🔍 Read the full analysis: Google’s First Data Center Satellite And The Limits Of AI In Space on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Google says Gemma 4 generated a poem aboard a satellite carrying a Google TPU, marking a small-scale test of its Project Suncatcher concept. The test does not establish that orbital data centers are economical or ready to scale: power, cooling, radiation, repair, communications and launch costs remain major obstacles.
Google says its Gemma 4 model generated a poem aboard a small satellite carrying a Google TPU, an early demonstration for Project Suncatcher, the company’s plan to test AI computing in orbit. The October test shows that a data-center-class chip can run a model in space, but it does not show that orbital computing is ready to compete with terrestrial data centers.
The satellite launched on October 1 aboard SpaceX’s Transporter-18 rideshare mission, according to Google. Built with satellite company Planet, it carries four TPUs but produces up to about one kilowatt of power. TechCrunch reported that power and heat limits mean it can operate only one chip at a time, in short bursts. Google said engineers spent about three weeks commissioning the spacecraft and running initial machine-learning tests.
For the first reported response, engineers sent a prompt during a ground-station pass as the satellite passed over Svalbard, Norway. Google said the model returned a haiku about space, which was then downloaded to Earth. Engineers also tested space-themed limericks and asked the payload for a technical status report. A Google spokesperson told TechCrunch that the satellite was “operating nominally.”
Google describes the result as the first response from a data-center-class TPU in orbit. It is not a claim that this was the first AI model run in space: Starcloud launched a satellite carrying an Nvidia H100 in November 2025, and CNBC reported that the startup later ran and queried Google’s Gemma model in orbit. Google’s distinction is its use of its own TPU and its longer-term design for satellites connected into computing clusters.
Infographic 01 · What flew and what it did
One satellite, four chips, one haiku
Google’s Project Suncatcher prototype is the first test of whether its AI data center chips can work in orbit. Here is what is on board and how it got there.
- 4 Nov 2025Project Suncatcher announcedResearch post and preprint; plan: two prototypes with Planet by early 2027
- 1 Oct 2026Prototype launchedSpaceX Transporter-18 rideshare, built with Planet; peer-reviewed paper in Joule
- Oct 2026About three weeks of commissioningHost and TPU trays powered up, early machine learning tests
- Week of 26 Oct“Hello World”First prompt to Gemma 4; the haiku came down during a pass over Svalbard, Norway
- NextHeat and radiation tests in orbitHow the TPUs behave under thermal stress and real space radiation
- 2027Twin interconnected satellitesMore power and laser links between satellites (Google, TechCrunch)
Scale check
The satellite vs. one power deal on Earth
890 MW = 890,000 kW, about 890,000 times the prototype (my arithmetic). The prototype bar is drawn at a visible minimum; to scale it would be invisible.
Not the first AI chip in orbit
Starcloud-1, launched November 2025 (CNBC)
- Carried one Nvidia H100, a data center GPU
- Trained a small language model in orbit
- Ran Google’s open model Gemma up there in December 2025
- Google’s claim is narrower: the first response from a data center class TPU in orbit
Sources: Google, 30 Oct 2026 · Google, 1 Oct 2026 · Google Research, 4 Nov 2025 · TechCrunch, 30 Oct 2026 · TechCrunch, 1 Oct 2026 · Google and Constellation, 6 Oct 2026 · CNBC, 10 Dec 2025.
Powered by Thorsten Meyer AI
A Test, Not an Orbital Data Center
The demonstration matters because it moves Project Suncatcher from a research proposal to an in-orbit hardware test. Google’s concept is to place solar-powered satellites in close formation and connect them with optical, or laser, links, allowing several spacecraft to work on one computing task. That is different from operating a single AI chip on an isolated satellite, but the current test did not demonstrate such a linked cluster.
Google’s stated motivation is access to solar power. The company’s research says solar panels in a suitable orbit could be up to eight times more productive than on Earth. That potential is relevant as electricity demand from data centers grows: the source material cites U.S. Department of Energy estimates that data centers used about 4.4% of U.S. electricity in 2023, with the share projected to reach 6.7% to 12% by 2028. Those figures describe terrestrial demand, not a forecast that space-based systems will replace it.
