Liquidity pressures are building beneath the surface of today’s AI boom. As the Fed adjusts policy and private credit funds absorb more data center debt, capital flows are tightening in ways many investors may not see. In this episode, Michael Lee, Founder of Michael Lee Strategy, joins us to break down the difference between a liquidity squeeze and a solvency crisis, what the recent selloff in SaaS really signals, and how AI infrastructure, energy constraints, and rate policy could influence markets in the months ahead.
03:27 – SaaSpocalypse: Hedge Funds Sell SaaS 03:27 – Private Credit & Data Crunch 08:29 – Will AI Replace SaaS? 10:56 – AI Bubble or Data Bottleneck 14:08 – Falling Inflation & Fed Signals 17:00 – Private Equity Risks: Liquidity vs Solvency 19:46 – U.S. Outlook: AI & Industry 22:18 – Energy Boom and Nuclear Shift in the U.S. 23:17 – Investing in AI Infrastructure ----- Let's Stay in Contact: • X (Twitter) - https://twitter.com/MktInsiderShow • Linkedin - https://www.linkedin.com/company/market-insider-show/ ----- The views expressed in this video are those of the guest and do not necessarily reflect the views of Market Insider. This content is for educational purposes only and should not be considered financial, investment, or legal advice. Always consult a qualified professional before making financial decisions. Some links in the description may be affiliate links.
AI Generated·May contain errors
Are we facing another credit crunch like 2008?
What's happening in the credit markets right now?
Our guest today is Michael Lee,
who's a 20 plus year veteran of Wall Street.
He's here to tell us what he's seeing in the credit markets.
He's going to give a deep explanation
of how these things work.
There's a tremendous demand on the capital markets
for money for data centers right now.
And given what's going on with software as a service,
you know, some of the key players
are going to have difficulty participating in it
and the sheer size.
So you're talking about a trillion dollars in debt
in a very short period of time.
So you're entering a period of what seems to be
a credit crunch a la 2008.
I'm Siamak, welcome to Market Insider.
Mike, it's great to have you on, welcome.
Yeah, thank you so much for having me,
really excited to be here.
We want to talk to you about liquidity issues.
The Fed just injected some liquidity into the markets.
Can you tell us what's going on
with global liquidity right now?
The Federal Reserve in the beginning of December
went from quantitative tightening to quantitative easing.
And so after the great financial crisis in 2008,
the Fed decided to monetize our debt
roughly 18, 24 months after 2008
with quantitative easing one,
and then Operation Twist and two and three.
And, you know, I lost track of how many we're on right now,
but you took a Federal Reserve balance sheet
that was several hundred billion, maybe seven or 800,
and it got as high as maybe,
it got north of 10 trillion.
Since 2021, the Fed has been letting bonds,
they've been actively selling bonds
and letting them run off their balance sheet.
Well, they stopped doing that
in where they're going to start buying $20 billion
with the bonds of the roughly $7 trillion balance sheet,
as well as using the dividends, the interest payments
and the principal maturities, reinvesting those.
So that balance sheet is going to grow over time
to 20 to $30 billion a month.
And so why does that matter?
It's because it affects risk asset prices
because the bonds that the Fed doesn't own
need to sit somewhere.
And a lot of times they end up sitting
on U.S. dealer balance sheets.
And so what I mean by dealer is,
if you were a primary dealer,
you're a bond salesman for the U.S. government.
When the U.S. government needs to issue bonds,
whether it's, you know, it goes to auction,
whether it's a bill, which would be a short-term
or a 10-year treasury or a 30-year,
you know, they go to the various dealers and they say,
hey, this is how much we want to sell.
Those dealers take in bids and then they sell.
But if there's any left over at the auction,
so if the auction fails,
the dealers have to hold these, right?
And so the dealers also end up owning older bonds
and what they call off the run,
which are not the most active traded,
but that eats up balance sheet.
And you can just think about it like shelf space.
And so what we saw happen with the U.S. government shutdown
is the treasury general account went from 500 billion
to 1.1 trillion.
So the treasury general account
is like the checking accounts
or the piggy bank for the U.S. Treasury Department,
separate from the Federal Reserve.
So part of that increase was tax receipts,
but a lot of it was the government just not spending money.
And so when the government doesn't spend money, right,
there's less money in the banking system
by the tune of hundreds of billions of dollars.
So that put pressure on asset prices.
And just kind of, as we started recovering from that
in December and the first few weeks of January,
we entered another period
that I like to refer to as SaaSpocalypse.
