So the bubbles are going to be more

likely to be showing up in those

sections. Many of those companies are

still private. The people are going to

be holding the bag. It's going to be the

VCs. So by itself, you say Chachi PT

Open AM and you need 234 trillion in

revenues eventually in this business to

justify this much being invested up

front in the LLMs themselves. As an

investor, you're going to get eaten

alive if you go into that space and

bring your valuation to justify what you

do cuz God help you. If you're going to

make an argument for a bubble, it's got

to be based on something that I'm not

seeing in the numbers. When the

correction to the big bubble happens and

people look back, they're going to start

writing off. There's going a lot of

impairments of AI investments. But guess

what? Nvidia will not be asked to return

the cash they got in the chips, right?

Because that's been spent already. Later

on, you'll hear more about the Fundrise

Income Fund and why sophisticated

investors are turning to higher yielding

assets like private credit. But for now,

let's get into today's interview. Very

happy today to be joined by Professor

Oswalt Deodorin, finance professor at

the New York University Stern School of

Business. Professor, it is wonderful to

be speaking to you here today. I want to

ask you about the valuation. How are you

assessing the valuation, whether it's

fair, under, or overvalued of US?

>> What do you mean by the valuation? The

valuation of what?

>> The S&P 500. It's a richly priced market

which is building an expectations

that the pathway is going to be a benign

one in terms of the economy and how

companies react to it. So it's no

different than it was at the start of

last year. What I said at the start of

last year is if to the extent that the

market is right, you're going to be able

to see the numbers delivered. The

challenges there are lots of things on

the horizon that could potentially trip

up the market. political, economic, war,

and those things don't seem to get get

be priced in. And so far, the market's

been right. You know, it it kind of blew

off the tariffs after an initial shock.

And in hindsight, it turns out that the

economy didn't collapse. Inflation

didn't come back. So, so far at least,

the market has been right in its benign

stance, but it's pricing in more benign

circumstances.

>> Is there a big market delusion in NI?

>> There always is. It's a given, right?

Anytime you have a big disruption cuz

think of what drives a big market

delusion it's human nature you see big

market you're an ambitious smart young

person or even older person what do you

think I want to be part of that big

market that's the entrepreneur mindset

so you start a business you go raise

capital who do you raise it from venture

capitalists and there's a selection bias

they listen to your story the venture

capitalists like your story put money in

so you've created a pod of overconfident

people looking at a big market thinking

they can conquer the market. If this pod

values itself, overconfidence is going

to show up as overestimated revenues and

growth and cash flows and over. So the

it's almost it it's a feature of big

change is you will get the big market

delusion. Too many businesses started

going after a big market individually.

Some of those companies are going to

become great companies or winners but

collectively these companies are going

to be priced too high.

So, it happened with the PC business in

the 80s, with the internet in the '9s,

with social media in the last decade,

and now it's happening with AI. And what

will it mean? There will be a correction

along the way. Does this mean that if

you sell short on AI stocks, you're

going to come out ahead? Not

necessarily, because a few of them are

going to be the big winners, but

collectively there's going to be a

cleaning up phase. And that's coming.

And I think it's healthy when it comes

because it is how we change as human

beings. we overreach and then we

correct. So there is a big market

delusion. There will be a correction as

a consequence. But I don't I think

that's part of being in a market. That's

how you let markets create change.

>> There's there's almost there's almost

always a big market delusion in every

big market

>> because human beings are over

optimistic. They're the ones who go

after it. So it's human nature to

overreach. In fact, one of the

expressions I use in my class when

people complain about bubbles and

overconfident entrepreneurs is I ask

them the question, would you want to

live in a world run by actuaries?

[sighs]

I'll tell you what that world would look

like. Would still be in caves because

actuaries base everything on expected

values, probabilities, and they're

saying this fire thing, it could get out

of control. So, let's make sure we have

all the numbers we need before we leave

the game. Thank God we weren't run by

actuaries, right?

For change to happen in economies, you

need people to overreach. The

consequence of that is you're going to

have market bubbles and market

corrections. They're a feature, not a

bug. You want to have a market with no

bubbles, fine. But that market is going

to create no innovation.

there's going to be no big change that

happens because you need chaos almost as

a feature of markets for this kind of

change to happen on the spectrum of an

extreme big market delusion to a minor

one. Where would you place AI and could

we divide that perhaps into the public

markets where the the revenue growth and

the earnings growth looks very good

particularly for Nvidia but also the

hypers scale companies? And no, wait,

that's not true. That's the architecture

of AI, right?

>> That business is going to that's like

Cisco was the first company with the com

boom. So that's actually the easy part.

You're building the architecture and

people are spending a lot. So if you're

thinking about Nvidia, you're think

about the data centers, the power

companies, constellation energy, you are

on the easy half of the equation, which

is the AI architecture is getting built.

There's no uncertainty there that people

are spending money, it's going to show

up. The uncertainty is so this is the

cost of building the factory. You know

what we're uncertain about? What product

that factory will produce and whether

anybody will buy the product. That's

where the delusions are forming. Not in

the architecture side but in the product

and service side. Companies that think

they can create a product that people

will want. They don't even know whether

there's a demand for it and are creating

businesses around this and attracting

capital to get in those businesses.

That's where you're going to see that's

why I think you're going to more likely

see a big market delusion with the LLMs

than you are with the architecture true

architecture companies because

collectively if you think about what

what LLMs do there's really you know if

you look at Chat GPT or Grock or Gemini

or Claude by itself it's not a

money-making device. You can charge a a

subscription but it's not going to be

enough to cover the huge cost. You've

got to figure out a way to make this a

commodity that other companies need to

produce products and services for

customers. And that's the challenge.

That's the part that's unknown. And

that's where I think people are

overreaching because they think they can

pull off stuff that I don't think on an

expected basis they can pull off. So the

bubbles are going to be more likely to

be showing up in those sections. Many of

those companies are still private,

[snorts]

right? So the people are going to be

holding the bag. It's going to be the

VCs in that space who are making these

bets right now on those companies.

>> And rough numbers, open AI has raised at

a $500 billion valuation is rumored to

attempt to be raising close to a

trillion 850 billion. Their uh annual uh

run rate of their, you know, their

current revenue annualized is 20

billion. Estimates 2026 of 30 billion.

