The alpha, twice a week
That was 3 min.
Get the best new moments in your inbox every Wednesday and Saturday. It's free.
Check your inbox
We sent a link to your address. Click it to finish signing up. Not there in a minute? Look in spam.
or
Prof G Markets · 3 min · Oct 27, 2026
A correction could hit chipmakers harder than hyperscalers
Robert Armstrong argues a pullback in AI infrastructure spending could hurt semiconductor companies while the cash-rich hyperscalers remain resilient.
The episode’s best minutes
Full analysis- 8:16Anthropic’s future compute costs dwarf its current losses9 min
- 30:11A correction could hit chipmakers harder than hyperscalersPro pick 3 min
- 3 more moments in the full episode
“even if inflation does go up a little bit and rates don't go up a little bit, you have a little cushion in there.”
Takeaways Pro pick
- A downturn in data-center spending could sharply hit chipmakers.
- Hyperscalers may benefit if they can redirect cash to shareholders.
- Treasury yields near 5% provide a cushion against some risks.
One of Pro's picks, open to you here. Try Pro free for 7 days
One of Pro's picks. The rest of the summary and the 3 takeaways are in Real Alpha Pro; the video, the quote and the transcript stay free.
Read it with Pro, 7 days freeAbout this segment
Robert Armstrong argues a pullback in AI infrastructure spending could hurt semiconductor companies while the cash-rich hyperscalers remain resilient. He also explains why Treasury yields near 5% offer investors some protection.
US Financial Commentator, Financial Times
Ed Elson, Host
From “Anthropic's $800 Billion Problem Is Hiding In The Fine Print”, Oct 27, 2026. Watch the full episode
Thanks. We'll look at and fix what's off.
Transcript
Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.
- 30:23
Let me start by taking the other side of the argument. One thing that if you're going to have a crazy investment euphoria, it's good to start with some companies who have a lot of money and have really good businesses. And so there's five, four or five of the hyperscalers that whatever happens to AI, they went into the situation having the most profitable businesses in the history of the world. Alphabet, Amazon, Meta, Microsoft, Apple. Apple's not participating, so take them off the list. But anyway, so you've got these cash machines at one end, and that's really good.
- 30:59
and to the degree that the ai data center bubble starts to deflate investors might look at those companies and say great they don't have to pay for this crap anymore they can just go back to churning out free cash flow buying back their own shares and paying dividends instead of sinking every penny into a data center down in louisiana right and and those stocks might go up that doesn't I mean, CoreWeave's not going up.
- 31:26
CoreWeave's going to zero under that circumstance. Yeah, but who cares? But who cares? So, like, the big pillars of the stock market might stay standing under that circumstance. The circumstance being, like, holy crap, we don't actually want to finance this much capacity for AI data centers. A lot of people will feel the pain, but those pillars might stay standing. But you do have another set of companies that have become incredibly important.
- 31:54
won't fare so well and that's the semiconductor complex NVIDIA broadcom and work work your way right down right those things would lose a massive amount of value right and that would be a shock right and uh a less biggest shock than if it happened to microsoft but uh you know NVIDIA right Now, Microsoft and the other hyperscalers, but NVIDIA right now is the most valuable company in the world still today, I think. And so that would be a nasty shock for those guys. So would it be a catastrophic 2008 style correction? My guess is no, because of the hyperscalers would probably do okay. Would there be some smoldering wreckage to pick through? Yes. Will you be glad you owned some 10-year treasuries with a 5% yield going in?
- 32:47
Probably yes again, you know? Certainly. And by the way, the nice thing about the five-year, the 10-year treasury, if I might just return to that topic here briefly. Yep. With a 5% yield on it, and you can get almost a 5% on the five-year, right? The nice thing about that is even if inflation does go up a little bit and rates don't go up a little bit, you have a little cushion in there. You know what I mean?
- 33:12
the 10-year yield at 2%, you got no insulation. Yes. Right? You know, once you have a 5% yield and some of that yield is real yield, you got a little give. It's not going to make you rich, but it'll protect you, you know?