So SaaS stands for software as a service, okay?
And software as a service company,
basically you got the highest multiples
of any other stock out there, both public and private.
And sometimes they were valued
based on their revenue, right?
Yes, because it is a software
that there's typically a long-term,
high-margin contract in place,
where it's a software that you're paying
a monthly, annual, multi-year subscription fee to.
So there's a very steady cashflow that's very high margin.
So that's easy to borrow against.
And because it's so steady and the reliability of it,
it was given a very premium valuation by the market.
So fast forward to January and the appearance of Claude Code.
So Claude is the frontier model,
the large language model of the company Anthropic.
So what Claude Code can do
is it can do a lot of software programming for you
through the use, through artificial intelligence, okay?
This has led Wall Street to believe
that these companies are just done for.
And so you now have the lowest exposure
of hedge funds to SaaS stocks in-
So they started selling all these companies
like Salesforce, all the SaaS-
Yeah, ServiceNow, great company,
just throwing the baby out with the bathwater.
Basically, I talked to a hedge fund manager last week.
Sell first, ask questions later.
And AI has gone in the market in certain sectors
to a very bullish factor,
to a great layer of uncertainty,
to where so many people are unsure
of how this is gonna play out for the various businesses
that stocks are just being sold
because they don't wanna be the one
sitting there holding the bag should the bottom fall out.
So how does that affect global liquidity?
Because there's a lot of SaaS companies
that get debt in the private markets,
in the private debt markets, so not public companies.
And these loans are sitting in the funds
of Apollo, very large private equity firm,
Blackstone, mega, mega private equity firm,
and another firm called Blue Owl.
So Blue Owl, they have many different strategies,
very smart credit shop, very smart guys,
but they have software-only debt funds.
So not a good time to be holding the bag there.
But the problem with that is
it's these firms themselves don't hold a lot of the debt.
Their funds do, but the firms are not
at great credit risk to this.
But the problem is, if there's all these problems
in these underlying funds, that when Oracle,
when Microsoft, when Google need to raise debt
to build data centers, their secondary debt source
after the large banks are-
It's gonna be these private equity funds.
They're tapped out.
But they're not available.
The funding is not there.
Yeah, because if all these funds,
if their performance is getting destroyed
due to fears about software as a service stock,
it's hard for them to go to external investors
and say, hey, I know I just lost you this money.
Well, come give me some more money
because this is a way better deal.
And so you combine that with the amount of debt
that some of these large companies are raising.
And so historically, Silicon Valley has funded
all of their operations from very high free cash flow
and equity.
Well, that's not necessarily the most efficient way to build
and most of these companies in Silicon Valley
have never had to build this level of infrastructure.
So there's a tremendous demand on the capital markets
for money for data centers right now.
And given what's going on with software as a service,
some of the key players are gonna have difficulty
participating in it and the sheer size.
So you're talking about a trillion dollars in debt
in a very short period of time.
And even if you're Wells Fargo or Citibank,
you can only lend a certain amount of money to Oracle.
Otherwise, you come up against a position limit.
You can only lend so much to one institution.
And so you're entering a period of what seems to be
a credit crutch, a la 2008.
I think the key difference here is this is not
a million dollar mortgage on a $500,000 home
that no one's gonna buy for $500,000.
Lending to data centers is a very, very good business
for these banks, which is why they're so aggressive into it.
They're just running out of money.
And it's so much so fast.
And the secondary players are under pressure from software.
It's put this cloud on top of the biggest stocks
in the world, the mag seven, if you will,
and that's pressuring all stocks.
And so some other ones have, there's been some rotation
and some other performance there,
but it's tough to watch for most people in their portfolios.
Now, before we get into the AI side of things
and the data centers, the SaaS space,
do you think the SaaS space is gonna get replaced by AI?
You know, I definitely think there's certain companies
that are gonna have a lot of trouble.
If you don't have a really unique software
with a lot of, the barriers to entry are definitely lower.
So when you have a stock that trades
at a premium market valuation,
and something like this happens,
you normally never regain that premium market valuation.
Right, and I'd say for now, some of this is for good reason
because if you don't have a large moat around your business,
why do you deserve that premium valuation?
But we were just discussing it,
and I brought this up yesterday.
Salesforce, right, the theory,
like Salesforce is the cheapest it's ever been right now
on an earnings base, earnings per share basis,
on a free cash flow basis, on a multiple of revenue basis,
like that it's ever been by 30 or 40%.
