Artificial Intelligence

[FiveSigma]

49 minutes ago, Dalal.Holdings said:

This is like thinking AOL would accrue the value of folks wanting access to the internet back in 2000.

 

Tell me you don't understand technology moats without telling me.

 

It's not a valid comparison and DeepSeek showed nothing. Enterprise customers are not deploying DeepSeek, they are deploying MS Copilot (integrated into the Office suite) and Google's AI Studio, not some bare LLM.

 

AI revenue growth of hyperscalers has not slowed down one bit.

 

AI software stack is not a commodity in the same sense as cable/fiber to one's home.

 

AI is being sold as a bundle (models + compute + orchestration + customizations + application APIs + support) and as such is closer to enterprise IaaS/SaaS and will have comparable future/economics.

 

The moat will once again (like with Cloud 15 years ago) come from a combination of scale, distribution, proprietary data, being the lowest cost provider (e.g., Google with TPUs), and bundling/integration into hyperscalers' other products.

Edited September 3, 2025 by FiveSigma
Fixed a typo

[Dalal.Holdings]

21 minutes ago, FiveSigma said:

 

Tell me you don't understand technology moats without telling me.

 

It's not a valid comparison and DeepSeek showed nothing. Enterprise customers are not deploying DeepSeek, they are deploying MS Copilot (integrated into the Office suite) and Google's AI Studio, not some bare LLM.

 

AI revenue growth of hyperscalers has not slowed down one bit.

 

AI software stack is not a commodity in the same sense as cable/fiber to one's home.

 

AI is being sold as a bundle (models + compute + orchestration + customizations + application APIs + support) and as such is closer to enterprise IaaS/SaaS and will have comparable future/economics.

 

The moat will once again (like with Cloud 15 years ago) come from a combination of scale, distribution, proprietary data, being the lowest cost provider (e.g., Google with TPUs), and bundling/integration into hyperscalers' other products.

 

LOL, sure tell yourself all the great stories you need to in order to justify massive investment that is unlikely to show the returns you are looking for.

 

You’re right though—they’re not like cable/fiber to one’s home. It’s even worse because it’s easier to switch AI providers than it is the cable company wired to your house (or at least it was in 2000). Even Cursor, one of the most common ways to write code with AI, allows you to choose the model you use with a simple click.

[Dalal.Holdings]

Times were great when all GOOG had to do was run a simple webpage and put ads in your face whenever you used it. The business required almost zero capital investment, it was winner-take-all, and the returns on capital were absolutely mouthwatering. 

 

The tech companies of today are not the tech companies of 10-20 years ago, but I don’t think most people realize that yet…

 

Edited September 3, 2025 by Dalal.Holdings

[widenthemoat]

1 minute ago, Dalal.Holdings said:

Times were great when all GOOG had to do was run a simple webpage and put ads in your face whenever you used it. The business required almost zero capital investment, it was winner-take-all, and the returns on capital were absolutely mouthwatering. 

 

The tech companies of today are not the tech companies of 10-20 years ago, but I don’t think most people realize that yet…

 

+1 on this, it seems very easy to switch AI providers and the economics for these companies appears to be changing dramatically (at least for now). Something else that I think is not really being taken into consideration is that regardless of whether the investment pays off or not, there is a much much wider range of outcomes for these companies now compared to the past given AI. Honestly I don't think anyone really knows how it all plays out, but there is no doubt that prior to AI the dispersion of outcomes was way lower. 

[FiveSigma]

18 minutes ago, Dalal.Holdings said:

The tech companies of today are not the tech companies of 10-20 years ago

 

Way to shift the goal posts from 'commoditized' to 'much more capital intensive'.

 

They are not, but it's arrogant, IMO, to insinuate that margins and ROIC will definitely be lower at Cloud and AI markets' maturity. The cost curves are extremely dynamic and even the foremost computing/AI experts are not making such prognostications.