So that's close to a 30 times forward

price of sales. You know, you you also,

you know, have the numbers of Anthropic

and XAI. When you look at those numbers,

what are you thinking?

>> Collectively, they're overvalued.

It's it's a there will be a winner

coming out of the space who might be

worth what you're investing in right

now, but if you took those companies

collectively and you put your money on

them,

there's going to be a correction in that

collective space. There isn't enough

revenue in the endgame to justify this

kind of pricing for what is essentially

part of the architecture. The LLM are

part of the architecture. They're not a

product of service by themselves. So,

collectively, I think they're

overvalued.

>> Tell me what you mean when you said

they're not a product by itself because

okay, a very elevated multiple. You can

perhaps justify that if you're going to

get a world where it turns into a

Google, it turns into a meta, it turns

into these.

>> No, but remember Ge Google uses Gemini

and it's advertising. It might get

revenues to the advertising business,

right? That's what I mean about a

product use. That's one place where an

LLM is being used to actually generate

revenues. So, think of that as what I

mean by the end revenue stream. Other

companies will have to license I don't

know what the exact business model will

be. licensed chat GPT as a vehicle not

to sell sad chat GPT but as a vehicle

for products and services that they will

then sell their customers based on that

cha chat GPT

you know the the the foundation so by

itself you say chat GPT open as I said

there's a small subscription model you

can build upon them but it's not going

to be $2 trillion and you need 234

trillion in revenues eventually in this

business to justify this much being

invested upfront

in the LLMs themselves.

>> And then perhaps on top of that, even if

the, you know, it met your revenue test,

it might not meet the profit test

because it's so subsidized and the

margins are so low. On that one, I'm I'm

I'm a little more upbeat because the

nature of these businesses seems to be

winner take all businesses, which is if

the revenues are that much, it'll end up

with three or four companies dominating

that space. And given the history of

what technology dominance has look like,

those companies are going to deliver

margins like the Metas and the Googles

in whatever space they operate. So, I

think the revenue part is actually the

tougher part. the margin part might be

more accomp might be easier to

accomplish because of the winner take

all networking component to how these

these company we don't know that yet but

that would be my prior is it's going to

look more like a technology business

than it is going to look like the

oldtime manufacturing business more

concentrated three or four players in

each game all of whom can you know can

collect big margins because there's

really no outside competition

>> what are the network effects for the

large language models for for Mastercard

Visa it's that everyone uses them for

Meta it's that you know no one wants to

go on a social media app that was there

are no users you your friends aren't on

there for the opens and and its

competitors what what is that

>> right now none right right now none

because we're using them in a very

casual way I use chat GPT to summarize a

letter I could ask grock to do it Gemini

to do it that's why I said this is the

early stage in the process in the

commercial phase it won't be chat GPT

you'll be using it'll be a product that

comes out of chat GPT a service that

comes out of chat GPT so you as a

consumer the person buying the product

or service are not interacting with chat

GPT you're interacting with the company

that use chat GPT to deliver this

product or service to you that's where

the dominance comes in is because those

companies will create their own versions

of you know networking benefits which

means if you're not on the platform that

they've created you will not be able to

get those benefits. I mean, think about

it now. Smartphone the the the Apple

iPhone is should be getting 40% of the

revenues for for that Door Dash makes

every year if in if this were a fair

word because you have a bunch of

businesses that are built on the

smartphone. So, Door Dash, Airbnb, you

can argue that all of these businesses

would be much smaller and much less

successful without the smartphone

operating as the delivery mechanism.

That's what I mean by LLM. LLMs will be

like the smartphone in the Door Dash

business. You're not dealing with your

smartphone, you're dealing with Door

Dash. The networking benefits come from

what Door Dash and Airbnb create in

their platform. And that's what I think

will be the endgame in the product or

service business that you see coming out

of AI. And I think you know we have to

wait to see how big that business is.

But it's going to be tough to sustain

how much we're investing up front in AI.

>> So you said open AAI and its competitors

anthropic XAI are overvalued. How much

overvalued are they and what valuation

would make sense if if 350 billion

doesn't make sense for Enthropic? 500 or

850 billion doesn't make sense for open

AI. What does make sense to you

>> at this point? Who cares, right? None of

them are public yet. There are VCs who

can only be long, they can't go short.

So your decision is should I invest or

not invest and basically collective

overvaluation means unless you have some

technological expertise that allows you

to pre-identify

which of these LLM is going to end up

the winner. this is a space you should

stay away from. Once they go public, the

decision will get a little more concrete

because now you're start going to start

to see the and early on it's still going

to be you're buying an option. You're

hoping it pays off as the winner they

call. But as the numbers start to roll

in, this will be less of a opaque game

and more transparent. If you're an

investor, my advice is stay until you

feel you're in that that that place

because right now getting into any of

these companies, you're playing the

trading game. It's mood and momentum.

And you can ride the momentum if you

want, but there's nothing to do with

investing in valuation.

>> You're playing the trading game and you

can't even trade because it's not

liquid.

>> No, but I'm saying once they go public,

the early part of going public, it'll be

it'll be completely a trader's paradise.

As an investor, you're going to get

eaten alive if you go into that space

and bring your valuation to justify what

you do because God help you that

valuation is not going to matter on a

dayto-day, week to week, monthtomonth

basis in terms of what it does for

prices.

