A space data center sounds like it solves cooling for free: deep space sits at about −270 °C. Google's own engineers say the opposite. In the post announcing Project Suncatcher's first launch, Google calls cooling orbital data centers "a crucial research challenge", and according to Ars Technica its first test satellite can run its TPUs for only about 15 minutes before it has to stop and let the radiators catch up. If you care where AI compute gets built, and who pays for the power, this is the physics behind the headline.
TL;DR
- On October 1 Google launches "MVP", a fridge-sized satellite built with Planet that carries four TPUs, on SpaceX's Transporter-18 rideshare. It is a test, meant to run for a few months.
- Space is cold but it is a vacuum, so heat cannot be carried away by air or water. It can only leave as infrared light from a radiator, and a perfect radiator at room temperature sheds about 460 watts per square metre.
- The satellite gets about one kilowatt of solar power, roughly a hair dryer. Per Ars Technica, citing The New York Times, the cooling system runs in "brief spurts of about 15 minutes". Then the TPUs shut down.
- Why bother: in the right orbit a panel sees near-constant sunlight and, per Google, makes up to eight times more power than on Earth, with no permitting board in the way.
- Also in this episode: Oracle owes investors for a New Mexico data center even without electricity, UK iPhones with two levels of encryption, and F-Droid 2.0.
Why is cooling a space data center so hard?
On Earth, cooling is contact: air or water touches a hot chip and carries the heat away.
In orbit there is nothing to touch. Google's post says it in two sentences:
Radiation is slow. The Stefan–Boltzmann law gives the power a surface can glow away: P = εσT⁴. For a perfect radiator (ε = 1) at about room temperature, 300 K, that is roughly 460 watts per square metre per side. The −270 °C of the cosmic background barely helps: the limit is what the radiator can emit, which depends on its own temperature.
The International Space Station is the working example. Its external active thermal control system is "designed to provide 35 kW of heat rejection per loop for a total capability of 70 kW", per NASA. It uses two rotating radiator wings with three panels each, and each panel is 23.3 by 3.4 metres, about the length of a tennis court. Six tennis-court-long panels, 70 kilowatts.
Now scale. Google's Ironwood TPU "scales up to 9,216 liquid cooled chips … spanning nearly 10 MW". Ten megawatts divided by 70 kilowatts is about 140 space stations' worth of radiators for one pod. A back-of-envelope version in Python, with the same caveats:
# back-of-envelope: ideal black-body radiator, one side,
# ignores absorbed sunlight and Earth's infrared
sigma = 5.67e-8 # Stefan-Boltzmann constant, W/(m^2 K^4)
T = 300 # radiator temperature, K (about room temperature)
flux = sigma * T**4 # ~459 W per m^2
pod = 10e6 # one Ironwood pod, "nearly 10 MW"
print(pod / flux) # ~21,800 m^2 of perfect radiator
Hotter radiators do much better, because the output grows with the fourth power of temperature. So treat "140 space stations" as a rough sense of scale, nothing you could build to. The direction is clear either way: in space the chips are easy to power and hard to cool.
Which explains one word in Google's post. Future satellites "will each carry dozens of TPU chips". Only dozens.
Project Suncatcher: what Google is launching on October 1
The announcement came from Travis Beals, Senior Director for Paradigms of Intelligence. Google is "launching a prototype satellite to evaluate how Google Tensor Processing Units (TPUs) perform in space", developed with Planet and flying on SpaceX's Transporter-18 rideshare. The New York Times got the first inside look: a Falcon 9 from Vandenberg Space Force Base, a satellite named MVP. For once, an honest product name.
Per Ars Technica, MVP is about the size of a refrigerator, carries four TPUs, and is built on a spacecraft Planet had already made for an early test. The original plan was two custom satellites in 2027; Google wanted to go sooner. The heat path runs from the chips through a thermal interface material into aluminium and copper heat pipes and out to a radiator, tested in a thermal vacuum chamber.
The other risks were tested too, according to Google:
| Risk | What Google tested | Result, per Google |
|---|---|---|
| Launch | ~10 minutes to orbit, sustained loads up to 10 g; shake tests on all three axes | held up |
| Radiation | Trillium TPUs in a proton beam at UC Davis while running AI workloads | survived a dose "greater than what they would receive during a five-year space mission" |
| Heat | heat pipes and radiators in a thermal vacuum chamber | the open problem |
The radiation result goes back to Google's November 2025 research paper: the first irregularities showed up in high-bandwidth memory at 2 krad(Si), about three times the shielded five-year dose, with no hard failures up to 15 krad. That paper also listed "thermal management" among the "significant engineering challenges" remaining. A year later it is still the headline problem.
Sundar Pichai announced it on X: "Can our TPUs survive and operate in space? Well, we're going to find out." It ended "One small step for TPUs...." Elon Musk quoted it with two rocket emojis, which is fair, since it is his rocket.
The 15-minute limit: how Google's first orbital data center actually runs
Here is the number Google's own post leaves out. MVP's solar panels produce "only about one kilowatt of power, The New York Times reports", Ars writes. "That's enough to power a microwave or a hair dryer." And the cooling can't keep up even with that:
So they survived the radiation in a proton beam and the rocket on a shake table. The limit is the heat. Right now physics is winning, and Google's first space data center works in shifts, with naps.
That is fine for a test: a duty cycle is a measurement too. It is a problem only for the phrase "data center".
