Skip navigation
Chamath explains how a data center slowdown turns the hyperscalers into kingmakers and squeezes the…
0:00 / 2:14

Chamath explains how a data center slowdown turns the hyperscalers into kingmakers and squeezes the…

Fireside Alpha

@firesidealpha39K followers

Visit on X
198 likesDownloadOriginal post

115,378 views

Chamath explains how a data center slowdown turns the hyperscalers into kingmakers and squeezes the neoclouds and foundation models: “Now, they’re doing smart deals because that’s what made sense. But I think what the incoming governors are going to say is these things are off the table. And instead, we want to have directly measurable economic impacts before we even entertain this.” “I think what it means is we’re going to see a slowdown. And I think what that does is it benefits the hyperscalers because they have existing capacity. I think it’s very bad for the neoclouds because they’re renters. And it’s pretty bad for the foundational models because they are so short this capacity and they need it to grow. So to the extent that these frontier models are so much better than everybody else’s product, they are the ones that need this capacity. It sets the hyperscalers to be a kingmaker, and they have this interesting game theory now.” [Jason: “Can you just say who the hyperscalers are, just so people understand the difference between them?”] “Amazon, Meta, Microsoft, Google, SpaceX. And what they’ll be able to do is make a decision. Path number one is continue to rent in the way that you’re doing it today. So the status quo, rent to some customers, large corporates, right? Rent to the neoclouds, rent to the foundational models. That’s what they’re doing today. Everybody does that.” “But then there’s the path less traveled, which all of a sudden becomes really interesting. You could decide to take all of the incremental compute and advantage your own products. So Google has a ton of capacity coming online. They could say, I’m going to give it to Gemini. What does that do? It actually allows Gemini to catch up in quality relative to the foundational models.” “Why? Because if you constrain the amount of assets that can actually generate compute, and a smaller surface area of it is available to the leading companies, those leading companies will have to pay more, the user experience will decay, and then other products can have time to catch up.” “And now that is important, because Anthropic and OpenAI are in a very fast recursive loop. They’re on a 60 to 90 day loop in improving their models. So now all of a sudden, without an infinite amount of supply coming online, you can slow these folks down.”

12 Comments

Comments from the original post on X (Twitter)

  • Panke's profile picture

    @spcxGOD

    I explained why SpaceX was the kingmaker months ago, glad these people are reaching the same conclusion now.

    3 likes

  • E1ev3n.'s profile picture

    @harden_369741

    Compute access has decided more AI races than model quality in the work I follow over the last two years. If the hyperscalers can throttle a rival by redirecting capacity, the moat is real estate as much as talent.

    2 likes

  • Fireside Alpha's profile picture

    @firesidealpha

    Ft @chamath Full video:

    1 like

  • Mr. Clicker's profile picture

    @MrClicker_1

    That's the real squeeze: if compute gets tight, the power plant owners win and everyone else rents at their mercy. Google keeping its best chips for Gemini is the quietest power move in AI.

    1 like

  • oj's profile picture

    @OlegJasnic

    So if you neocloud with capacity that is already online who does that become less valuable?

  • The SI Maximalist's profile picture

    @thesimaximalist

    New capacity may already be spoken for. Google could prioritize Gemini with uncommitted compute, but how hard that squeezes rival SI labs depends on their contracts and how quickly they can secure capacity elsewhere.

  • Raúl Nogales 🇪🇸🇺🇸's profile picture

    @raulnogales

    The flaw in that game is assuming compute is the whole race. Slowing Anthropic and OpenAI buys Gemini time, but users are already locked into workflows and habits. Google would also be trading rental revenue and a lot of goodwill for a lead that still has to be earned in the product.

  • G S A's profile picture

    @GSA199452

    Demand will not change. HypScaler need the data center. Nothing makes sense what you say

  • Prohumanity's profile picture

    @Prohumanistx

    SpaceX king of orbital racks will be very VERY fine. Hyperscalers too

  • Recovering Bag Holder's profile picture

    @BtcAvi

    Neoclouds like Iren own the full stack…

  • Cole Nichols's profile picture

    @ColeNichol99434

    I disagree with this, his scenario 2 made more sense where spcx cancels their 90 days agreement with Anthropic and forces them to be a price taker

  • Rick Santos's profile picture

    @RickSantosCliff

    This ignores enterprises shifting towards ACIE, which grew faster than hyperscaler capex over the past year.