A fun question. In 5 years time, what looks better? The US’s enormous bet & capex on AI? Or China’s equally enormous bet and capex on renewables?
China has made cheap, clean energy available in huge quantities. The world should take the win econ.st/4oqFszB
Photo: Eyevine
Photo: Eyevine
November 7, 2025 at 7:10 AM
A fun question. In 5 years time, what looks better? The US’s enormous bet & capex on AI? Or China’s equally enormous bet and capex on renewables?
Interesting to see GOLDMAN — which has been pretty constructive on the sustainability of the AI/capex spend — start to fold in some caution:
“.. history suggests that today’s tight credit spreads may be vulnerable not just to downside risks to the economy
but to increased use of debt ..” 👀
“.. history suggests that today’s tight credit spreads may be vulnerable not just to downside risks to the economy
but to increased use of debt ..” 👀
November 10, 2025 at 4:47 PM
Interesting to see GOLDMAN — which has been pretty constructive on the sustainability of the AI/capex spend — start to fold in some caution:
“.. history suggests that today’s tight credit spreads may be vulnerable not just to downside risks to the economy
but to increased use of debt ..” 👀
“.. history suggests that today’s tight credit spreads may be vulnerable not just to downside risks to the economy
but to increased use of debt ..” 👀
Interesting thing I read today is that Anthropic is taking the lead in enterprise AI subscriptions and in a capex heavy business, that stability of revenue is more important than having more subscribers.
November 11, 2025 at 2:49 PM
Interesting thing I read today is that Anthropic is taking the lead in enterprise AI subscriptions and in a capex heavy business, that stability of revenue is more important than having more subscribers.
Michael Burry has some concerns about AI accounting sherwood.news/markets/mich...
My TL;DR: it’s 50% scaremongering, 50% a good launching pad for some potential issues with AI capex and ROI.
My TL;DR: it’s 50% scaremongering, 50% a good launching pad for some potential issues with AI capex and ROI.
November 11, 2025 at 7:36 PM
Michael Burry has some concerns about AI accounting sherwood.news/markets/mich...
My TL;DR: it’s 50% scaremongering, 50% a good launching pad for some potential issues with AI capex and ROI.
My TL;DR: it’s 50% scaremongering, 50% a good launching pad for some potential issues with AI capex and ROI.
“.. META seems to have finally crossed the line of CAPEX (and OPEX) too high along sort of a ho-hum quarter.”
- JPM desk, this week
- JPM desk, this week
BREAKING 🚨: Meta
$META has now plunged more than 17% since last week's earnings report 📉📉 Thanks a lot Zuck 🫡
$META has now plunged more than 17% since last week's earnings report 📉📉 Thanks a lot Zuck 🫡
November 6, 2025 at 9:39 PM
“.. META seems to have finally crossed the line of CAPEX (and OPEX) too high along sort of a ho-hum quarter.”
- JPM desk, this week
- JPM desk, this week
This is an impossible market.
It's very unlikely any AI companies will see meaningful revenue growth with a clear ability to scale in any space other than AI coding tools in the near future.
This will never be able to justify the insane capex spending of big tech and investors.
It's very unlikely any AI companies will see meaningful revenue growth with a clear ability to scale in any space other than AI coding tools in the near future.
This will never be able to justify the insane capex spending of big tech and investors.
November 10, 2025 at 10:46 PM
This is an impossible market.
It's very unlikely any AI companies will see meaningful revenue growth with a clear ability to scale in any space other than AI coding tools in the near future.
This will never be able to justify the insane capex spending of big tech and investors.
It's very unlikely any AI companies will see meaningful revenue growth with a clear ability to scale in any space other than AI coding tools in the near future.
This will never be able to justify the insane capex spending of big tech and investors.
Actually, most of the commentary I see from academics deplores AI as one might deplore Hallmark movies (if romcom capex was ~2% of U.S. GDP). It's creating a community of affect, but not making a policy argument that needs to be engaged.
Every debate about AI on here about whether it is good or evil, verboten or not, presumes it can be prevented, and that seems so out of touch I don't even feel like reading to the end of a post, even the rebuttals.
November 9, 2025 at 12:52 PM
Actually, most of the commentary I see from academics deplores AI as one might deplore Hallmark movies (if romcom capex was ~2% of U.S. GDP). It's creating a community of affect, but not making a policy argument that needs to be engaged.