The Hacker News thread (196 points, 396 comments) got there quickly. The third top-level comment, from zactato, is one line: "How are they solving the heat dissipation issues?"
Why build data centers in space at all?
Because heat is half the problem, and power is the other half. Google's post: "In low Earth orbit, satellites can access near-constant sunlight, generating up to eight times more solar power than on Earth."
The economics come from the 2025 paper. It bets launch prices "may fall to less than $200/kg by the mid-2030s". At that price, Google says, launching and operating a satellite is "roughly comparable" to what a data center on Earth pays for energy per kilowatt per year. The design sketched there is an illustrative cluster of 81 satellites within a one-kilometre radius, linked by lasers. Two more satellites are planned for 2027.
And on Earth, power has turned into a permission problem, which brings me to Oracle.
Also in this episode: Oracle, UK iPhone encryption, F-Droid 2.0
Oracle pays for a data center with no power. Oracle's 1,400-acre "Project Jupiter" campus in Doña Ana County, New Mexico, at the centre of its $300 billion compute contract with OpenAI, ran into permitting delays and local opposition. Oracle sent the developer, a Blue Owl Capital unit, a force majeure notice over the power supply. The Financial Times reports it still owes a "carry cost" for up to three years if the site is not ready by the first-phase deadline: "The power delivery risk sits with Oracle, not the investors." On HN, ForHackernews: it's "lowkey insane that it will end up cheaper to shoot your datacenter into space than get it past the county board permitting process."
Same iPhone, two encryption tiers. macanorak describes Alice and Bill in the UK with identical iPhones. Alice turned on Advanced Data Protection before Apple withdrew it for new UK users in February 2025, so her backups and photos are end-to-end encrypted. Bill can't turn it on. The cause is a secret Technical Capability Notice under the Investigatory Powers Act. Apple is gagged from confirming it exists and, per the Telegraph as cited there, has asked the tribunal to lift the gag; Liberty and Privacy International called it "farcical". The author wrote to the Home Secretary before publishing. No reply. (HN)
F-Droid 2.0. The open source Android app store shipped its biggest update ever: rebuilt in Kotlin Compose, updates installed in the background by default, rolling out "over the coming weeks". The new unified installer uses Android's pre-approval API, "thanks largely to pressure from the EU's Digital Markets Act (DMA)". The same page opens with a banner: "F-Droid is under threat. Google is changing the way you install apps on your device." It was the top story on HN that day with 1,280 points.
Verdict: NEEDS REVIEW
I stamped Google's space data center NEEDS REVIEW. I would ship the test flight tomorrow: the radiation and launch results are real, and measuring TPUs in orbit is exactly what a prototype is for. But four chips on a hair dryer's worth of power, napping every 15 minutes, are still an experiment, a long way from a data center. The headline goes back for review until the radiators can keep up.
FAQ
What is Google Project Suncatcher?
Google's research project to run TPUs on solar-powered satellites. The first prototype, MVP, launches October 1, 2026 with four TPUs.
Why is it hard to cool a data center in space?
A vacuum has no air or water to carry heat away. Heat can only leave as infrared radiation from radiators, about 460 W per square metre for a perfect radiator at room temperature.
How long can Google's TPUs run in orbit?
Per Ars Technica, citing The New York Times, about 15 minutes at a time. Then the TPUs shut down so the radiators can catch up.
Are space data centers cheaper than ones on Earth?
Not yet. Google's 2025 paper says they could be roughly comparable if launch prices fall below $200 per kilogram by the mid-2030s.
Sources
- Google blog, "Behind Project Suncatcher" (Sep 24, 2026): https://blog.google/innovation-and-ai/models-and-research/google-research/google-project-suncatcher-facts/
- Ars Technica (Sep 24, 2026): https://arstechnica.com/google/2026/09/googles-first-suncatcher-orbital-data-center-test-launches-october-1/
- The New York Times (Sep 24, 2026): https://www.nytimes.com/2026/09/24/technology/google-suncatcher-ai-data-center-space.html
- Google Research, space-based AI infrastructure design (Nov 2025): https://research.google/blog/exploring-a-space-based-scalable-ai-infrastructure-system-design/
- NASA, ISS active thermal control system overview: https://www.nasa.gov/wp-content/uploads/2021/02/473486main_iss_atcs_overview.pdf
- Google Cloud, Ironwood TPU: https://blog.google/products/google-cloud/ironwood-tpu-age-of-inference/
- Stefan–Boltzmann law: https://en.wikipedia.org/wiki/Stefan%E2%80%93Boltzmann_law
- Sundar Pichai on X: https://x.com/sundarpichai/status/2103209164072010051
- Hacker News, Suncatcher: https://news.ycombinator.com/item?id=49830606
- Financial Times on Oracle's Project Jupiter: https://www.ft.com/content/a96bf05a-a299-4d6a-a753-b298dd0f4016
- macanorak, "Two-tier encryption in the UK": https://macanorak.com/two-tier-encryption-in-the-uk/
- F-Droid 2.0: https://f-droid.org/2026/09/24/f-droid-2.0-a-new-chapter-for-android-freedom.html
This article expands on an episode of **The Daily Diff, a five-minute daily video on what shipped and what broke in tech.
Watch the episode · Subscribe on YouTube · the written diff lands in your inbox every morning at thedailydiff.dev.


Top comments (0)