Google tests orbital AI compute with one 1-kilowatt server riding a SpaceX rocket
Google flies one server's worth of TPUs on 1 kilowatt of solar power on Oct. 1, its first hardware test for orbital AI data centers. The flight tests chip survival and cooling while SpaceX's and Blue Origin's far larger constellations still wait on FCC review.
The Investor · Invest desk
What happened
- Google will launch a satellite carrying its TPU AI chips into low Earth orbit on Oct. 1, aboard a SpaceX Falcon 9 on the Transporter-18 rideshare, in a project built with Planet.
- The satellite has the computing power of one data-center server, runs on 1 kilowatt of solar power to answer simple AI queries, and is expected to operate for a year.
- SpaceX has asked the FCC to approve up to 1 million data-center satellites and Blue Origin close to 52,000, and both applications still await FCC review.
- Google plans to put two more Project Suncatcher satellites into orbit in 2027.
Why it matters
- What's now possible capability If the satellite lasts its expected year, Google will hold a year of orbital operating data on its own chips by about the time SpaceX says its Nvidia-powered Starmind satellites start deploying, in the fourth quarter of 2027.
- What it limits constraint With the chips reportedly limited to about 15 minutes of running before a cool-down, a satellite cannot answer AI queries continuously until the heat pipes and radiators work in orbit.
- What it costs cost The Motley Fool tells investors to expect no near-term financial contribution from Suncatcher, so Alphabet is funding the program as research with no revenue line attached.
Next story · Invest · Oct 1
Google's first orbital TPU test puts one server's worth of chips on a SpaceX rideshareConfirmed4 publishers
Google has already tested the launch half of the experiment on the ground. It says it shook a satellite on all three axes to mimic a launch and put the TPUs through a proton beam, and the chips survived [8]. The ride itself lasts about 10 minutes, with components taking 50 to 100 times the force of gravity [4]. What the flight adds is the year after: radiation and heat swings over months, and the first in-orbit trial of the heat pipes and radiators Google built to cool the chips [2].
Scale is easiest to compare in kilowatts. The Electric Power Research Institute estimates data centers use 4% to 5% of US electricity now and could use as much as 17% by 2030 [16]. Eric Schmidt, Google's former chief executive, told a House committee that data centers could need 29 more gigawatts than they do now by 2027 [14]. At 1 kilowatt a satellite, meeting that would take 29 million of them [15]. SpaceX's entire filing, built to the same size, would add 1 gigawatt, about a twenty-ninth of Schmidt's figure [17]. A production satellite need not be a one-server box, and the filings as reported do not state power per satellite.
Valuations and timelines for orbital compute are far apart. Musk announced in February a merger of SpaceX and xAI in a deal valued at $1.25 trillion, and said "space-based AI is obviously the only way to scale" [10]. Bezos has called timelines of two or three years a "little ambitious" [18]. Google's own analysis, published with its 2025 announcement, put the cost of an orbital data center roughly level with one on Earth by the mid-2030s [11]. James Manyika, Google's senior vice president for research, told The New York Times: "We don't expect, to be perfectly frank, that we'll have anything usefully operational in the next few years" [19].
If the TPUs run and the cooling works, Google carries the only flight data on its own chips into its next launches. Should the chips survive and the cooling fail, the eight-times solar yield Google cites for low Earth orbit [5] buys power the processors cannot use for long. A third outcome makes the orbital data secondary. If the FCC clears SpaceX's application and SpaceX builds at anything near the million satellites it asked for [7], launch rate decides the contest, and one server's test settles little.
I think the October flight is a component test of Google's own chips, and the contest it opens is over cooling, where Google is the one company of the three flying test hardware, according to Cryptopolitan [13]. Google's effort goes into chips and cooling, with the launch bought as a slot on a rideshare [1]. Schmidt bought the launch firm Relativity Space [20]. This view is wrong if a year of orbital data lets Google move its cost-parity estimate earlier than the mid-2030s [11].
What to watch
- Whether Google reports its TPUs running past the reported 15-minute limit once the heat pipes and radiators are working in orbit.
- FCC action on SpaceX's million-satellite application and Blue Origin's application for close to 52,000 satellites.
- Whether the two 2027 Suncatcher satellites keep their date, or change in size, after the first satellite's results come in.
Information, not investment advice. This story is general information about companies and markets, drawn from cited reporting. It is not a recommendation to buy, sell or hold any security; check the cited sources before acting.
