AI infrastructure / hyperscaler capex basket — what the besties said

Where they stand now

Chamath argues AI capex spending of $1+ trillion has no near-term revenue justification — only toy apps to show for it — and predicts a reckoning for startups with inflated valuations, with mismatches between spend and ROI likely playing out over years not quarters.

E187 · Jul 12, 20241 call ↓

Sacks

Commentary

Sacks is more bullish than his co-hosts on AI infrastructure investment, drawing historical parallels to broadband and railroad buildouts where initial bubbles were followed by justified long-term value creation, arguing the arms race nature of GPU procurement and early real revenue from OpenAI validate the spend.

E187 · Jul 12, 20241 call ↓

Friedberg

Commentary

Friedberg acknowledges the AI bubble is partially bursting for public market tech stocks in the near term due to unproven ROI, but remains optimistic that improving model economics over the next 24-36 months will ultimately justify the infrastructure buildout, with the key risk being whether payback comes before the next capex cycle.

E187 · Jul 12, 20241 call ↓

The discussion

The hosts are divided on AI infrastructure/hyperscaler capex, with no clear consensus. Chamath is the most bearish, arguing at high conviction that over $1 trillion in AI spending is unjustified given only nascent, "toy" applications to show for it, and warns of a multi-year reckoning for inflated valuations. Friedberg holds a mixed view, conceding a near-term mini-bubble is deflating in public markets due to unproven ROI, but remaining cautiously optimistic that model economics will improve enough over the next 24–36 months to eventually justify the buildout. Sacks is the most bullish, drawing historical analogies to the broadband and railroad overbuild cycles to argue that today's GPU arms race and early real revenue from players like OpenAI validate the long-term investment thesis, even if a short-term correction occurs.

How they got there

Chamath1 mention since Jul 12, 2024

NegativeE187Jul 12, 2024⚑

Chamath argues AI capex spending of $1+ trillion has no near-term revenue justification — only toy apps to show for it — and predicts a reckoning for startups with inflated valuations, with mismatches between spend and ROI likely playing out over years not quarters.

you can't spend a trillion dollars and only have toy apps to show for it. So I think we've all been amazed by these AI demos. I think the four of us have talked pretty incessantly about it now for 18+ months ever since we saw the first Open”26:25

Sacks1 mention since Jul 12, 2024

PositiveE187Jul 12, 2024⚑

Sacks is more bullish than his co-hosts on AI infrastructure investment, drawing historical parallels to broadband and railroad buildouts where initial bubbles were followed by justified long-term value creation, arguing the arms race nature of GPU procurement and early real revenue from OpenAI validate the spend.

I'm much more bullish than you guys about this investment that's being made. Remember that when the internet got started in the '90s, it was via dial-up. I mean, you literally had to have a modem and you would dial up the internet and it wa”33:18

Friedberg1 mention since Jul 12, 2024

MixedE187Jul 12, 2024⚑

Friedberg acknowledges the AI bubble is partially bursting for public market tech stocks in the near term due to unproven ROI, but remains optimistic that improving model economics over the next 24-36 months will ultimately justify the infrastructure buildout, with the key risk being whether payback comes before the next capex cycle.

perhaps the first AI mini bubble is bursting a bit, and particularly with respect to the accelerated expectations that public market investors had for public market technology stocks, that perhaps now is the time for a bit of a reckoning,”23:49

iAbout these quotes

Quotes are machine-transcribed from the episode audio — use the Listen links to verify any take against the source, or the ⚑ link to report a problem. Takes marked unverified, low-conviction, or commentary-only never move stances, the index, or the funds.