And so the argument is that cloud code
from Anthropic is gonna take it out.
Well, if you go on Indeed right now,
and you look at job listings for Anthropic,
they're hiring a Salesforce administrator
to run Salesforce for their organization so that, so.
So they can use it to sell their own product.
Yes, so it's, you know, the fear and the luck.
This is a great conference of money show.
I had privilege of speaking to CEO
of a publicly traded data center company last night,
and I said it to him.
The lack of underlying knowledge as to how AI works,
how the tech stack works,
where the demand is coming from and where it's going.
I've never seen a bigger disconnect on Wall Street
in my entire 25 year career.
Normally, most people that are playing with money
become very, very quick studies,
and they learn these things inside and out
so that they can make informed decisions.
The stuff that I hear from very sophisticated investors
talking about this stuff,
it's like, where have you been the last three or four years?
Like, have you not done any research
on how the underlying technology works
and where it's going to go?
So I think this is a speed bump
and maybe some really big speed bumps,
but we ultimately end up much, much higher
because the demand for AI and for tokens
and for the infrastructure and what it's going to do,
it's just at a stratospheric exponential pace.
You mentioned the data centers.
A lot of people are assuming or thinking
this might be a bubble, the AI spending,
and do you think these data centers will get built?
I think they will eventually,
and I love that you bring up that point
because those that think this is a bubble
are the same people that think cloud code
are going to take out all the SaaS companies.
And so why do I bring up that point?
When you look at a large language model,
so when you do a search on ChatGPT or Grok,
what I primarily use,
it uses a certain number of tokens.
A token is three quarters of a page
of a large language model.
When you start going to the agentic AI, the agents,
which is what cloud code would be,
you use an exponentially greater number of tokens.
So there's no way AI can take out these software companies
without way more infrastructure that we have right now.
And you are in wildly logistically-challenged moments
for AI right now.
In the US, they're starting to figure out
how to stand up these data centers,
but they haven't figured out how to get the energy.
And so the energy is the logjam.
And so Elon is out there saying
that he doesn't think chips are the logjam.
For every chip Nvidia sells,
they can sell another 10 or 15, right?
So there is so much demand
because as the availability of compute grows,
the demand is going to grow at the same rate,
if not greater, right?
So it's called Javan's paradox.
So as the cost of something,
the availability of something comes down,
the demand for it goes up.
So we're so far away from being done.
I think maybe at some point we'll end up in a bubble,
but that's years, years, years away.
And this data center stuff,
it's good business.
These are good loans for banks.
There's just not enough money right now
so that these hyperscalers can build what they want.
You know, that flies in the face of the bubble theory
that like, you know,
if you have 20 data centers that you want to build,
but logistically you can only build 12,
how do you get in a bubble, right?
We're not hitting our marks
in terms of what we're planning out to build.
And so the hyperscalers have announced,
I think it's $700 billion of CapEx next year, right?
It's estimated we're at 2027.
I don't think they'll hit those numbers.
Like, I think they'll be short.
I don't know if it's 10 billion
or a hundred billion or 200 billion.
There's just not enough.
There's not enough real estate.
Something like 90% of all data centers being built
have public opposition.
And 40% of public projects that have public opposition
don't end up getting built.
So it's a real,
it's going to take time to build these out
and the log jams and the supply chain
are going to continue this scarcity.
And so that scarcity is going to preserve pricing power,
right?
And it's going to, you know, preserve the demand.
So it's like, we have a long way to go in my opinion.
Now, in terms of liquidity crisis,
you mentioned that the Fed is going to inject more capital.
How's that going to impact the funding for these projects?
So, great, great question.
So I think, you know, we have a real problem here
in the U.S. with our government data, right?
And I think pre-COVID, you know,
there were problems in the data that existed,
but they'd be uniform.
So if you looked at the jobs data from August of 2018
to August of 2015, you get a similar error
and you can make apples to apples comparisons.
We're now, I don't even know what we're looking at, right?
So, you know, during the last years,
the Biden administration, the jobs number
was like it would initially beat estimates
and by the second revision would be below
like 22 out of 24 times.
Okay, the math on that is like a 12 standard deviation event.
Like there's a higher probability
you get struck by lightning twice walking to your car
than what happened with our jobs number.
And like, you know, maybe it's nefarious, maybe it's not,
but as an investor, it's just hard to rely on that.
And so tying this back into liquidity,
you know, there's a company I follow Trueflation, right?