 

Case in point - Google lowering power consumption for their GenAI queries by 33x in just one year!

 

CapEx side is easy to see for a value investor used to flying by 10Ks, but the revenue side requires imagination and mental exertion.

 

 

 

Edited September 3, 2025 by FiveSigma

[Dalal.Holdings]

22 minutes ago, FiveSigma said:

 

Way to shift the goal posts from 'commoditized' to 'much more capital intensive'.

 

 

Pay attention. A few posts back, I literally said in this thread:

On 8/29/2025 at 12:32 PM, Dalal.Holdings said:

 

Yep. Most of the Mag 7 are becoming worse businesses (more capital intensive) with likely lower future returns on capital. Then you also have the insane compensation (many are in “non-cash” equity so cash flows mask it) packages for people skilled in AI…

 

Mom & Pop who are shoving everything into an index fund (over 1/3 of investment in S&P now goes to Mag 7) might not be aware of that fact…

 

22 minutes ago, FiveSigma said:

 

 

CapEx side is easy to see for a value investor used to flying by 10Ks, but the revenue side requires imagination and mental exertion.

 

 

Hah yeah, good luck investing based on your "imagination" !

[Dalal.Holdings]

The numbers keep going up and these companies look less and less like software businesses and more like railroads (with the added lucrative stock based comp)...

 

I'm sure the $50-100B in annual capex will work out well !

 

They probably are like Railroads, except Railroads in the 1800s when every investor threw money at them and rail lines that ran parallel next to each other (going to the same places) were built. The returns for investors in railroads at the time (which were a great innovation and fundamentally changed society) were...not great to say the least

 

 

 

 

Edited September 3, 2025 by Dalal.Holdings

[FiveSigma]

14 minutes ago, Dalal.Holdings said:

 

Pay attention. A few posts back, I literally said in this thread:

 

 

Hah yeah, good luck investing based on your "imagination" !

 

The screenshots you posted are in no way a counterargument to my replies to your points:

1) You cannot compare AI stack economics to that of an internet cable

2) Claiming that ROIC will be lower (which you did explicitly and implictly) is arrogant even for someone with domain expertise, and a lot more so for someone without

 

Thanks for the wishes, not all investing edge is in 10Ks and the last 15 years abundantly demonstrated that, so I choose to evolve with the world.

Edited September 3, 2025 by FiveSigma

[Dalal.Holdings]

6 minutes ago, FiveSigma said:

 

The screenshots you posted are in no way a counterargument to my replies to your points:

1) You cannot compare AI stack economics to that of an internet cable

2) Claiming that ROIC will be lower (which you did explicitly and implictly) is arrogant even for someone with domain expertise, and a lot less so for someone without

 

Thanks for the wishes, not all investing edge is in 10Ks and the last 15 years abundantly demonstrated that, so I choose to evolve with the world.

 

Over the past 10 years, MSFT ROIC has hit levels north of 30%.

 

Do you contend that it is "arrogant" to say ROIC will go down as capex has risen ~8-fold to ~$65B a year? AMZN and META capex has risen about 20-fold over 10 years. What generally happens to returns as invested capital rises? Will 30% ROIC be easier or harder to achieve going fwd ?

 

Ask Warren if it was easier making high returns with smaller starting capital or larger starting capital.

 

Total Assets over 10 years:

AMZN: $65B --> $625B

MSFT: $175B --> $620B

META: $49B --> $276B

GOOG: $147B --> $450B

 

I guess if you use your imagination though, anything is possible.

 

Edited September 3, 2025 by Dalal.Holdings

[Dalal.Holdings]

The real winner here will likely be someone like Apple who has remained model agnostic and will likely just incorporate whichever model it wants...

 

Total Assets over 10 years:

AAPL: $290B --> $365B

 

AAPL capex 2015: $11.2B, ttm: $12.3B

 

Amazingly, Apple's capex has barely changed and is actually down in real terms. Its asset growth is also much lower than the other tech giants. That's because they made a decision not to build an AI model. I think it will prove to be a smart decision (though Apple's success is also tied to the longevity of its ecosystem).

