Follow me on LinkedIn:
https://www.linkedin.com/in/leo-cui-ph-d-cfa-59531439/
I am a venture capital investor at FoundersX Ventures, an AI and deep-tech VC fund based in Menlo Park, Cambridge and NYC. I previously did my Ph.D. at Harvard, postdoc at MIT.
All infographics and the full presentation from this video are available to download for Research tier members — plus members can propose my next research topics and enjoy an ad-free version of every video.
---
Somewhere near you, there may be a building with no windows, a very large electricity bill, and a fence that makes it clear you're not invited inside. Behind that fence, tens of thousands of processors are training the models that technology companies believe will reshape the economy. We call it "the cloud," which makes it sound weightless. It's a warehouse that consumes extraordinary amounts of power and turns nearly all of it into heat.
And the industry is running into a wall — not a money wall. Electricity that can be delivered to the right place. Transformers that arrive on time. Land that can be permitted. Water communities are willing to share. Transmission lines that take years to build.
So the proposal is exactly as strange as it sounds: if you can't bring the power to the computers, move the computers to the power. Put the data center in orbit, where sunlight is uninterrupted and nobody objects to the substation.
Google, SpaceX, and a startup called Starcloud are all working on it. NVIDIA hardware has already run in orbit. But between a working demonstration and a gigawatt of orbital compute sits a series of physics problems that don't care how much money you have — and the most misunderstood one is cooling. Space is not cold in the way people assume.
In this video, I follow one processor from a terrestrial data center into orbit and ask what actually breaks — then separate what has genuinely been demonstrated from what exists only in press releases.
You'll learn:
- Why data center growth is now limited by grid connections rather than capital
- Why "data centers in space" describes two completely different businesses — and only one is close to real
- What a data center actually provides a processor beyond a rack and a power cable
- Power: the one genuine, unambiguous advantage of orbit
- Cooling: why the vacuum of space cannot cool a hot chip, and what the Stefan-Boltzmann law demands
- Radiation: why encouraging test results are not a service-level agreement
- Networking, latency, and the missing technician
- The proof ladder — what has actually flown and worked, step by step
- Google's Project Suncatcher, and what its 2027 demonstration will and won't prove
- SpaceX's Starmind and AI1 — and how this connects to Starship economics
- The regulatory land grab: spectrum and orbital slots
- Scaling from one processor to one gigawatt — the honest arithmetic
- When the economics could work, and why speed may matter more than free electricity
- Orbit is not empty: debris, collision risk, and astronomy
- Who might benefit — and why most of the obvious names have almost no exposure
⏱️ CHAPTERS
0:00 Cold Open — The Pitch
2:17 Act 2 — Why Anyone Is Considering This
5:51 — Two Businesses Hiding Under One Name
8:41 Act 3 — Follow One Processor Into Orbit
11:02 — Power: The Genuine Advantage
15:30 — Cooling: Why Cold Space Doesn't Cool a Hot Chip
18:38 — Radiation: Tests vs. Guarantees
20:30 — Networking, Latency, and the Missing Technician
25:18 Act 4 — The Proof Ladder: What Has Actually Happened
33:11 — Google's Project Suncatcher
37:29 — SpaceX: Starmind and AI1
40:56 — The Regulatory Land Grab
44:45 Act 5 — From One Processor to One Gigawatt
46:27 — When Could the Economics Work?
48:41 Act 6 — Orbit Is Not Empty
52:50 Act 7 — The Investor Map: Who Might Benefit
57:32 Final Verdict — Is This Actually Realistic?
Companies discussed include Google (Alphabet), SpaceX, Starcloud, NVIDIA, AMD, Blue Origin, Amazon, Microsoft, Rocket Lab, Redwire, Planet Labs, TSMC, SK Hynix, Micron, Samsung, and many more.
Disclaimer: Nothing in this video is financial advice. I am a venture capital investor, and my fund invests in AI, space, and deep-technology companies. Some companies discussed may overlap with areas I evaluate professionally. This video is for education and reflects my own analysis. Always conduct your own research and never invest money you cannot afford to lose.
The infographics and images in this video were created with the assistance of AI tools. While every effort was made to verify their accuracy, they may contain errors. Technical specifications, cost estimates and project timelines are based on sources available at the time of production and change over time. If you spot an error, please let me know in the comments — corrections will be pinned.
#DataCenters #Space #AI #SpaceX #Google #Starcloud #NVIDIA #OrbitalComputing #Investing #WatchBeforeInvesting 
Comments (0)