>> Earlier, professor, I I brought up uh

Nvidia, which is building the

architecture, the other chip companies,

as well as the the hyperscalers, and you

said, "Oh, well, I'm focusing on the

private markets." That's the real

question, the LLM. But a lot of the

earnings growth in the the publicly

traded companies that the chip

architecture that the semic companies

and the hyperscalers is because of the

private LLMs right so

>> being their customers right

>> yeah they are their customers

>> it's their customers so basically in a

in a sense it doesn't for if I'm

investing in Nvidia

do I care that companies are

overspending on chips no in fact it's a

good thing for me right

>> so I don't see how you know that plays

out the only way this process starts our

feedback effects is when the correction

to the big bubble happens and people

look back they're going to start writing

off there's going a lot of impairments

of AI investments but guess what Nvidia

will not be asked to return the cash

they got on the chips right because

that's been spent already so in many

ways Nvidia and the chip companies and

the data centers are the not just the

front end of AI they're the most

protected the money's already been spent

the mistakes are already sunk costs

and even if in hindsight we decide this

bubble burst, we're not going to go back

and collect our money back from them but

it will impact their continuing growth

because the new money coming in which

they need to to justify will start to

level off. Earlier you brought up Cisco

and said okay Nvidia is the Cisco of

this cycle even though I'm sure the the

investors who lost the most money at

least in percentage terms were the

investors in the profitless in some

cases revenueless VC companies the

pets.com of the world in investors in

Cisco lost money investors in Microsoft

love lost money and I believe it was

only over the past 12 months or so that

Cisco hit its uh um its high in 2000

although with dividends it probably did

that before that in in total return. So,

you know, if you think the LLMs, you

know, I'm just connecting the dots here.

If you think LLMs are overvalued, you

didn't use the the bubble word. Uh, how

are you assessing the risk that this is

going to stop? Because if if Nvidia is a

beneficiary of all this capex and the

capex stops, not only is the revenue

growth going to go down, the profit

growth going to go down, but I assume

the multiple is going to go down. You

know, you have the exact numbers. I'm

just going to take a rough guess. Say

trailing 12 months, PE for Nvidia 50,

forward roughly 30. you know, it

justifies that growth. If if it

continues growing, it's it's it's grow

there. Don't get me wrong, if the if

there's a correction, Nvidia is going to

drop in price. There's no way around it.

The question is percentage drop. And

it's not just Nvidia and AI companies.

The market is going to take a drop

because this is such a big part of the

market. There's no place to hide in

stocks if this correction comes because

there will be punishment meed out in

your portfolio even if you don't hold a

single AI stock in your portfolio. Not

one, right? The nature of the beast. So

there there will be a cleaning up phase

where everybody up and down the cycle

and even people outside are going to get

hurt when that correction hits, which is

one reason people complain about

bubbles. Look at how much money I lost.

They don't have the perspective to think

about when they entered the game, how

much money they made before they lost

the money. They they focus just on the

losses. But there will be cleaning up as

I said and that cleaning up is going to

be painful for everybody involved.

And the reason you mentioned Cisco being

a bad investment, part of the reason is

for a decade after the com bust, they

seemed to not recognize that the bust

had happened. They kept acting like this

was a a temporary phase that they could

keep doing what they had been doing

which is buy another 15 companies every

year small internet companies and keep

growing. I have a feeling that if they

recognize this earlier in the game,

you'd have seen much a much smaller drop

off in market cap than you did. So the

lesson for Nvidia is when that

correction happens bring you know bring

your ambitions down. Don't try to keep

doing what you used to when times were

good and there was this, you know, 30,

40, 50 billion being invested by big

companies and AI every year. You got to

live with the moment. And listening to

Jensen Wong, he strikes me as a sensible

man that he would adjust to what's

happening out there that he does he

doesn't strike me as leading with his

ego.

>> So that is the risk factor. You

described the left tail. How are you

assessing that risk versus I guess your

your base outlook versus the right tail

risk where all these you know venture

capitalists have all these podcasts and

you know if what they're saying is going

to come true Nvidia's revenues is going

to be you know a trillion dollars in a

few year where how are you assessing

that that that spectrum

>> I mean it's a sense the world is a big

place people have different opinions

>> I don't see I really don't care what

other people think because it's my money

that I'm investing if they want to

invest based on a right tail of course

They they should now I'm not going to

talk them out of it.

>> Right. So sorry to be clear I was asking

about about your opinion on the Nvidia

the the publicly traded companies that

are beneficiaries of this whether it's

the Nvidia or the other semi companies.

I think you said you had invested in it

since since 2018 in a YouTube video we

can we can link to that came out

recently maybe a month ago. You saw you

thought it was overvalued but I believe

you said that you still owned it. It's a

pl there and there there's a plausible

path I said for it to be the value to be

justified where and so at this point if

you ask me would I buy Nvidia at today's

price I wouldn't I think where you know

you need too much to go right to break

even but that's me and that's my money

if you ask me should I buy Nvidia I'm

going to give you the numbers I'm going

to give you the distribution and say you

make your own judgment it's your money

to invest so I think that it is richly

priced it's been richly priced for the

last four years. But remember, as it's

been richly priced, it delivered huge

returns for investors. So, I think that

too much has to go right for me to buy

at today's price. And it also draws a

distinction between how you treat

something you're planning to buy versus

how you treat something that's already

in your portfolio. I mean, in valuation

books, it seems like a 01, right? If

something is overvalued, you don't buy

it. But if something is overvalued, you

have it in your portfolio, you sell it

right away. And I think that misses a

lot of realw world issues. Taxes for

instance, time horizons, you know, what

are you going to do with that cash? That

mean that you're far slower to pull the

trigger on something you already own

than you would if you were buying the

same thing, right? So I you know I sold

my last quarter of Invidia at the end of

last year but it was staggered over four

years

in terms of how much I sold each year.

And is it possible that I would regret

it? No. Price could go up but I don't

believe in looking back and saying I

wish I'd done this. I wish I'd done that

because that's a pathway to all kinds of

damage in your portfolio. Now, I'm I'm

okay now with what I I mean, Nvidia has

been incredibly good for me as an

investor and the fact that I didn't make

that extra 25%. Now, I'm okay with

leaving that on the table in return for

the passing the sleep test.

Sophisticated investors don't typically

park their cash in high yield savings

accounts. Instead, when they want yield,

they turn to one of the fastest growing

asset classes for institutional

investors, private credit. These private

credit strategies were once reserved

exclusively for wealthy institutional

investors, but now they're available to

investors of all sizes thanks to the

Fundrise Income Fund, which has more

than 600 million invested and a 7.97%

distribution rate. With most traditional

savings yields falling, it's no wonder

why sophisticated institutional and

individual investors are turning to

higher yielding asset classes like

private credit. Click the link in the

description or visit fundrise.com/m

to invest in the Fundrise Income Fund in

just minutes. The fund's total return in

2025 was 8% and the average annual total

return since inception is 7.8%. Past

performance does not guarantee future

results. Current distribution rate as of

12312025. Carefully consider the

investment material before investing

including objectives, risks, charges,

and expenses. This and other information

can be found in the income funds

perspectives at fundrise.com/income.