 

Their ROIC has only grown over the past decade to over 50% now and will probably remain high while the others see it crumble

 

 

Edited September 3, 2025 by Dalal.Holdings

[FiveSigma]

In general - maybe.

 

In technology world - too hard, even with a great deal of domain knowledge.

 

Not every market is the same either - chances of another 200K rail miles built in the US (let alone the same ROIC) are not the same as AI becoming the next $1T market.

 

My point is it's foolhardy to throw unresearched (and observably false, like the AOL comparison) analogies the way you did.

 

Investing game has evolved and many value investors that got stuck in "Mr. Buffett" quotes and "buy low P/E" formulas have not a great time. Yet you snidely reduce my words of caution to relying on fantasy. This does seem to be the MO of this board (together with Politics section ranting), which is why a lot of thoughful posters left it for X FinTwit community.

 

Edited September 3, 2025 by FiveSigma

[Dalal.Holdings]

19 minutes ago, FiveSigma said:

 

 

Investing game has evolved and many value investors that got stuck in "Mr. Buffett" quotes and "buy low P/E" formulas have not a great time. Yet you snidely reduce my words of caution to relying on fantasy. This does seem to be the MO of this board (together with Politics section ranting), which is why a lot of thoughful posters left it for X FinTwit community.

 

 

 

Words of caution? I am not short AI or big tech stocks, so what exactly are you cautioning me against?

 

If anything, everyone and their mother is over-indexed to these big tech stocks (remember, 33% of the S&P), so the cautious thing to do is to think about what can go wrong with these stocks, not use your imagination to think the sky is the limit.

 

Everyone and their mother thinks AI is the next big thing/money maker. That is not the cautious view to take at this point.

 

Edited September 3, 2025 by Dalal.Holdings

[tnathan]

For what it's worth, I actually do think over a 10 year period the hyperscalers will make a very good return on the capex spend. Companies building models without the ability to bundle with an overall consumer + enterprise offering will suffer. And if the bubble bursts and nvda chips get cheaper, it'll just cost less to build. But i have no doubt that the market will be many times bigger than people are imagining and to win you need to be building now

[Milu]

1 hour ago, tnathan said:

For what it's worth, I actually do think over a 10 year period the hyperscalers will make a very good return on the capex spend. Companies building models without the ability to bundle with an overall consumer + enterprise offering will suffer. And if the bubble bursts and nvda chips get cheaper, it'll just cost less to build. But i have no doubt that the market will be many times bigger than people are imagining and to win you need to be building now

Yes I agree with this. I think most naysayers here will be continually surprised at how much bigger big tech gets over the next 10-20 years. 

[Milu]

4 hours ago, Dalal.Holdings said:

Everyone and their mother thinks AI is the next big thing/money maker. That is not the cautious view to take at this point.

Not fully sure about this. Everything I read seems to be along the lines that AI is a bubble and that the mag 7 are just dumping hundreds of billions into low returning cap ex. I’d say the contrarian view is that this isn’t a bubble and that the leaders of these companies are seeing great ROIC from these investments so are continuing to plough more money into it. So far we are over two years into this buildout and returns on capital for alphabet, Amazon, Microsoft, meta seem to holding up if not improving, maybe this will change at some point.

[widenthemoat]

7 hours ago, Milu said:

Not fully sure about this. Everything I read seems to be along the lines that AI is a bubble and that the mag 7 are just dumping hundreds of billions into low returning cap ex. I’d say the contrarian view is that this isn’t a bubble and that the leaders of these companies are seeing great ROIC from these investments so are continuing to plough more money into it. So far we are over two years into this buildout and returns on capital for alphabet, Amazon, Microsoft, meta seem to holding up if not improving, maybe this will change at some point.

