🔍 Read the full analysis: Does Google’s Haiku-Writing Satellite Change How We Think About Space AI? on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
A Google-built TPU aboard a prototype satellite generated a haiku after a prompt was sent during a pass over Svalbard, Norway. The test shows that a data-center-class chip can run AI in orbit, but the satellite is small, operates in short bursts and does not establish that space-based AI is economical or practical at scale.
Google says a prototype satellite carrying one of its Tensor Processing Units generated an AI-written haiku in orbit this week, a first test of the company’s plan to explore computing clusters in space. The October 30 result shows that Google’s Gemma 4 model could respond to a prompt from orbit, but it does not show that space-based AI can yet operate at commercial scale or compete with data centers on Earth.
The satellite was built with satellite company Planet and launched on October 1 aboard a SpaceX rideshare mission. After several weeks of commissioning, engineers sent a prompt while it passed over Svalbard, Norway. Google said the satellite returned the three-line poem: “Blue marble below / Data streams in silent grace / Starlight feeds my core.” Engineers also tested space-themed limericks and asked the payload to report its technical status.
The spacecraft carries four Google TPUs, but power and heat constraints mean it can operate only one chip at a time. TechCrunch reported that the satellite can generate up to about one kilowatt of power and that the chip ran in bursts of roughly 20 minutes during the reported test. Google described the event as its “Hello World” moment for Project Suncatcher; a spokesperson told TechCrunch the satellite was “operating nominally.”
Google calls this the first AI response from a data-center-class TPU in orbit. That wording is narrower than saying AI has never run in space: Starcloud launched a satellite carrying an Nvidia H100 GPU in November 2025, and CNBC reported that the company later ran and queried Google’s Gemma model in orbit. Google’s proposed distinction is its longer-term design: satellites linked by optical, or laser, connections so multiple machines could work together.
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 Small Test of Orbital AI
The result matters as a hardware demonstration, not as proof of a new commercial computing platform. Running a model on a TPU in orbit tests whether AI hardware can function through launch and under the thermal and radiation conditions found in space. Google says further tests will examine those stresses. That information could help establish whether specialized computing payloads can work reliably beyond Earth.
Google’s case for eventually moving some computing to orbit centers on solar energy and data-center demand. The company’s research says solar panels in a suitable orbit could be up to eight times more productive than on Earth. Separately, the U.S. Department of Energy has estimated that data centers used about 4.4% of U.S. electricity in 2023 and could account for 6.7% to 12% by 2028. Those figures provide context for the energy challenge; they do not show that orbital facilities would reduce electricity costs or emissions.
The gap between this test and a useful orbital cluster is substantial. A satellite producing about one kilowatt is a tiny system compared with terrestrial facilities, and a working constellation would need many spacecraft, reliable links and ways to manage heat, damage and replacement. Google’s own cost estimates also make launch prices a central factor. The haiku confirms a functioning experiment; it does not settle the economics.
Products Worth Considering
Logitech MX Master 4 for Mac, Ergonomic Bluetooth Mouse with Haptic Feedback, Ultra-Fast Scrolling, USB-C Charging, for MacOS, iPadOS, Windo
Dell 34-Inch S3425DW Curved WQHD USB-C Monitor
Dell XPS 13 9350 Copilot+ Laptop, 13.4-inch 120Hz Display, Intel Core Ultra 7 256V, 16GB RAM, 512GB SSD, Ubuntu Linux 24.04 LTS
As an affiliate, we earn on qualifying purchases.
From Suncatcher Plan to Prototype
Google introduced Project Suncatcher in November 2025 as a research effort to study solar-powered satellites carrying TPUs. Its proposed architecture involves groups of spacecraft flying close together and exchanging data through laser links, with the aim of making them operate more like a single computing system. The company’s published research sketches clusters of 81 satellites at roughly 650 kilometers altitude, separated from neighboring craft by about 100 to 200 meters.
The October flight arrived ahead of the schedule described in Google’s initial announcement. At that time, Google said it was working toward a learning mission with Planet involving two prototype satellites by early 2027. The company now says the first prototype launched on October 1 and has completed early commissioning. Google also published a peer-reviewed version of its research in the journal Joule on that date.
Other companies have already tested AI hardware in orbit, including Starcloud’s H100 satellite. The distinction between projects matters: the reported Google test concerns a TPU response, while Suncatcher’s broader ambition is to evaluate interconnected orbital computing. The source material reports a 1.6-terabit-per-second bench test between a pair of optical transceivers, but a bench result is not the same as proving a satellite network can sustain that performance in orbit.
““Hello World” moment”
— Google, describing the reported test
Costs and Orbital Risks Remain
Commercial viability remains unproven. Google’s research estimates annual launch costs of about $14,700 per kilowatt at current prices, compared with about $810 per kilowatt if launch costs fall to $200 per kilogram. These are estimates based on stated assumptions, not observed operating costs for a commercial orbital data center. The source material says the lower-cost scenario would require roughly 1,800 Starship launches; the schedule, feasibility and cost of such a launch campaign are not established.
Technical questions are also open. Google has said its next tests will examine thermal stress and radiation exposure. The source material identifies heat removal in a vacuum, repairs, communications with ground stations and orbital debris as unresolved challenges, including about 47,000 tracked objects in orbit. It is not yet clear how long the prototype can operate, how its hardware will perform over time, or whether optical links between satellites can support sustained AI workloads in orbit.
The available report does not give independent test results, detailed reliability data or a full accounting of launch, maintenance and replacement costs. The haiku is evidence that one prompt-response test succeeded, not evidence that the larger system can train or serve AI models economically.
Thermal Tests and Twin Satellites
Google says the prototype will undergo tests of TPU performance under thermal stress and radiation exposure. The company has identified interconnected satellites as the next major milestone, with a planned two-satellite mission in 2027. TechCrunch reported that the pair is expected to generate more power and test optical links between spacecraft.
Those tests should help clarify whether separate satellites can exchange data reliably while operating in orbit and whether the hardware can withstand the environment. Google has not established a date for a larger constellation or provided a confirmed commercial deployment plan in the material available. For now, the next evidence to watch is the outcome of the environmental tests and whether the 2027 link demonstration proceeds as described.
Key Questions
What did Google’s satellite do?
Google says its prototype satellite received a prompt during a pass over Svalbard and returned a haiku generated by Gemma 4 running on a Google TPU. The company also reported tests involving limericks and a request for the payload’s technical status.
Was this the first AI system in space?
No. The source material says Starcloud launched a satellite carrying an Nvidia H100 GPU in November 2025 and later ran and queried Google’s Gemma model in orbit, according to CNBC. Google’s claim is specifically about a response from a data-center-class TPU in orbit.
Can the satellite run a full data center?
No. The reported prototype generates up to about one kilowatt and can operate only one of its four TPUs at a time, in bursts. It is a small experiment, not a full data center or a demonstration of large-scale AI computing.
Why is Google testing AI computing in orbit?
Google’s research explores whether satellites could use abundant solar energy and connect through laser links to form computing clusters. The company has argued that solar panels in a suitable orbit could be more productive than on Earth, but whether this can lower costs or address data-center power demand remains unproven.
What happens after the haiku test?
Google says it will test the TPU’s performance under radiation and thermal stress. Its next stated milestone is a pair of interconnected satellites in 2027 to test optical links, according to the company and TechCrunch.
Source: ThorstenMeyerAI.com