This is a paid advertisement. Back to

the interview. What about Microsoft?

>> Microsoft I've owned since 2014. it's

still intact fully in my portfolio

partly because the plausible pathway I

think is more reachable for Microsoft.

So with Microsoft too I wouldn't buy at

today's price but I don't need as much

to happen to justify leaving in my

portfolio. So it's a question of degree.

To me, Nvidia, the degree to which you

got to deliver great stuff is far

greater than it is in Microsoft. And

Microsoft since it's already in my

portfolio. And given that I live in

California with the taxes that I have to

pay in capital gains, I can live with a

slight with an overvalued Microsoft in

my portfolio as long as that

overvaluation is within within range

within it's something I can live with.

How do you assess the growth in its

cloud business which has been ve very

robust but the customers are the LLM

companies that that you think are

overvalued and so is is the is the

growth engine for Microsoft the the the

venture capital funded companies that

>> I think this has become an excuse for

old revenue growth at tech companies

it's coming from AI would love to see

how much of the growth in the cloud

business actually comes from AI I mean

let's face it Microsoft didn't have a

stagnant cloud business which suddenly

took off with AI. This has been the

fastest growing component of Microsoft

because we're all on the cloud. You are

in the cloud. I'm on the cloud. I mean

the cloud has become almost part of not

just business life but individual life.

So, you know, much as you know, so much

as I might worry about people

overspending, I don't think it's as

dominant a piece of the cloud business

at any of these companies as people make

it out to be. In other words, tomorrow

AI spending dropped. I I'm not sure that

you're going to see cloud revenues at

the at any of these companies decline.

The growth might get lower, but I don't

think you're going to see a drop off in

revenues. How do you assess the cloud

business in you know do you think it's a

it's a good business or a bad business?

It's become a necessity, right? It's

like a utility, right? We need it. We

need it because so much of everything we

do has to go on the cloud because we

have so much data that carrying it

around in our phones or putting it on

hard drives has become almost impossible

to do. the way we transmit information,

the way we store information and the

cloud also has relation I mean because

so many businesses you know they're

connected to apps that are connected to

the I mean we're all you know dependent

on the cloud working as we re realize

the minute one of these clouds goes down

we realize how much of the world comes

to a stop because so much it so I think

it's a necessary business the question

is what will the margins look like as it

matures as a businesses because there'll

be a point in time where even the cloud

business levels off and the question

then is will the margins stay where they

are which are very healthy numbers for

both Microsoft and Google or will they

start to decline because it is a mature

business and competition is going to

kick in.

The other thing I'm not quite clear

about that somebody can is how sticky

these businesses are. the customer picks

Amazon cloud over Google cloud for their

business. How difficult is it for them

to disentangle themselves from one and

go to the other and my guess is these

[clears throat] companies create

stickiness in their models which makes

it difficult for customers to leave once

they're on a particular cloud cloud

platform. So that creates the stickiness

and the networking benefits that leave

you in a particular cloud platform even

if the costs look a little lower in a

different platform. Switching might not

be easy in this business.

>> What about Google?

>> Google is a company that's still so

dependent advertising. You got to wonder

how a company with so many smart bright

people and it has never been able to go

beyond the search box in terms of

actually delivering value. I mean that

much talked about alphabet is really six

dwarves

and one giant, right? Because none of

the bets have really been able to stand

alone as businesses. A couple of years

ago, I wrote a piece on on sugar daddy,

the what I call the sugar daddy problem.

And what I was talking about is

is corporate VCs where you know and

sovereign funds what do they share in

common? They operate in a business where

there are regular VCs and regular

investment funds. But the difference is

a corporate VC when things go bad can go

to the parent company. The parent

company says oh you poor people here

take another two billion. Right? Same

thing with sovereign funds. You lose

money people don't fire. to stay the

sovereign fund because the Saudi

government is not going to get rid of

the Saudi sovereign fund. So what

happens is when these business when

these investors make mistakes for to be

a truly effective VC you got to be

ruthless in terms of cutting your losses

and walking away from bad businesses.

But that's tough to do for corporate VCs

because they keep getting this money to

push more and more because many of the

businesses are within the company,

right? You don't want to cut the So I

think there is with Alphabet the the

question is can you change the way

Alphabet invests to allow all this money

they're investing to actually show up as

value. I think that with Gemini, they've

actually played the game much better

than they have with their other

investments

and maybe way more will be the other

opening they can use to kind of get into

automated driving. But I think a lot

remains to be seen as to whether they

can build businesses in those areas that

actually deliver value on their own.

>> And you have a YouTube channel far

bigger than my own YouTube channel. So

you know that the the power of Google I

don't there's some parts of YouTube

that's not disclosed obviously you know

Google's own by the same parent company

Alphabet as Google. How do you how do

you value YouTube?

>> I value all these companies as

ecosystems

like Facebook, Instagram, WhatsApp are a

part of an ecosystem. The entire

ecosystem delivers the advertising

revenues. So even though the revenues

look like they go to Facebook, it's

somebody using WhatsApp who sent a link

to somebody else using WhatsApp to go on

a Facebook page. So So these companies

just want you to stay in that ecosystem

and Google keeps people in the YouTube

has become a very effective device for

Google to keep in the ecosystem. It's

one of the few success stories of a

standalone business as well because you

also have the YouTube subscriptions,

people getting their TV channels through

YouTube. So I think you know of all the

things Google has invested in YouTube is

probably the greatest single investment

that billion dollars invested in YouTube

has paid off 100fold in terms of value

created for the company. So I you know

to me it is a free platform for Google

though there are benefits they get their

ecosystem they're not doing it for

charity. So, you know, it's it's it's I

think been something that's added to

Google's value, just like Amazon Prime

has value way beyond just the Prime

subscriptions on Amazon. I think YouTube

has value way beyond what they collect

as subscription and actually direct

revenues on YouTube. And now we turn to

Meta, which I believe unlike a lot of

the other Magnificent 7 companies is is

a vast majority of their spending on AI

is for themselves and for their own

product. you know, Microsoft is spending

money and but for other customers, same

with Amazon, whereas Meta, they're

they're spending uh their own money for

their own product to improve that that

model. So, I think there we could see a

more direct ROI or ROI, you know,

>> and you know what I would track? No, FA

Meta actually keeps track of how much

time a meta user spends in the meta

system. At least they do it at the

Facebook level.