I’m not sure I follow this - if the contrarian view is that we are not in a bubble than why are we in the bubble? Wouldn’t people sell if the consensus was we are in a bubble? A bubble can only exist if people are buying these stocks expecting huge gains or, alternatively, they just want to play the game and get out before the music stops. Is your argument that it is the latter?

Edited September 4, 2025 by widenthemoat

[treasurehunt]

7 hours ago, Milu said:

Not fully sure about this. Everything I read seems to be along the lines that AI is a bubble and that the mag 7 are just dumping hundreds of billions into low returning cap ex. I’d say the contrarian view is that this isn’t a bubble and that the leaders of these companies are seeing great ROIC from these investments so are continuing to plough more money into it. So far we are over two years into this buildout and returns on capital for alphabet, Amazon, Microsoft, meta seem to holding up if not improving, maybe this will change at some point.

 

ChatGPT was released to the public on Nov 30, 2022. From that day till yesterday, the S&P 500 had a total return of about 58%. AAPL, which one might consider the most prominent non-AI tech stock, returned just over 63%. In contrast, here's how the five biggest tech companies that are usually considered beneficiaries of AI did over the same period.

 

Ticker	Nov 30, 2022	Sep 3, 2025	Total Return
NVDA	$16.91	$170.62	909%
MSFT	$249.65	$505.35	102%
AMZN	$96.54	$225.99	134%
GOOG	$101.45	$231.10	128%
META	$117.46	$737.05	528%

 

It seems clear that the market is not pricing in enormous wasted capex at these companies.

[Milu]

18 minutes ago, treasurehunt said:

 

ChatGPT was released to the public on Nov 30, 2022. From that day till yesterday, the S&P 500 had a total return of about 58%. AAPL, which one might consider the most prominent non-AI tech stock, returned just over 63%. In contrast, here's how the five biggest tech companies that are usually considered beneficiaries of AI did over the same period.

 

Ticker	Nov 30, 2022	Sep 3, 2025	Total Return
NVDA	$16.91	$170.62	909%
MSFT	$249.65	$505.35	102%
AMZN	$96.54	$225.99	134%
GOOG	$101.45	$231.10	128%
META	$117.46	$737.05	528%

 

It seems clear that the market is not pricing in enormous wasted capex at these companies.

How do you know the cap ex is wasted? 

[FiveSigma]

12 hours ago, Dalal.Holdings said:

 

Words of caution? I am not short AI or big tech stocks, so what exactly are you cautioning me against?

 

 

Pay attention.

 

Words of caution are for you, when you say and I quote "with likely lower future returns on capital".

 

When a business can lower a key input costs by 5x in a single year (source below) and another key one (power) by 33x, it should give a prudent person pause when assuming ROIC declines.

 

You move the goal posts again, now claiming the argument is about whether AI being "the next big thing" is consensus.

 

Also, nowhere did I state that "sky is the limit".

 

I'm saying it's a very difficult question to throw uneducated takes on.

 

Source:

https://www.seroundtable.com/google-sge-ai-answers-cost-80-less-37326.html

[Milu]

54 minutes ago, widenthemoat said:

I’m not sure I follow this - if the contrarian view is that we are not in a bubble than why are we in the bubble? Wouldn’t people sell if the consensus was we are in a bubble? A bubble can only exist if people are buying these stocks expecting huge gains or, alternatively, they just want to play the game and get out before the music stops. Is your argument that it is the latter?

I don’t really believe bubbles can be called until after the fact, and even after the fact it’s tough to say. For example if tech stocks drop 80% during dot com crash I’d be comfortable saying that it was a bubble, but what if stocks drop 50%, 40%, 30%? Where is the transition point from overpriced equities correcting down to regular price vs bubble crash. Also if somebody today says that Mag 7 are in a bubble and they double in price and then crash 40% then all the bears would be proclaiming how they called the bubble. Yet in reality the post crash low is higher than the price they were shouting bubble. Alphabet, one of the best positioned stocks has been trading at 20 pe until the past few days. Meta has trailing pe of less than 27. Microsoft is quite pricey at 37 times trailing but again these are the largest most cashflow generative businesses in history. This isn’t Cisco trading at 35 times revenue like in 2000. 
 

so in summary while some pockets of the market are overvalued I don’t necessarily think we are in a bubble and seeing as most people do think we are in some sort of bubble, the contrarian view would be to hold on and not sell any msg 7 stocks as the could have a long runway ahead.

