollars to build, they
use a technique known as distillation, where you train your model on the outputs of the
better ones and get most of the capabilities for almost free.
But here's the thing, NVIDIA's own Jensen Huang actually defended distillation.
Distillation, learning from AI, learning from other sources of knowledge, is fundamental
to intelligence.
We're constantly learning from other people.
All day long we're learning from each other.
And so AI has to learn from something.
And then 25 other American big tech companies told Washington to back off.
So let's get into what happened.
Let's start with the stocks.
These tech companies today are fundamentally different investments than they were just
over the past decade.
And you need to recognize this change.
So you see, Google used to be a low capex money printer.
Mostly software, search ads, cloud.
High margins.
In fact, it was less of a tech company and more of an advertising media company.
Great blue-chip investment.
A nice stock to round out any retirement portfolio.
Well that version is dead.
And capex has doubled to nearly $45 billion in just one quarter.
More than the entire business generated in operating cash.
The full year guidance for just this year raised to $200 billion of capex.
So they are spending insane amounts of money just to keep up.
They're not even beating the competition.
And the CFO said free cash flow will stay under pressure.
We continue to expect our capex to increase significantly in 2027.
And we'll provide more details at a later date.
Google did not even name a number.
But analysts are estimating even higher.
Maybe $250 billion of spending.
Even $300 billion next year.
So this is no longer just a software company.
It is a high capex hardware bet that they can face the likes of OpenAI, Anthropic, and
every Chinese lab out there.
And they're already losing despite all the capex Google admitted.
They are behind on coding.
There are areas where we've acknowledged we need to improve coding and agent decoding
is an example of that.
Coding should be their bread and butter.
But Anthropic is leading here.
And the next version of Gemini continues to be delayed.
The future continues to be just vaporware that we are betting on that this is somehow
going to win.
Now, sure, this bet may pay off or it may not.
It could go horribly wrong.
It is not the same investment profile.
And they have had more than a year to catch up.
And yet the likes of Anthropic and OpenAI remain in the lead.
Even the China open source models are outperforming Gemini in benchmarks.
And to make matters worse, Anthropic just released Cloud Opus 5, a new state-of-the-art
model near fable-level intelligence at lower costs, overshadowing Google's own models
released earlier this week, Gemini Flash 3.6.
But for Google, they must spend here.
They must compete in this arena because this threat is existential to their entire business.
If they do not compete, they lose search entirely.
Their back is against the wall here and they are forced into the spending.
And that is the quiet part.
Now Tesla.
The stock crashed nearly 15% just in one day.
Their free cash flow flipped negative for the first time in two years because CapEx
suddenly jumped over 140% to $5.8 billion.
Earnings per share came in at $0.33 against expectations at $0.55, nearly 40% below expectations.
Operating income down 57%.
Their margins at just 1.4%.
I mean that is grocery store margins.
And the large reason for this is Optimus, another AI hardware play.
Hardware just happens to be insanely CapEx heavy and insanely hard to get right.
Traditionally, VCs do not invest in hardware.
It's one of the hardest things to solve, to make a autonomous humanoid robot that can do tasks.
But you've probably seen lots of impressive demonstrations of robots on the internet.
But those demonstrations you're seeing are pre-programmed or remote controlled.
Meanwhile, the robotaxi fleet remains too unreliable and Musk is not willing to risk
the legal liability headlines of a self-driving Tesla running someone over, and for good reason.
But what a moonshot bet.
Speaking of moonshots, Washington is panicking over Moonshot AI's Kimi K3.
Trump's science advisor publicly accused Moonshot of large-scale industrial
distillation of Anthropic's latest frontier models.
The US Treasury followed with considerations to sanction these models.
But the rest of big tech is opposing this.
Why?
Because they would like nothing more actually than to see Chachi, BTN,
Claude lose their market position so that they too can get a piece of the action.
Microsoft itself just published the defense of open waste and distillation as essential.
For American leadership, they call distillation a legitimate technique.
And they warn that restricting open-source models risks driving the next wave of innovation
overseas.
NVIDIA co-signed, Dell co-signed, Meta, Palantir, IBM, CrowdStrike, A16Z, Y Combinator.
25 companies all co-signed.
Jensen Huang made his first post ever on X to promote this.
Well, guess who did not sign?
Some of the biggest vendors in AI.
OpenAI, Anthropic, Google, XAI.
These closed-source companies are being attacked on all sides, foreign and domestic.
And here's my take.
It may be a bit of both.
On the consumer side, it's likely we'll be using some closed models.
A few models will win out because it's just more convenient to have somebody hosting the
models and serving them up at speed, scaling it all out, upgrading the models, keeping
everything maintained.