If using AI keeps you in the meta

ecosystem, I think LA the last estimate

I saw was 57 minutes in the day. You

know, typical human being spends in some

part of the meta ecosystem. If that can

become an hour and a half, that's a huge

plus in terms of advertising revenues.

So there the ROI will not show up as

direct money you're paying for AI

products and services. it might show up

as you spending more time on Instagram

because AI has made it more attractive

to you. They found a way to make it more

addictive to you and that then shows up

as more advertising. So I I think that's

what I would watch is time that people

spend in the meta ecosystem is it going

up because if AI is being effective

that's why you'll see the effects show

up.

>> So the the free cash flow for for meta

has been declining. I you know on your

website you wrote about how you you

think that lease liabilities

I actually forget but they should count

as liabilities like I think it's not on

its balance sheet yet but Oracle has

half a trillion dollars of of lease

liabilities like there we are going to

find if you know we we

>> Oracle's gone a little off I mean they

they've left the they've clearly gone

off the map in terms of their AI

spending they've I mean they've made a

huge bet so I would actually set Oracle

separately from the companies we've

talked about. I think Oracle is overb

partly because they want to be a

trillion. They want to be in that club.

They wanted to be mag eight with Oracle

as a in many ways they they they remind

me of a gambler who puts all his chips

on one number hoping it pays off because

if it pays off they become part of the

trillion dollar club. So, I think

they're they're investing very

differently from everybody else because

they seem to want to get this big pot

that puts them into that club. I think

for Meta, cash flow, the free cash flow

is down for all of these companies

because that capex on AI comes out of

your free cash flow. But here's the

difference between the com bubble

investment and the AI investment. The

dot bubble investment was far smaller

than the AI investment. You didn't have

tens of billions of dollars being

invested. And second, the companies

making those investments are often money

losing companies.

So the bubble hit doubly hard because

when the bubble burst, not only did the

the the shine come off those

investments, you now had money losing

companies that were struggling to make

it. That's why it's such the big AI

investments. I'm I'm not there are of

course exception. Big AI investments now

are being made by some of the most

cashrich companies on the face of the

earth.

And the debt they take on, I know there

are debt stories. At least in these

companies, the debt is such a small

fraction that they could pay off all the

debt they owe with one year of free cash

flow. None of them is going to go

bankrupt. None of them is going to see

their ratings drop to see because they

have so much cash cow cash flow coming

in. So with a Meta or

Microsoft, even if there's an AI

collapse, there will be an impairment.

But you know what? They make so much

money that their impairment will still

leave them with positive earnings. This

is not the kind of impairment that

created huge losses and default risk. So

will there be regrets if this works out?

Absolutely. But those regrets don't have

to translate into default risk and

distress like they did in the dotcom

bubble. Amazon came very close

to going under in 2001. The only thing

that saved them was Jeff Bezos raising a

billion dollars towards the start of

2001 and leaving it as a cash balance

and that basically allowed them to get

through the hard times. In this case, I

don't think you have that kind of

default risk hanging over. But there'll

be a subset of the AI companies that

have borrowed money to find AI and those

companies are going to be in trouble

when the correction hits. what you said

about how the the the co the the

customers spending on AI are some of the

most profitable companies uh that have

ever existed, the largest companies that

have ever ever existed. I have said that

almost almost word for word myself and I

believe it to be true technically

because Microsoft and Amazon are so

profitable. However, a lot of their

customers as as we talked about earlier

are these VC backed companies that are

burning a lot of money and um you know

there was one out

>> customers for I'm sorry customers for

what

>> for for AI cloud like like you know

Microsoft is buying

>> the cloud is a different business the

cloud is a business that's independent

of AI right it's a business that AI has

helped but it's a business it's a

function of the times we live in right

so we we shouldn't act like the cloud

business was created by a portion of the

cloud business benefit from AI. A big

chunk of the cloud business has nothing

to do with AI. It's got to do with the

way we store and use data. And we live

in a world where there's so much stuff

to use and store that we need the cloud.

So I think we're overstating how much of

Amazon. So let's say we took Amazon's

total revenues and we took the portion

of the revenues that the cloud business

that the AI business is creating for

cloud. I'll wager that fraction is a

tiny number. I might be wrong because

they don't actually break it out, but I

wager it's a tiny number. So, I think if

you're worried about clients collapsing

and that affecting their revenues, it

will, but at the margin, it's not going

to cause their revenues to drop by 20%.

What it'll do is it'll make their

revenue growth go from 8% to 4%. That's

where you're going to see it show up,

which is bad, but it's not catastrophic.

So I think that that there is a segment

which is particularly exposed to AI

collapse

but these companies have other other you

know other things on the table that will

allow them to generate cash flows and

value that will keep them going well

past that collapse.

>> I bet you're probably right about Amazon

and I think it really is important to

look at what is the reality. Like a lot

of people thought oh well uh these

tariffs are going to be horrible for

China. let's sell China or short China.

A big stock in China in the Chinese

market is Alibaba. I looked at it and

yeah, the the Alibaba's revenue to the

US is around 10% or maybe even less. I

it's totally important to look at you

know beyond the narrative what are the

actual numbers but professor wi with

Microsoft I think you know to the extent

that their capex has exploded their

capital expenditure I think a lot of

that is because of the AI and they're

buying GPUs

>> capex that's the bad side right I'd love

to know how much of Microsoft's revenues

comes from pilot or whatever other AI

stuff you think because there again I

think the percentage of revenues that

Microsoft gets that's AI related is is a

small fraction. So, I'm not saying that

it's not a big chunk of the cost. That's

the negative. The revenue side, which is

where you're going to see the impact of

a correction, is going to be much

smaller than we think it is because

these companies get a relatively small

percentage. They have big revenues, so

relatively small percentage is still no

billions or tens of billions of dollars.