And they take a million real time data points
and they translate it to an inflation.
It's been leading the government inflation numbers
by anywhere from three to six months, okay?
Over the last three or four months that has cratered
and like it's a daily reading
and sometimes it's coming in below 1%.
So I think over the next six months, inflation in the US,
we're gonna be getting readings
in the one, one and a half percent range
from the two, two and a half percent range we are now,
which is gonna clear the way for short-term rates
on the front end of the curve
to come down much more than the street predicts.
I think the street is,
we're not quite at two more rate cuts by the end of the year.
There's still some debate about it.
I think it's 75% odds, not a hundred percent.
I think we could end up seeing like a year from now,
you know, two and a quarter percent front end,
which represents, you know, six cuts, right?
So if that happens,
because all of these loans are SOFR based,
it just, it makes it easier to service the debt.
Like it just, and then the,
I don't think the 10 year treasury
and the 30 year treasury will necessarily move down as much.
So you're gonna end up with a very steep yield curve
and that steep yield curve is awesome for banks.
It's for the profitability, for the net interest margin
and for other people looking to take on leverage.
So like if you're the Blackstones of the world,
like you're borrowing,
your borrowing costs are much closer to SOFR.
And then, so you can go out and borrow at SOFR plus 50,
turn around, sell that money at SOFR plus 300,
make that 250 basis points,
something along those lines, right?
So it should solve a lot of that liquidity log jam
that we're seeing right now,
or at least grease the wheels
to where things can operate a little bit better.
Some people are concerned about the private equity space.
We have had a couple of bankruptcies in that area
and then people might not be sure
what assets they're holding and how it looks like.
So in the private equity, private credit space,
if some of these data center deals fail
or some of these other private,
so the private credit markets, right?
Where you don't have insight,
there's massive fraud that went out there
with fake invoices and stuff like that.
It's really hard to combat fraud.
But my experience with a lot of these guys
in the private equity, private credit space
is they are some of the smartest people walking around.
They don't want to lose money.
They don't want to take excess risk.
So there's probably some landmines
and some other nefarious things going on,
but it's important to know that that was fraud,
not the sign of a bubble.
That doesn't mean something else isn't waiting out there.
And what my fear in the short term is,
is that if a data center deal fails
and somebody's left holding the bag
because they priced it at,
it's priced at par and the interest rate is, say, 7%.
They can't sell it for par at 7%.
They got to sell it at 90 cents,
which represents, I don't know, an 8% or 9% yield,
something along those lines.
That means everyone else that holds similar debt
will have to take a massive markdown on their book, right?
With assets that aren't necessarily marked very often,
it's not like these-
So if the value of one drops,
then everybody else should-
It's going to create a domino cascading effect, right?
Which will cause a little bit of a panic,
but that's, to me, again,
there's a difference between a liquidity crisis
and a solvency crisis.
And so if you look at what happened in 2008,
the majority of the banks in the US
paid the government back with interest
as soon as they humanly possibly could, right?
And so when the Fed stepped in and injected liquidity
in during COVID,
again, you backstopped the short-term liquidity crisis,
right, whereas a bank like Citigroup,
and there was a solvency crisis
where their stock still hasn't recovered, right?
And so where the losses were greater than,
you know, there's nothing.
So I don't think that's the case,
the best of my knowledge,
that you're in a solvency crisis
and there's a lot of defaults sitting around the corner,
especially in this data center space,
knowing what I know about demand
and where demand is going to go
and what people are willing to pay for.
So I'm not so worried about that,
but every time there's a negative headline like this,
it puts pressure on stocks.
And for someone like, you know,
most people out there are excited about their stock portfolio
and if you've been investing in these AI and tech stocks
for the last three years, you've done awesome.
So anything that puts pressure on them,
it's like, ah.
So how do you see the future in this space and the economy?
I think the US is just absolutely poised
to just kick you know what.
So I'll explain why.
You know, we just did away with the endangerment finding,
okay, which is something
that the Obama administration put in place,
which allowed the government to regulate greenhouse gases.
Okay, in addition to all sorts
of other environmental regulations,
especially with the Chevron ruling
we got out of the Supreme Court.
So the ability for the US to add costs for manufacturing
and for other companies operating in the US,
that's gonna come off over time, okay?
You look at us being the leader in artificial intelligence,
the leader in robotics, okay?
And you add that to all the additional energy
that's gonna need to come on grid
over the next five to seven years,
you fast forward that five to seven year timeline,
we are now a manufacturing juggernaut, right?