The economic hurdle remains substantial. Google’s paper estimates that, at current launch prices, putting a kilowatt of capacity in orbit would cost about $14,700 a year in launch costs. Its estimate falls to about $810 a year if launch costs reach $200 per kilogram, a scenario the company compares with power costs for U.S. data centers. The paper’s calculations are projections, not operating costs established by this prototype; reaching the lower-cost scenario would require a major change in launch economics.
Products Worth Considering
Jackery Explorer 1000 v2 Portable Power Station
Anker SOLIX S2000 Portable Power Station, 2,010Wh, 1,500W
DJI Osmo Mobile 7P Phone Gimbal with Mic Mini Transmitter, Infinity Black
As an affiliate, we earn on qualifying purchases.
From Research Plan to Flight Test
Google introduced Project Suncatcher on November 4, 2025, in a research post and preprint. Its initial plan described solar-powered satellites carrying TPUs, flying close together and communicating through laser links. Google then said it aimed to launch two prototype satellites with Planet by early 2027. The October 2026 launch put a first, smaller prototype in orbit ahead of that stated milestone, but it is not the same as demonstrating the planned network.
Google’s research sketches clusters of 81 satellites at roughly 650 kilometers altitude, with neighboring spacecraft 100 to 200 meters apart. It also reports a bench test of a single pair of optical transceivers reaching 1.6 terabits per second. That is a ground test of communications hardware, not a demonstration of a full orbital cluster. Google published a peer-reviewed version of its research in the journal Joule on October 1, according to the source material.
Space also imposes constraints that terrestrial data centers do not face in the same way. The source material identifies heat management in a vacuum, radiation exposure, limited repair options, communications with Earth and orbital debris as unresolved challenges. It cites about 47,000 tracked objects in orbit; that number refers to objects being tracked, not a count of all debris.
““the first-ever AI response from a data center class TPU in orbit””
Cost, Cooling and Cluster Risks
The prototype’s successful response leaves open whether TPUs can operate reliably over longer periods under thermal stress and radiation. Google has not yet shown that several satellites can maintain the high-speed links needed to function as one computing system in orbit. Its planned close formations would also have to manage spacecraft coordination and the risk of collisions.
The economics are also unresolved. The launch-cost figures in Google’s paper depend on future prices and assumptions; the current satellite is a test platform, not evidence that orbital AI is cheaper than data centers on Earth. The source material does not give a complete cost for manufacturing, launching, operating, replacing or deorbiting a large constellation. It also does not establish how much of the projected solar advantage would remain after those costs and engineering limits are counted.
Google’s “first” wording is specific to a response from a data-center-class TPU. Starcloud’s earlier orbital GPU and reported Gemma tests mean that broader claims about AI running in space would be inaccurate. How Google’s approach compares in performance or cost with other orbital-computing efforts remains unknown.
Two-Satellite Tests in 2027
Google says it will test the current satellite’s TPUs under heat and radiation exposure. If those tests proceed as planned, the company’s next stated milestone is a pair of interconnected satellites in 2027. TechCrunch reported that the pair is intended to generate more power and test optical links between spacecraft.
Those results should offer a clearer measure of whether the system can move beyond a single-chip demonstration. They will not, by themselves, resolve the cost of scaling to the much larger constellation described in Google’s research. The company has not established a deployment schedule for a commercial orbital data center, and the performance and economics of any larger system remain to be tested.
Key Questions
What did Google demonstrate in orbit?
Google says its Gemma 4 model generated a response on a satellite carrying a Google TPU. Engineers received the poem during a ground-station pass as the spacecraft passed over Svalbard.
Was this the first AI model to run in space?
No. The source material says Starcloud launched a satellite with an Nvidia H100 GPU in November 2025, and CNBC later reported that it ran and queried Gemma in orbit. Google’s claim is narrower: its reported first response from a data-center-class TPU in orbit.
Can the satellite run all four TPUs at once?
No. Although it carries four TPUs, TechCrunch reported that the satellite can operate one at a time because of its power and heat limits. Its output is up to about one kilowatt.
Why is Google exploring AI computing in orbit?
Google argues that solar panels in a suitable orbit could be up to eight times more productive than on Earth. The company is investigating whether that advantage could support computing in space, but it has not shown that an orbital system is cheaper or practical at scale.
What is Google testing next?
Google says it will examine how the TPUs perform under thermal stress and radiation. The company’s next stated milestone is two interconnected satellites in 2027 to test optical links, according to the source material.
Source: ThorstenMeyerAI.com