Clarity's read
What the record supports and how the coverage leans. The claims behind it follow.
Reality
- Evidence60
- Adoption4
- Hype gap+35
- Incentives55
- Confidence62
Perspective Coverage
5 publishers- Builder
- Builder 41%
- Operator
- Operator 22%
- Investor
- Investor 37%
Claim ledger
Ranked by verification strength, evidence, and original report placement.
- [1]
Google will launch a satellite carrying its TPU AI chips into low Earth orbit on October 1 aboard a SpaceX Falcon 9 as part of the Transporter-18 rideshare mission, in a project developed with the imaging company Planet.
- [2]
Travis Beals, the senior director leading Project Suncatcher, has reportedly stated that the chips can run for only about 15 minutes before they need to shut off and cool down; Google is testing an orbital cooling setup built around heat pipes and radiators.
- [3]
Google plans to put two more satellites in orbit in 2027.
- [4]
A trip to low Earth orbit lasts about 10 minutes, and individual components can face forces of 50 to 100 times gravity.
- [5]
Google says satellites in low Earth orbit see near-constant sunlight and can generate up to eight times more solar energy than panels on the ground.
- [6]
The satellite will have the computing power of one data center server and operate on 1 kilowatt of solar power to answer simple AI queries; it is expected to operate for a year.
- [7]
SpaceX has asked the FCC to approve up to 1 million data-center satellites to train xAI models; Blue Origin has asked the FCC to deploy close to 52,000 solar-powered satellites for AI computing; both are still awaiting FCC review.
- [8]
Google said it already mimicked a launch by shaking a satellite on all three axes; the TPUs were also put through a proton beam and survived.
- [9]
Project Suncatcher is a research effort Google announced in November 2025.
- [10]
Elon Musk announced in February that he would merge SpaceX and xAI in a deal valued at $1.25 trillion, arguing that "space-based AI is obviously the only way to scale."
- [11]
Google suggested in an analysis published alongside its 2025 announcement that running a data center in orbit could become roughly comparable in cost to running one on Earth by the mid-2030s.
- [12]
SpaceX plans Starmind AI satellites powered by Nvidia chips, which SpaceX believes will begin deploying in the fourth quarter of 2027.
- [13]
SpaceX and Blue Origin are both still at the regulatory stage, while Google is the one actually flying test hardware next week.
- [14]
Eric Schmidt told a House committee that data centers could need 29 more gigawatts of power than they do now by 2027.
- [15]
At 1 kilowatt per satellite, supplying 29 additional gigawatts would take 29 million satellites.
- [16]
Data centers are estimated to consume 4% to 5% of US electricity today, expected to increase to as much as 17% by 2030, according to the Electric Power Research Institute.
- [17]
One million satellites at 1 kilowatt each would supply 1 gigawatt, about one twenty-ninth of Schmidt's 29-gigawatt figure.
- [18]
Jeff Bezos has said orbital data centers are a reasonable long-term goal but thinks timelines of two or three years are a "little ambitious."
- [19]
"We don't expect, to be perfectly frank, that we'll have anything usefully operational in the next few years," James Manyika, Google's senior vice president for research, said in an interview with The New York Times.
- [20]
Former Google CEO Eric Schmidt has bought the launch firm Relativity Space.
- [21]
Investors shouldn't expect any near-term financial contributions from Project Suncatcher.
- [22]
A satellite launched Oct. 1 in the year after Suncatcher's November 2025 announcement and operating for its expected year would still be generating data around October 2027, the start of the fourth quarter in which SpaceX expects Starmind deployment to begin.
Sources
5 independent publishers whose own reporting we read for this story.
- cryptopolitan.comGoogle jumps ahead of Elon Musk's SpaceX in space data center race
1 article · September 24, 2026
- fool.comGoogle Is Taking a Major Step Forward Toward Space-Based AI Data Centers | The Motley Fool
1 article · September 24, 2026
- livemint.com'One small step for TPUs': Google CEO Sundar Pichai announces Project Suncatcher to test AI compute in Space | Mint
1 article · September 24, 2026
- thedailyupside.comGoogle Takes Data Centers to the Stars to Solve Earthbound Problems
2 articles · September 24, 2026
- verdict.co.ukGoogle's Project Suncatcher to test AI chip performance in orbit - Verdict
1 article · September 25, 2026