A relatively small percentage of those

revenues.

>> Yes. And and I guess you're part part of

your point is that AI bubble or not, you

know, Xbox and uh you Microsoft Word are

still going to be tremendously

profitable businesses

>> and the cloud and the Microsoft cloud is

still going to be immensely profitable

after you've written off that AI

segment, right? Because the demand for

the cloud doesn't go away just because

AI is has crashed and burned or has

become a much smaller story.

>> What about Tesla?

>> Tesla is interesting, right? I mean,

it's a company that I'm not even sure

what it is anymore. And even the people

who are the strongest advocates for

Tesla hardly ever mention cars anymore.

It's almost like it's become an

afterthought. It's robots and it's

automated driving and ride sharing and

all these potential businesses. And

that's always been Tesla's power, right?

To keep shifting the narrative from what

you think it is to something different

and being able to carry enough investors

to buy into the new story.

You know, I used to think it was a

because it was a young company with a

charismatic CEO, but this is now a a

trillion dollar at least at points in

time, a trillion dollar company where

the CEO has been around enough that we

know who he is in terms of his pluses

and minuses and it's still able to do

it. So, you're writing a case study

about a company that can shift

narratives and carry investors with it.

Tesla would be a perfect case study

which is you need investors who are not

just investors they're fans they love

your company they love and you're able

to convince them that you're now a

different company and they go along

right whether it's you're now an energy

company you're now a robotic so I think

people are buying not into the product

you're making but some underlying

strength they think you have to enter

new businesses

disrupt them change them and end up

dominating them the forward Forward

price to sales for Tesla is is 15. The

forward price to earnings is 220.

>> You know what? It's the only way you can

justify Tesla's valuation is by mapping

out a way it gets into a different

higher margin business. It cannot

sustain this value as a car company and

Tesla and that's where I think you you

know even Tesla bulls will agree with

you on that. The question is what is

that business? Will it be big enough?

Can they make enough money? As I said, I

don't feel enough confidence in any of

these businesses, robotics, automated

driving, that I'm willing to make a bet

on it. But if you're a Tesla investor,

you might have the confidence that one

of these businesses or maybe both of

these business are going to be big

enough for you to jump in. So, I would

agree to disagree on that. And I, you

know, I also don't like my stocks to

have a political component to them. And

Tesla, for better or worse, has become a

political stock. What I mean by that is

people buy or sell your products based

on what they perceive your politics to

be. You're in a dangerous place as an

investor. And I and I said I don't want

that. I don't like that. The reason I

don't I'm underinvested in China is

precisely that reason. The government is

a player in every story. And Tesla, for

better or worse, that's become part of

the story. The politics has become part

of the business narrative. And I'm not

comfortable with business narratives

that have politics in them. So it's part

of the reason Tesla was the first of the

Max 7 that I exited from and it's gone.

It's doubled in price since. So again,

I'm not going to look back and say, you

know what, I should because I did it

because the reasons I specified and

those reasons still hold

>> and you may have been wrong on on the

price. I don't know when you sold, but

in terms of fundamentals, I should say

the revenues and the margins have gone

down over the past, you know,

as an investor. You don't care about

revenues and margins. You care about the

price, right? I sold at about 300. I

know at least at the start of this

month, it was close to 500. So, I mean,

I've left $200 in the table.

>> Professor, when it comes to AI, how much

do you buy the narrative that it is

going to improve margins and you know,

regardless of any company, but

>> collectively across all companies? No,

it'll reduce margins.

>> If everybody has it, nobody has it.

Which means that if I can give you as a

grocery store, if I can give you an AI

product that actually makes it easier

for your customers to buy stuff, you're

saying, "This is good. It or to cut

costs on inventory, saying, "This is

good. My costs are lower. I should have

higher margins." Here's what you're

missing. I'm selling that same product

down the street to your competing

grocerers. They're all lowering costs.

And you get the what happened in online

retailing. If everybody's cost structure

goes down, you don't end up with higher

margins. You end up with lower prices.

That's what competition does. And you

have lower prices. Everybody walks away

with lower cost structures but also

lower margins. So collectively across

all companies, I don't see this as good

for profits. But there will be

individual companies that make money

from selling you the products. They will

make money. But on the companies buying

these AI products and services, I'm not

upbeat about that converting into higher

margins and higher cash flows. So I'm

just taking what you said that every

company or most companies their costs

will go down but their margins will go

down is a consequence of that logically

that the

>> consumers will benefit right I mean who

are the biggest beneficiaries in the

online retailing

>> it was you and I as consumers because we

were able to get a choice and a cost we

would never have been able to get before

online retailing came along. So be good

for us collectively as a society but

might not be great for the businesses

that deliver us

>> and deflationary

>> that I don't know because that's going

to be a function of fiscal policy. So

there's so many other things that go

into it. So the fact that you have lower

prices on individual products and

services doesn't mean that the rest of

your life isn't is is being controlled.

So it might be deflationary for some

segments of what you do. Right? So when

I say lower prices, not that your

eggplant is cheaper than it is today,

because somebody's got to grow the

eggplant and get it. So the cost savings

are not going to be that the eggplants

are cheaper to the groceryer. It's once

the eggplants come in, AI allows them to

keep the inventory more. So the costs

you're controlling are at the margin in

the store, not the cost of the inputs

that come in. If those inputs go up, you

can still have double digit inflation in

a world where your margins are lower,

right? you're charging the lower price

than you would have in a in a in a

different setting, but that's neither

here nor there.

>> And so a year ago, you thought the

market was overvalued. And with the

earnings up 12% and the S&P up roughly a

little bit more than that, the valuation

is a similar story. So still overvalued,

but not grotesqually more overvalued

than it was last year.

>> It's not a I mean, it's not bubble

overvalued. it's overvalued like it's

been for pretty much of the la I mean

you could you could tell the same story

at the start of every year for the last

10 which should be a cautionary note

because if you view that overvaluation

as a signal you should get out of stocks

in hindsight being a terrible signal so

if you're going to make an argument that

it's a bubble that's a different

argument but I don't see it in the

numbers

>> do you think that this double digit

earnings growth is sustainable I think

the current analyst estimates of growth

you have for 2026 is around 10%.