And we just pay a small subscription fee and let them take care of the rest of it for us.
This is what happened with the mobile era, where we saw a duopoly of Apple and Google,
two mostly big winners who just dominated the whole sector.
But there were a bunch of other phones.
Now, on the other side, on the enterprise business side, it would probably be nice not
to have every company be at the mercy of OpenAI and Anthropic or whichever closed-source companies
win.
They should be able to run something open-source like Linux or whatever it is to protect their
data, their IP, to lower their costs, to continue innovating.
But in either case, 25 of the biggest companies in America are arguing that this moat should
not be protected because a commoditized frontier is good for their own business.
And that is the bubble.
Not that AI fails, but that it works, it gets cheaper, and it gets commoditized such that
the biggest spenders see their investments go to zero.
Now, let's take a look at the actual debt here.
Nikkei ran the analysis and found this shadow debt.
They hold off-balance debt, obligations of $1.65 trillion of additional debt in addition
to the $1.35 trillion they've already reported.
So what is this?
Data center leases that have not started yet, GPU supply contracts, accounting tricks on
future leases.
There's a second timing problem here as well.
These balance sheets are fudged.
So the way this works is that the hyperscalers are depreciating their hardware costs over
five to six years.
So for now, the chips look like assets on their books.
So this year's revenues look strong across the board precisely because the future expenses
have not hit that income statement yet.
But let's try waiting another year and the chips from this year already start looking
outdated.
They got to depreciate that maybe faster than they've been doing.
What Meta did was they just played with the books a bit, extended the so-called useful
life of its servers from last year.
And this one accounting trick added about $3 billion to their earnings.
Not from selling anything, just from changing a number.
And so here's a real picture.
What used to be positive cash flow and buybacks of company stocks are now negative cash flow,
dilution of stocks, hidden debt larger than the reported debt, hidden depreciation, circular
funding, insane capex spending, higher risk, and 168,000 tech layoffs just this year.
Not because business is bad, but to free up cash for this arms race.
What I found is the investment profile of these stocks is just totally different fundamentally.
It's not to say they're not going to work, but it is a higher risk bet.
And when's this all going to end?
Is there even an end in sight?
We're already going into negative cash flow.
We're already diluting the stock.
Is there an end to intelligence?
You can always train a bigger model.
You can always become smarter, but you cannot burn money forever.
Now, undoubtedly, AI is a transformational technology.
Most people will and should use it.
But the thing is, this is nowhere near profitable.
The spending is running nearly 10 times revenue.
Now the build-out is real and necessary, but you don't want to be in this game.
In my opinion, unless you're absolutely forced to be in this arena, such as like Google,
they have to play or they lose their entire business.
Personally, I don't know why SpaceX got themselves involved in this game,
but XAI has been a black hole money pit on their balance sheet.
Without this, Starlink would have been very profitable.
Starlink made $4.4 billion in profit.
The AI segment lost everything for them over $6.4 billion in losses.
As for Microsoft and Meta, I think what they saw was in the last mobile war,
Google and Apple won the platform wars as a duopoly.
And so they want to compete this time around.
The question is, could it cost them everything?
Unlike in the past, real hard tech CapEx is nothing like the media software SaaS businesses
that were low CapEx, high margins.
That was the old deal.
This new deal is concrete power contracts and chips that depreciate.
Now what happens next?
Well, we've seen this movie before.
In 2000, the dot-com boom was supposed to change everything, and it did.
But here's how that played out.
First, the company spent like there was no tomorrow and got themselves bankrupt and wiped out.
Internet 1.0.
Many of those companies don't even exist today.
And what happened next was capital actually fled into treasuries and consumer staples.
The dollar strengthened.
As it plays out, what may happen after that is recession, unemployment, job losses, a
slower economy.
And so the government tries to stimulate it by injecting additional cash, lowering rates,
and we may see liquidity then flooding the system.
Inflation could follow and then hard assets like gold, maybe Bitcoin could then run afterwards.
So I would say hedge your bets, hold dry powder in treasuries and high quality consumer staples.
Watch the dollar and gold for that liquidity response.
And for more market analysis and trading signals, check out techleadpro.com, our private trading
group.
We've been calling out some good trades there making passive income.
Actually, we released a few crypto trading arbitrage bots on there as well.
The code there has been backtested, confirmed to generate hundreds of dollars per day, really,
in passive income.
Simple strategy written in Python.
Check it out if you're interested.
Free money, really.
But that'll do it for this video.
Let me know your thoughts on the AI bubble in the comments below.
But the spending here is no longer under anyone's control.
The pursuit of intelligence is endless.
Your capital is not.
See you in the next one.