Edited September 4, 2025 by Milu

[treasurehunt]

42 minutes ago, Milu said:

How do you know the cap ex is wasted? 

 

I don't know this. In fact, I don't have a strong view about this at all. You said this in your post though: "Everything I read seems to be along the lines that AI is a bubble and that the mag 7 are just dumping hundreds of billions into low returning cap ex." I was just pointing out that the market isn't pricing in this view that apparently permeates everything you read.

[Eldad]

1 hour ago, FiveSigma said:

 

Pay attention.

 

Words of caution are for you, when you say and I quote "with likely lower future returns on capital".

 

When a business can lower a key input costs by 5x in a single year (source below) and another key one (power) by 33x, it should give a prudent person pause when assuming ROIC declines.

 

You move the goal posts again, now claiming the argument is about whether AI being "the next big thing" is consensus.

 

Also, nowhere did I state that "sky is the limit".

 

I'm saying it's a very difficult question to throw uneducated takes on.

 

Source:

https://www.seroundtable.com/google-sge-ai-answers-cost-80-less-37326.html

Where is the moat? Everyone and their uncle can churn out a LLM. If prices are coming down at such a fast rate, then isn’t there a massive risk that they are overbuilding? 

[Milu]

1 hour ago, treasurehunt said:

 

I don't know this. In fact, I don't have a strong view about this at all. You said this in your post though: "Everything I read seems to be along the lines that AI is a bubble and that the mag 7 are just dumping hundreds of billions into low returning cap ex." I was just pointing out that the market isn't pricing in this view that apparently permeates everything you read.

Maybe we are confusing each other here, maybe I'm confusing myself .

 

Is your opinion that the market is currently pricing in epic returns from the massive cap ex build out. If you look at the valuation multiples of Alphabet or Meta do they look elevated in anyway against their average multiples over the last 5 years, or last decade. If the market was pricing them in a bubbly way I'd have expected significantly higher multiples to earnings, cash flow etc, but I don't see anything different to valuations before this AI capex ramped up. I can't speak to Nvidia or Microsoft as I don't own those stocks but maybe they are bubblier.

 

I'm not sure whether we are agreeing with each other or not.

 

Edited September 4, 2025 by Milu

[Eldad]

3 minutes ago, Milu said:

Maybe we are confusing each other here, maybe I'm confusing myself .

 

Is your opinion that the market is currently pricing in epic returns from the massive cap ex build out. If you look at the valuation multiples of Alphabet or Meta do they look elevated in anyway against their average multiples over the last 5 years, or last decade. If the market was pricing them in a bubbly way I'd have expected significantly higher multiples to earnings, cash flow etc, but I don't see anything different to valuations before this AI capex ramped up. I can't speak to Nvidia or Microsoft as I don't own those stocks but maybe they are bubblier.

 

I'm not sure whether we are agreeing with each other or not.

 

PEs are maybe similar but price to FCF is massively higher and every year going forward you get a bigger and bigger depreciation hit from the servers. The Capex is effectively lowering future earnings while the profit side is still in doubt. 

[Eldad]

MSFT

10 year average P/FCF 31x

Today 52x

 

META

10 Year 31x

Today 39x

 

GOOGL

10 year 30x

Today 42x

 

AMZN

10 year 60x

Today 186x 

 

Amzn already went down this road when for some strange reason they started buying warehouses, trucks, and grocery stores, but now they have gone to another level.