Discussion

• @patrick_oshag Patrick OShaughnessy on x

  This is my fifth conversation with @GavinSBaker. Gavin understands semiconductors and AI as well as anyone I know and has a gift for making sense of the industry's complexity and nuance. We discuss: - Nvidia vs Google (GPUs + TPUs) - Scaling laws and reasoning models - The [video…

• @patrick_oshag Patrick OShaughnessy on x

  “The most important thing in the next 3-4 years is data centers in space. In every way, data centers in space, from a first principles perspective, are superior to data centers on earth. In space, you can keep a satellite in the sun 24 hours a day. The sun is 30% more intense, [v…

• @nickadobos Nick Dobos on x

  We are in the process of building the worlds biggest computer an AI super giant neocortex around the entire planet. Data centers in space will be real. Star link will make a bajillion dollars cuz they have the only working fast internet connection to the planet.

• @splehman Sam Lehman on x

  Listened to this plus Gavin's ai thoughts post. He seems very confident in pre-training scaling laws holding and I'm just... not so sure? The argument is very focused on advancements in compute pushing pre-training but, definitionally, there needs to be commensurate increases in

• @philipjohnston Philip Johnston on x

  “The most important thing that's going to happen in the world in the next 3-4 year is data centers in space” - @GavinSBaker We have the first Nvidia H100 operating in space at @Starcloud_Inc_ 🚀 [video]

• @patrick_oshag Patrick OShaughnessy on x

  Gavin on how Blackwell and GB300 will end Google's position as the lowest-cost producer of tokens: “I think it's important that Google has been the lowest cost producer of tokens. And this is really important because AI is the first time in my career as a tech investor that [vide…

• @philipjohnston Philip Johnston on x

  Thanks for the @Starcloud_Inc_ shoutout @GavinSBaker! 👌 “Especially when there is a working, albeit very small, datacenter in space *today* - Starcloud's orbital setup just successfully trained an LLM. Great name btw.”

• @gavinsbaker Gavin Baker on x

  Starlink v3 will be 20 kilowatts and should have no thermal dissipation issues. Elon's stated plan is 100 kilowatts for each AI satellite; slightly less than a full Blackwell rack. Starship should be able to lift 10 to 15 megawatts to sun-synchronous orbit per flight.

• @deryatr_ Derya Unutmaz on x

  Amazing conversation with @GavinSBaker! I highly recommend it to anyone interested in the future of AI scaling!

• @clawrence Craig Lawrence on x

  I really don't get this. Terrestrial power via gas & solar & batteries is very cheap. Build generation onsite and you don't have to pay for transmission. These are 30+ year assets. How is it remotely possible that this will be cheaper in space?

• @jackfarley96 Jack Farley on x

  Very good interview. In Gavin's view, “by far most plausible and scariest bear case” to AI & data center CapEx is Edge AI (running AI on devices, not from data centers)

• @patrick_oshag Patrick OShaughnessy on x

  Gavin explains that the bear case for AI capex spend is on-device inference: “In three years, on a bigger phone, you'll be able to run a pruned-down version of Gemini 5, Grok 4, or ChatGPT. And that's free. This is clearly Apple's strategy - we're going to make it privacy-safe [v…

• @elonmusk Elon Musk on x

  Yeah

• @elonmusk Elon Musk on x

  @GavinSBaker Fools are determined to be fools. Trying to stop them from being so is futile.

• @gavinsbaker Gavin Baker on x

  Deeply amused by all the confident commentary that datacenters in space do not work from a physics and engineering perspective. Elon operates two of the largest coherent GPU clusters in the world, SpaceX is responsible for over 90% of mass to orbit and SpaceX operates the

• @patrick_oshag Patrick OShaughnessy on x

  Gavin on why it's a mistake for SaaS companies to resist AI because it has a lower margin structure: “When there's a transformative new technology customers are demanding, it's always a mistake not to embrace it. If you're trying to preserve an 80% gross-margin structure, you [vi…

• @pgelsinger Pat Gelsinger on x

  Unless you have a technology that naturally operates at temperatures of 2-3 Kelvin and at extremely lower temperatures, the plan for data centers in space makes no sense. Oh, we have some good news — that is the natural temperatures of superconducting devices like

• @drphiltill Phil Metzger on x

  It's crazy how quickly opinions flipped on this.