Amy

10 TIMES NVIDIA To put 100 GW/per year of AI computers in perspective, Grok estimates Nvidia is presently at 10GW/per year in AI compute. IE 10X Nvidia. 100 TRILLION COMPANY Now you see why Elon Musk says a 100 Trillion market cap company may be possible within 10 years. SpaceX, xAI and Tesla will all play major roles in AI Space Compute.

Chairman (king) of orbital data center age. He's in favor of quote "weird space stuff". I think we are all going to be alright

@diana_dukic @elonmusk I agree about the instability and not seeing the posts of my friends, or even my twin.

@elonmusk The instability is the biggest hurdle. Between unpredictable reach and a 'For You' feed that feels disconnected from my actual community, the ROI on high-quality content has plummeted. Creators need a reliable feedback loop, not a void. And bigger payouts would be nice too lol

Went from scrolling 24/7 to not even wanting to log in. X just hasn’t been hitting the same lately.

@ourteslafuture He plans how to hyperscale first, while others get surprised by the constraints later….

When we start to imagine what 10x and 100x the energy generation compared to our current earth based infrastructure could look like it is obvious that free electricity in space becomes the next step forward for humanity and Elon will continue to lead multiple years ahead of the rest of the terrestrial bound players in this space, pun intended. Taking it to 100x is what Elon does best. Choosing the correct goals and mission statements helps a lot…

FYI: There won’t be enough energy on Earth to power the AI Race. So this means that SpaceX wins.

@SawyerMerritt @SpaceX SpaceX will ultimately deliver millions of tons to the Moon to build a self-growing city there and same for Mars

@Kristennetten I hadn’t either!

@_SFTahoe I didn’t realize there are so many templates 😂

Try the Grok Imagine > “Helicopter Escape” template 🚁

@JK_Lundblad Great writing, the litigation tax in the US is unfortunately real.

@_SFTahoe It does sound sci-fi but so did Starlink...yet here we are. I agree with you on the litigation piece. American legal environment is dysfunctional and prevents anything from getting done. I wrote this last week on that very issue: risknprogress.substack.com/p/scales-skewed

Good point. SpaceX will provide launch to orbit for other companies too. Win/win for SpaceX as it creates revenue and allows a increased launch cadence. And Like Yahoo’s Alibaba investment it will prove very profitable for Google. There will be a number two, three, four etc. But they, IMO, will be competing with each other for 2nd place as SpaceX - through it’s vertical integration - can scale faster and cheaper.

@_SFTahoe Well then It's a good thing Alphabet owns a piece of SpaceX 😅

It isn’t just permitting, but also nuisance lawsuits, growing expense of energy, and hard limits on things like gas turbine blades - that are sold out through 2030. Earth isn’t a realistic option for compute at scale, IMO. Only 16% of the world’s population right now uses AI, and the majority of compute is being used for training. That will flip to 99% of compute being for inference using the models, as billions of people start using AI frequently, many companies are replaced entirely by AI, robots, etc. This will be an exponential increase. Energy providers have long lead times, and China doesn’t face the same immediate energy restrictions as the West, which would allow China to win the AI race short term, and I don’t think that is an acceptable outcome for most, and it is a huge national security risk. In contrast, space takes the two largest expenses (real estate & energy) and makes them negligible. While a difficult engineering challenge it is doable to put compute in Space, and in fact has already been done, by both China and US Startup Starcloud. The limiting factor for Space will be rocket fuel at around a Terawatt of annual compute, but then more compute could be manufactured and launched from a moon base and mass driver. Sounds sci fi, but so did airplanes at one time.

@_SFTahoe Perhaps. It's an interesting concept but might we be better served by improving the chips, the algorithms, and tackling the permitting issues on Earth?

@TeslaAlgo Tesla and SpaceX and xAI are executing on every level.

@pbeisel Probably more like 160k wafers/month, factoring in yield

@jhong The limiting factor will shift from chips to energy on Earth, then back to chips when space solar (star) power is unlocked

@peterrhague Mass driver on the Moon ftw

@extropic @Gfilche @aaronburnett @SpaceX @Tesla @TeslaLarry Wow, now I am really curious. Can’t wait to see what you come up with @extropic

@Gfilche @aaronburnett @SpaceX @Tesla @_SFTahoe @TeslaLarry 👀🛰️

Terafab & the Inevitable @SpaceX + @Tesla merger ⚡️ I sat down with @_SFTahoe & @TeslaLarry to discuss Tesla's plan to build chips, and how that will be the key partnership for a full merger with SpaceX/XAI 🎙️ 0:45 SpaceX to $100T 2:30 SpaceX Plans To Be 10X Bigger Than NVIDIA 5:20 The Terafab is Coming! Impact & Timing 22:36 Digital Optimus & MARCROHARD Project 37:03 SpaceX/Tesla/xAI Cash Needs 41:40 The SpaceX IPO Process 48:08 The Biggest Merger/Company of All-time? 54:17 Grok Is Helping With All of This 🙏 Huge thanks to Larry & Amy for coming on the show! Two of the most knowledgable voices on these topics!

@wholemars So psyched.

Wow, they’re really doing it.

NEWS: Tesla has started hiring for its Terafab (chip manufacturing) project in Austin, Texas! Job listing: "You'll own end-to-end program scoping—including factory design/construction from concept through execution, ramp-up, and production readiness. The ideal candidate brings a strong background in technical program management, semiconductor facility engineering, multidisciplinary design leadership, value engineering, capacity modeling, and cross-functional execution for large-scale programs."

TESLA’S FSD “14.3 is where the last big piece of the puzzle finally lands”. Elon Musk Elon Musk just confirmed that Tesla’s v14.3, (that he’s described as feeling sentient) is in final testing with wide release expected in just a few weeks.

TESLA’S ADVANTAGE=Vertical Integration “AI5 will punch far above it’s weight, because the entire Tesla AI software stack is designed to make maximally effective use of each circuit.” Elon Musk ♦️Be you own best customer: Then chip design & AI software are optimized to work together. This is also a protection against your AI being stolen, as it is designed to work on your hardware. ♦️Tesla’s Terafab will compound this advantage as Tesla & SpaceX engineers can sit together and optimize the chips. Iteration is faster & COGs are lower. ♦️A Terafab doesn’t mean Tesla & SpaceX (or a joint merged entity) won’t still take all the Chips Nvidia, Samsung and TSMC can provide. Tesla typically asks their suppliers how much they can get - and produce the projected shortfall in house. Elon Musk just confirmed: “SpaceX Al and Tesla expect to continue ordering Nvidia chips at scale.”

@DustinWalper Great goal.

I'm a simple man. All I want is to build a generational company that brings heavy industry back to America at the largest possible scale.