Real Alpha
Watch mode

Cheeky Pint · 2 min instead of 2 h 15 m · Oct 12, 2026

Why Meta's compute bet is less risky than the $100B headline

Mark Zuckerberg

Founder & CEO, Meta

  • John CollisonHost · Co-founder & president, Stripe

    Sold his first startup, Auctomatic, at 17, then co-founded Stripe in 2010.

About this segment

Zuckerberg argues the downside is capped because other buyers would take the compute and Meta has other uses for it. For the bet to fail, AI overall would have to fail.

From “Muse Wonder: Zuckerberg and Wang's $100B AI”, Oct 12, 2026. Watch the full episode

“if AI works overall, plenty of other people want all our compute. So they're willing to pay.”
Mark Zuckerbergat 58:03 · Cheeky Pint58:03

Takeaways

  1. Spare compute can be sold to others if AI demand holds up
  2. Failure would require AI overall to fail, not just Meta
  3. Product demand was built ahead of the product existing
Watch 2:14 from 56:35

In this episode

Full episode
  1. 51:35Training a frontier model means committing $100B before you know it works4:16

  2. 56:35Why Meta's compute bet is less risky than the $100B headline2:14

Up next · All-InQuantum computing has one real trick and no business case2:41

Transcript

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

  1. 56:39

    know we want to train the model, because this is key, but how much product demand do we think we're going to have? So now it's like, alright, well, the product doesn't exist. like and as you're trying to build data centers to serve Muse for a product that doesn't exist years in advance this is probably where when we were talking about it last year there was a moment where I think a bunch of people were like yeah maybe we should just be a little more conservative no we're doing this we're going to make it work this is going to work but it was like I mean even I mean, I too am an entrepreneur and founder and have, you know, at various points in my life taken risk.

  2. 57:22

    I remember the visceral feeling at the time was like, this is a fuck ton of risk. Like, I don't know if I remember. It's only risk. Yeah, yeah, it's only, well, I do think this is like one of the like things that I think, Mark, is very impressive about you, which is I think you have really internalized this concept of, you know, it's about how big it is when you win, not like, you know, failure is okay.

  3. 57:46

    It's just like how big is the win when you do win? And this was, I think... Mark has fully internalized power laws in a way that many people have not. I think quite deeply, yeah. Well, I think the other pieces... I don't actually think it's quite as risky as this. I mean, the numbers are big. But I think... Meta has a backstop in terms of how it can use compute. Yeah, so I think there's two things.

  4. 58:05

    One is if AI works overall, plenty of other people want all our compute. So they're willing to pay. I mean, if the cost is whatever, 15 billion a gigawatt, we could get a lot more for that. So in order for it not to work out, it's not just that we have to fail. AI overall has to fail. And then even when that happens, we still have all these other uses to productively use compute. So in the grand scheme of things, we were definitely building out this compute for inference ahead of knowing that the product was going to be ready.

  5. 58:41

    But your AI pilled enough that you said, GPUs online empowered will be useful to someone. Yeah, but also we believed in what we were building.