Cheapest energy in the developed world, okay?
Best regulatory environments, okay?
We've always had the best rule of law
and the best capital markets, right?
So it's a lot to be excited about.
I mean, how AI is gonna end up playing out,
I mean, we'll see.
I tend to think it's gonna be a lot of upheaval
in the short term in terms of white collar jobs.
And those people that have white collar jobs
and have a college education,
they may have supported Trump,
but if they're not making the money
or things don't go that well,
and someone like Mondami shows up
and promises that in the world,
even though they're not gonna deliver,
like they defect and they'll vote socialists,
leftists, communists before you can blink an eye.
So I think the administration
really needs to get their arms around this turmoil period,
because in every major technological advancement,
it's always created more new higher paying jobs, okay?
I think the speed with which AI is coming,
there's gonna be a transition period
where that might not be the case.
And if that's not handled correctly,
like it could be a disaster.
So that kind of scares me,
but the other side of this, right?
The other side of this, it really,
it excites me because you're now gonna have
an industrial base previously thought unimaginable
on top of the services economy
we already have that's very good, all right?
And it's like, it's all these things
kind of coming together to make things work over here,
work really well.
Do you see an energy boom,
even if AI doesn't go the way it's planned?
The risk to AI is that it's faster and bigger,
not that it's a bubble that explodes, my opinion.
And energy is coming.
This administration has taken
a completely different attitude towards nuclear,
which is the cleanest, most efficient power in the world.
And like every time you turn your head,
there's somebody smart starting a nuclear startup,
whether it's a mini reactor for military bases
or Oklo building one for Meta in Ohio.
There's gonna be all sorts of new energy that comes online
and like the attitude towards nuclear energy
in this country has fundamentally shifted
for the first time in a long time.
I think that's because we're a long way
from Three Mile Island.
But this lowering of the energy costs, right?
And this movement away from green nonsense, right?
Is a huge net positive like for the humanity
and for the United States.
Do you have any recommendations for people
in terms of assets or what they should do
with their finances?
My favorite stock, and it's come down substantially,
so I think it's a very good entry point right here,
is Palantir.
So what Palantir does,
they're the glue that holds AI together, right?
So to be able to get all your information
from all your various systems and all your data sources,
structured and unstructured,
to be able to feed that into like a cloud,
a open AI, a chat GPT, or a grok.
Like you need software to organize
all those things together.
And then when it comes out the other side,
you need to be able to implement it in their model.
And so Palantir has been working on these sorts of solutions
for over 20 years, right?
And like, you know, they're growing at an insane rate.
The way that they structure their business
is they'll start with their software
in one of your business units
and then spread like wildfire to the rest.
So with their existing clients,
they keep getting bigger and bigger contracts, right?
As well as all the defense applications
and everything that they've done.
They are the best positioned company in history
for what is happening right now, right?
I think it's the most transformational company
I've ever seen.
It's run by extraordinarily smart people.
And just to touch on an earlier point,
they don't, the executives there, Shyam and Alex Karp,
don't believe that AI is going to take out millions of jobs.
They think it's going to be a creative,
help people do more.
I hope they're right.
I think they may be wrong in the short term.
I also, I like that all the chip makers,
because you're just in a situation
where we just don't have enough GPUs to build these.
So that's your Nvidia, your Broadcom.
So Broadcom makes TPUs for Google
and Google's cloud is growing.
They're building tons of data centers,
just did a huge bond deal.
And then Marvell is a little bit under the radar.
Marvell's making GPU, custom GPUs
for both Amazon and Microsoft.
They're ideally going to triple their revenue
in the next couple of years, but it's good business.
So I like, from the software,
I don't know how this SaaSpocalypse is going to play out.
You know, Salesforce might actually be the deal
of the century right here,
but I've used Salesforce before.
I don't like, as a user, I don't like it.
So I'm biased against the stock.
So I won't own Salesforce.
I'll own Palantir in the software space.
Microsoft, the stock has gotten destroyed for no reason.
They're unbelievably well-positioned.
They're already in in all these enterprises.
It's an exciting time
where you can buy big, well-known stocks,
kind of hold them for a few years,
and I think make outsized returns.
Mike Lee, it was great to have you on Market Insider.
Yes, thank you for letting me talk the entire time.
What do you think of this discussion with Michael Lee
at the Money Show conference?
Do you agree with Michael?
And if you have any comments, leave them below.
I'm Siamak.
This is Market Insider.
We'll see you next time.