AAPL and META last few weeks really illustrates why i view markets through a narrative prism and not a valuation prism. market has shifted to questioning insane capex spend without any real ROI
November 11, 2025 at 8:10 PM
AAPL and META last few weeks really illustrates why i view markets through a narrative prism and not a valuation prism. market has shifted to questioning insane capex spend without any real ROI
This is British capex, Bernard
November 7, 2025 at 8:54 AM
This is British capex, Bernard
TSLA spending $100bn of capex to try and build a leading-edge node semi foundry would be incredible entertainment
November 6, 2025 at 10:42 PM
TSLA spending $100bn of capex to try and build a leading-edge node semi foundry would be incredible entertainment
“Massively ramping capex through purchase of Nvidia chips/servers on a 2-3 yr product cycle should not result in the extension of useful lives of compute equipment. Yet this is exactly what all the hyperscalers have done.”
'Big Short' investor Michael Burry accuses AI hyperscalers of artificially boosting earnings
Burry alleged that "hyperscalers" are artificially boosting earnings by extending the useful life of their computer equipment.
www.cnbc.com
November 11, 2025 at 7:17 PM
“Massively ramping capex through purchase of Nvidia chips/servers on a 2-3 yr product cycle should not result in the extension of useful lives of compute equipment. Yet this is exactly what all the hyperscalers have done.”
It's kinda wild how, decades into this fairly mature industry where 8 figure capex for a project is commonplace and 9 figure capex for a project isn't all that remarkable, outcomes suggest that project management is still completely un-figured-out.
Grand Theft Auto VI delayed to Nov 19, 2026 www.rockstargames.com/newswire/art...
November 6, 2025 at 10:03 PM
It's kinda wild how, decades into this fairly mature industry where 8 figure capex for a project is commonplace and 9 figure capex for a project isn't all that remarkable, outcomes suggest that project management is still completely un-figured-out.
The likelihood of AI CAPEX forcing Oracle to default on its loans is very low. But it’s no longer extremely low.
November 7, 2025 at 5:09 PM
The likelihood of AI CAPEX forcing Oracle to default on its loans is very low. But it’s no longer extremely low.
AI boosters are beginning to crash out
November 6, 2025 at 5:04 PM
AI boosters are beginning to crash out
Simpler than that. It's the debt.
The talking point was that hyperscalers paid for CapEx out of cashflow. It had been clear for a while Oracle overextended. Then, META (not even an LLM player) shows up with $150b in off-balance sheet debt and lots of IG paper. Hyperscalers issued $75b Sep-Oct alone
The talking point was that hyperscalers paid for CapEx out of cashflow. It had been clear for a while Oracle overextended. Then, META (not even an LLM player) shows up with $150b in off-balance sheet debt and lots of IG paper. Hyperscalers issued $75b Sep-Oct alone
November 9, 2025 at 12:23 PM
Simpler than that. It's the debt.
The talking point was that hyperscalers paid for CapEx out of cashflow. It had been clear for a while Oracle overextended. Then, META (not even an LLM player) shows up with $150b in off-balance sheet debt and lots of IG paper. Hyperscalers issued $75b Sep-Oct alone
The talking point was that hyperscalers paid for CapEx out of cashflow. It had been clear for a while Oracle overextended. Then, META (not even an LLM player) shows up with $150b in off-balance sheet debt and lots of IG paper. Hyperscalers issued $75b Sep-Oct alone
It continues to amaze me how the money taps for capex are limitless, but we are too stingy to actually operate the infrastructure we have built
November 7, 2025 at 7:18 PM
It continues to amaze me how the money taps for capex are limitless, but we are too stingy to actually operate the infrastructure we have built
Burry, Berkshire e Universa vedono profilarsi (non si sa con che orizzonte) una crisi. Le valutazioni sono estreme, la compiacenza è quasi totale (Move), i fondamentali sono sospetti (Capex) e i profitti potrebbero essere gonfiati.
Il ritorno alla realtà sarà doloroso. Per quasi tutti.
11/11
Il ritorno alla realtà sarà doloroso. Per quasi tutti.