>> You know what? It's been double digit

for the last 15 years on average. This

is not an aberration, right? There's

something that's changed in the

economics of business in the century.

And I think there are two things. One is

companies seem to be able to deliver

earnings through good times and bad

times through crisis. Things that used

to bring earnings down in the last

century, a recession, a slowing economy,

so don't seem to have the same

consequences. The other is the overall

level of earnings. Individual companies

see earnings going down have actually

become more stable in this century than

they were in the last one.

And if you're wondering why, there are

two reasons that used to be given. One

is globalization that companies even if

the domestic economy were doing badly

were able to to get their earnings from

overseas. But this year that was put to

the test because if your if that was the

reason for earnings being solid and

stable, this should be the year that

came apart because of tariffs and the

global trade wars that have so I think

after this year I'm not sure

globalization is the answer. The other

is the S&P 500 has become an

increasingly techdinated index. 30% of

the market gap comes from tech

companies. you're saying so what tech

companies have be are much more flexible

much more adept at dealing with things

changing around them and that could

explain why they're able to keep the

earnings going even during crisis and

the second is the biggest tech companies

that account for a big chunk of the

earnings are money machines their

earnings are actually not just high but

incredibly predictable and stable

because they dominate their markets so

what they're talking about Google or

Facebook you look at their earnings from

advertising ing incredibly resilient in

the face of what's happening around

them. So I think part of this is a

fundamental story and maybe those of us

who value the market are the ones who

are not adjusting quickly enough to a

changing world and we got to leave that

door open. So, it's entirely possible

that stocks are overvalued, but it's

also possible that a structural shift

has happened in the economy and markets

that could explain why they're priced at

24 times earnings as opposed to the 16

or 18 times earnings we thought was the

norm prior to this century.

>> And which way do you lean? You looking

at your equity risk premium ERP roughly

for January, you had it at 4.23%.

In what percentile is that historically?

>> It's probably the median. It's around

the m middle of the distribution. So the

equity risk premium is roughly the

median for the last 60 years. I mean of

the computation going back to 1960. It

is low by post 2008 standards. Post 2008

we saw a jump in the equity risk premium

partly because of the 2008 crisis and

these rolling crises since. So this is

low if you frame it against 2008

numbers. It's pretty much what you'd

expect if you frame it against a much

longer time period. And this is why you

can have this wild debate where

everybody looks at the same numbers and

comes to very different conclusions

about where the market is. The one group

of people that I'm not sure what they're

looking at to make judgments are people

who essentially throw the word bubble

into the mix, but don't have anything to

back it up. So, if you're going to make

an argument for a bubble, it's got to be

based on something that I'm not seeing

in the numbers right now.

>> That's that's interesting. So you talked

about the three Ps possible, plausible,

and probable. The is it sounds like you

don't think a bubble is probable. Is it

plausible? Is it possible?

>> Of course, a bubble is always plausible.

It's been true for every year of the

stock market, right? There is always a

pathway for stocks to collapse. That's

why we demand an equity risk premium. So

that's a very weak test to apply on the

market. So, I think that a bubble is

always plausible, but to invest on that

plausible, in this case, not invest

based on that plausible. Keep your money

out of stocks to me is a horrendous

mistake to make because in hindsight,

you're going to end up with returns that

lag somebody who doesn't try to time the

market, goes through the bubble, loses

money, and comes back again. So, I think

that it's not the statement that your

market's in a bubble that bothers me.

It's what you do in reaction to it

that can create cost for you in the long

term.

>> I suppose people saying it's a bubble,

they may rest their case just on the uh

price to earnings ratio where I think

the PE right now is the highest it's

been saved for maybe 99 and

>> this is about the laziest justification

for a bubble. Chpt could do better than

that. If this is your basis for your

expertise and market timing, it's time

for you to find a different profession.

I mean, come on. I mean, that's like a I

I'm me going into a doctor's office and

they're taking my temperature and

essentially passing my entire judgment

on my health based on my temperature.

It's a very rough metric. Price earnings

ratio is the most volatile and messy of

all pricing ratios. to make a judgment

about the entire market just because a

PE is higher than the average. I I mean

I I don't even know what to tell you

because I mean that's I mean let's face

it a gentleman won a Nobel Prize

building on that, right? Robert Schiller

basically built an entire story about

market bubbles built on the Schiller PE.

I mean sometimes you get accidental no

bells I guess because you know in

hindsight even he accepts that you're

missing a lot of stuff when you look at

the piece. So he actually has changed

his tune but there are a whole bunch of

people who still put their hang their

hats on chiller PS and that's their

entire basis. No. So if you want to make

an argument for markets being

overvalued, make it a more reasoned one

that goes past the PE ratio and brings

in, you know, know if your story is that

margins are at all-time highs, that

competition is coming, that AI is going

to reduce margins, and that I I'm

willing to listen, that's a story built

on fundamentals, but a PE ratio story,

I'm sorry, there's nothing there. When

you you look at the equity risk premium

which at the peak of the dotcom bubble I

believe was in the low twos maybe high

ones you you tell me the equity risk

premium is now double that is the

current equity risk premium being higher

than that is is that because the

expected earnings growth is higher or is

it because

>> it's got nothing to do it's very little

to do with earnings growth it's got

everything to do with the fact that the

cash being returned by companies in the

S&P 500 is so I mean last year there was

more more than a trillion dollars in

buybacks and this is not unique right

this has been going on for a while you

add that to the dividends collectively

US companies are returning about 85% of

their earnings in cash and if you add

that on to the dividend your cash yield

the dividend yield looks like it's at a

historic low right one and a half to 2%

but we added the buybacks you're looking

at a 4% cash yield and that's well

within the norm of what companies return

as cash to their investors. [snorts]

>> So, it's the cash flows that are holding

it up.

>> And so, the DCF that you prefer looks

not at EPS or or net income, but at cash

flows, which is dividend.

>> No, wait, wait. You can't do a DCF with

net income. Net income is not cash flow,

right? So, to begin with, if you do a

DCF and use earnings, you've already

broken a fundamental first principle.

