SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Strategies & Market Trends : Ride the Tiger with CD
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
From: Gib Bogle8/24/2025 1:05:18 PM
1 Recommendation

Recommended By
onepath

   of 312653
 
OT: Kuppy is sceptical about AI, as am I:

(from here: uraniuminvestor.substack.com

In this substack I will summarise Harris Kupperman’s key points in a nearly 50 minute long interview he did with the market huddle. Here are the key points:

1)Investing into data centers are a terrible investment and will fail. Despite the huge growth in AI demand, there is too much investment so it still wouldn’t make sense. AI demand would soar but just not as much as the market expects.

2)Investing into data centers is like investing into global crossing which decided to invest for the growth of the internet very early on but they overestimated the demand for fiber despite the internet being very huge today. The story shows even during huge demand increases you need to also look at the supply side.

3)AI will do well but valuations are too high and unjustified leading to low or negative returns- better to invest elsewhere.

4)Telecom spending peaked in inflation adjusted dollars in the early 2000s despite huge growth in demand.

5)AI companies have no moat just like telecom companies- customers will just switch to competitors if they get better.

6)Data centers cost $400 billion a year. Over a 10 year depreciation curve they need $40 billion a year. If gross margin is 25% at least $160 billion a year needed to break even. This is super unrealistic since very profitable subscription services that also have a lot of customers do not make near enough like Microsoft Office which makes about $100 billion and Netflix at about $30 billion (each per year.)

7)If you also include a 20% return on capital needed to make the investment make sense in risk adjusted terms. Next year could be doubling the pace of build out. Doing all the math about $1.5 trillion a year in revenue is needed which is super unrealistic.

8)Some even assume a 5 year depreciation curve due to the pace of changes in AI meaning even more revenue is needed to make the investment worthwhile.

9)If we assume a 5 year depreciation curve $3 trillion a year is needed which is 10% of global economy which is super unrealistic.

10)AI is like railroads or restaurants you see them making huge amounts of money but it’s usually not the guy founding the company making the profits. It’s usually the guy buying it up for pennies on the dollar later on that makes the money.

11)AI spending is artificially making US economic growth seem stronger than it is and likely represents over half of economic growth.

12)Wealth effect is also artificially making the economy seem stronger than it will be (and according to Kuppy, the wealth affect matters more than fiscal or monetary policy) when valuations reach proper intrinsic amounts. Both of these things are distorting the economy.

13)Outside of AI, include wealth effect, multipliers, downstream affects etc USA is likely in a recession. This definitely means too much is being invested into AI.

14)Wealth affect is way more important than previously due to increased wealth inequality and the fact that wealthy individual take loans against the value of their investments to consume goods.

15)prices of oil, plastics and lumber suggest demand in the “real economy” (non VC economy) is doing crap. Market is a bubble.

16)Kupperman believes USA is headed towards a recession.

17)Trump’s tariff policies are too volatile to work out as businesses would take a “wait and see” approach to decide what to do with investing.

18)Nobody knows much about exact impact tariffs will have as a longer term policy (once volatility gets reduced) so federal reserve can take pretty much any approach but leans dovish (he ended up being right- the interview was before Jackson Hole.)

19)In the longer term interest rates will go much higher because they USA has a similiar fiscal policies to African policies so should have similiar borrowing costs (we also made a similiar point to Kuppy on one of our paid posts.)

20)The increase in bond issuance will crowd out stock market investments lowering stock market valuations.

22)Overall he is bearish on AI investments as there is too much competition and the economy is doing much worse than it seems.
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  

HomeMailHotSubjectMarksPeopleMarksKeepersSettings
Terms Of UseContact UsCopyright/IP PolicyPrivacy PolicyAbout UsFAQAdvertise on SI
© 2025 Knight Sac Media.  Data provided by Twelve Data, Alpha Vantage, and CityFALCON News