11/11
November 11, 2025 at 3:57 PM
Burry, Berkshire e Universa vedono profilarsi (non si sa con che orizzonte) una crisi. Le valutazioni sono estreme, la compiacenza è quasi totale (Move), i fondamentali sono sospetti (Capex) e i profitti potrebbero essere gonfiati.
Il ritorno alla realtà sarà doloroso. Per quasi tutti.
11/11
Il ritorno alla realtà sarà doloroso. Per quasi tutti.
11/11
what I really hate about it is the scale, NIH/NSF funded orders of magnitude than what Europe and Canada are pledging
the only winner here is China, I’m getting tons of interest from a lot of firms based there because their Capex is nuts compared to the US shooting itself in the dick because MAGA
the only winner here is China, I’m getting tons of interest from a lot of firms based there because their Capex is nuts compared to the US shooting itself in the dick because MAGA
November 12, 2025 at 7:54 AM
what I really hate about it is the scale, NIH/NSF funded orders of magnitude than what Europe and Canada are pledging
the only winner here is China, I’m getting tons of interest from a lot of firms based there because their Capex is nuts compared to the US shooting itself in the dick because MAGA
the only winner here is China, I’m getting tons of interest from a lot of firms based there because their Capex is nuts compared to the US shooting itself in the dick because MAGA
When or if the AI bubble deflates the driver will probably be the customer being bored by something the technologists adore. It's like the very clever stylus in my phone that I have to force myself to use. Except with trillions in capex... 3/
November 7, 2025 at 7:07 AM
When or if the AI bubble deflates the driver will probably be the customer being bored by something the technologists adore. It's like the very clever stylus in my phone that I have to force myself to use. Except with trillions in capex... 3/
JPM DESK: “.. Getting a lot of Qs on $MSFT weakness. .. increasingly a source of funds with Azure growth a bit less inspiring
“.. We continue to see rotations into $AMZN and $GOOGL ..
..” $META seems to have finally crossed the line of CAPEX (and OPEX) too high along sort of a ho-hum quarter.”
“.. We continue to see rotations into $AMZN and $GOOGL ..
..” $META seems to have finally crossed the line of CAPEX (and OPEX) too high along sort of a ho-hum quarter.”
November 5, 2025 at 11:15 PM
JPM DESK: “.. Getting a lot of Qs on $MSFT weakness. .. increasingly a source of funds with Azure growth a bit less inspiring
“.. We continue to see rotations into $AMZN and $GOOGL ..
..” $META seems to have finally crossed the line of CAPEX (and OPEX) too high along sort of a ho-hum quarter.”
“.. We continue to see rotations into $AMZN and $GOOGL ..
..” $META seems to have finally crossed the line of CAPEX (and OPEX) too high along sort of a ho-hum quarter.”
Sam really is remarkably deceptive. He pretends to answer the question of how he is going to pay for his gigantic CAPEX spend by talking about OpenAI revenue. But he loses money on every dollar of revenue. So how is that revenue paying for even a ham sandwich?
November 8, 2025 at 2:35 AM
Sam really is remarkably deceptive. He pretends to answer the question of how he is going to pay for his gigantic CAPEX spend by talking about OpenAI revenue. But he loses money on every dollar of revenue. So how is that revenue paying for even a ham sandwich?
The layoffs continue with no hiring, but the stock market... well, they're finally starting to realize that Sam "hey, can you like, guarantee you'll pay for these data centers" Altman and the rest of the AI bros might be building something that will never pay for its CapEx never mind profit?
November 7, 2025 at 2:17 PM
The layoffs continue with no hiring, but the stock market... well, they're finally starting to realize that Sam "hey, can you like, guarantee you'll pay for these data centers" Altman and the rest of the AI bros might be building something that will never pay for its CapEx never mind profit?
An important truism - delays in getting revenue from CapEx quickly screw your ROI.
November 10, 2025 at 2:49 PM
An important truism - delays in getting revenue from CapEx quickly screw your ROI.
UBS: "we now forecast global AI capex to reach USD 423bn this year (from our previous estimate of USD 375bn) and USD 571bn in 2026 (from USD 500bn). By 2030, we expect overall spending to hit USD 1.3tr"
November 5, 2025 at 12:52 PM
UBS: "we now forecast global AI capex to reach USD 423bn this year (from our previous estimate of USD 375bn) and USD 571bn in 2026 (from USD 500bn). By 2030, we expect overall spending to hit USD 1.3tr"