It's got to be a cash flow. The choice

is do you want to go with the

traditional definition of cash flow

which is dividends a dividend discount

model or do you want to go with this

augmented version of cash flow which

includes buybacks with a dividend

discount model I'll save you the trouble

the market is overvalued by 80%. It's a

bubble

but that misses the fact that US

companies have shifted away from

dividends to buybacks. It's a cash

return policy and if you added the c the

buybacks then you get more reasonable

cash flows and that's at the basis of

the 4.2% premium. In fact, I computed

the implied equity risk premium using

just dividends. It's about 1.5%.

So if you focus just on dividends,

you've been out of stocks now for 15

years waiting for dividends to come

back. I'll make a prediction. You will

be waiting for dividends to come back

for a really long time because they're

not coming back. I think buybacks are

the way in which companies will turn. I

think of them as flexible dividends and

I think they actually make more sense

from an equity cash flow perspective

than these fixed dividends that you get

locked into in good times and in bad

times.

>> Why is dividends plus buybacks a better

metric for valuation than earnings?

>> Earnings are not cash flows. It's not

even in the in the competition. You can

look at a PE ratio, but you're doing a

discounted cash. You can't take the

earnings out of a company, [snorts]

right? I mean, it's it's you you've got

to bring in all these adjustments to

make it into cash flows, including how

much you have in depreciation, non-cash

expenses, you subtract. That's the big

big difference, right? You can't return.

So, any investor who's discounting

earnings, just stop.

>> Don't even try to do DCF. If that's the

first principle, you're failing, then

stick with PE ratios. There's nothing

wrong with pricing. I'd rather that you

do an honest pricing than these messy,

dishonest DCFs. And a lot of DCFs that I

see out there really shouldn't even be

in the in the in the ring because they

just break down on fundamental first

principles.

>> What about a DCF with free free cash

flow,

>> which is what I mean ultimately that's

basically where dividends and buybacks

come from, right? That's basically what

you're trying to estimate. Now you could

take individual companies and estimate

free cash flows. There are two problems

with free cash flows. One is that they

swing a lot from year to year, right?

Because they don't, you know, you don't

smooth out things like capex and so you

can have big plus years, big minus

years. And the second is computing free

cash flows for financial service

companies which remain a pretty

significant segment of the S&P 500 is

impossible to do. Impossible. Why?

Because what's capex for a bank? What's

working capital for a bank? Right?

Right? So you end up with this messy

definition of So we're using dividends

and buybacks not because we think that

they're the the right they're they're a

proxy for free cash flows because

ultimately we're assuming that's where

the cash flows to buy back stock and

dividends come from. But it's got to be

free cash flow equity. Again don't you

know you can never stop free cash flow

without specifying to equity or to the

firm. Equity is after debt payments to

the firm is before debt payments. So you

can use free cash equity but it's really

messy to do that for 500 companies many

of which have are financial service

companies or some of which are financial

service companies

>> and the the big companies in the market

I think with the exception of probably

Tesla the vast majority of their

earnings has have gone up a a lot over

the past year. What about free cash

flow? A because free free cash flow a

negative into free cash flow is capital

expenditure. A lot Microsoft, Meta, all

of these companies are spending a

tremendous amount on capital

expenditure. A lot of which goes to

Nvidia as revenue and profit. Are you

you have concerns there? I I believe you

know

>> the concern. It's not about what they're

spending. It's about whether they'll get

the returns. I mean there's we want

these companies to grow, right? You

can't have your cake and eat it too. You

want these companies to grow. If you

want them to go, they have to reinvest.

So understand the zeal they have to be

in the next big space which say CSAI.

The challenge there is they're investing

this upfront on a hope and potential.

There's really nothing on the ground

right now right now that you can use to

justify the investment. So you're an

investor in those companies. It's not

the capex per se that should bother you.

It's whether that capex will pay off as

higher earnings in the future. And

that's the big unknown. And there I

think that Apple is probably in a better

shape than the companies that are

jumping in with both feet and investing

tens of billions of dollars because we

really don't know yet what form the AI

business is going to take and whether we

need these huge investments in chips and

data centers and and power and water to

kind of produce things that might be

producible at a much lower cost with

much less technology.

[snorts]

And how does the decline in free cash

flow impact your valuation analysis?

Because you know Oracle I think I still

think their dividend buybacks still

looks pretty good.

>> I think this is ane I mean you're using

five companies which are big meta

investors. There are 500 companies the

S&P 500. The collective cash flow last

year was 7% or maybe even 10% higher

than it was the previous year. So, it's

not affecting cash return yet. Maybe it

will start to show up, but we're not

seeing it in the numbers yet. Perhaps we

did a true free cash flow last year,

you'd get a much lower number than we

got what it got as dividends and

buybacks. That by itself doesn't bother

me because that's that's the upfront

investment. If they keep investing these

amounts, I think it's eventually going

to hit the cash flows through the they

they don't have the cash to do the

buybacks anymore. And it'll get even

worse if they borrow the money to make

these investments because they create

risks that they that they historically

haven't faced that's allowed them to

deliver these nice stable earnings.

>> Professor, if if I can, Professor, what

do you think about gold? How do you

value it?

>> It's what you go to when you lose trust.

And we're in a world where people have

lost trust. And there's a segment of the

market that believes we are heading for

a cliff. It's it's your it's your

insurance against catastrophe. So I

think the fact that that segment is

bigger than it's been historically and

includes people who historically would

never have been in there. I heard Ray

Dalio say that he might I mean 20 years

ago that would not have been the case.

So there are people in that crowd who

are unusual members of that crowd. So I

think that that should send us all a

message about the potential for

catastrophe that it's higher now than

it's been historically for a variety of

reasons.

and that you might need to buy some

insurance. Whether that takes a form of

gold or not, I will leave in your hand.

>> Professor, thank you very much for for

joining us. We will link to your to your

ex and your YouTube channel as as well

as your website. Thank you everyone for

watching. Please leave a rating and

review on Monetary Matters for Apple and

Spotify. Thank you.

>> Take care.

>> Thanks for watching. Interested in the

Fundrise Income Fund? Click the link in

the description to learn more. Until

next time.