SpaceX to Launch Google AI Chips and Satlyt Software in Orbital Data Center Test

The uncrewed Transporter-18 mission will carry Planet Labs satellites, including a solar-powered prototype equipped with Google tensor processing units, CNBC reported. It marks the first orbital test for Project Suncatcher, a Google “moonshot” initiative Alphabet first revealed in November 2025. In a post announcing the launch, Alphabet said the project is a means of “exploring whether space could one day host scalable machine learning infrastructure.”

Alphabet said that low Earth orbit offers advantages for AI infrastructure. Satellites there can “access near-constant sunlight, generating up to eight times more solar power than on Earth,” the company said. Eventually, Alphabet expects to “link together multiple constellations of satellites, allowing them to manage larger AI workloads while in orbit.” The Google parent has already tested TPUs running AI workloads at a facility at the University of California at Davis, but it does not yet know how the chips will perform in low Earth orbit.

Alphabet is a significant investor in SpaceX, which went public in June in a record IPO, and its stake in the company is valued at more than $82 billion, CNBC reported. The businesses are close partners even as their AI divisions compete. SpaceX, led by Elon Musk, has also announced plans to build and launch its own orbital data centers, made up of satellite swarms developed in Redmond, Washington, equipped with graphics processing units and solar arrays it plans to produce with Tesla. Musk said earlier this year that data centers in space would be the cheapest way to train AI, “and that will be true within two years, maybe three at the latest.” SpaceX COO Gwynne Shotwell said in September that the company will deploy “supercompute in space” in 2027.

Industry experts caution that space-based data centers remain a distant prospect, if they become feasible at all. Rocket launches are still capacity-constrained and expensive, and orbital data centers would require cooling systems and chips that can withstand extreme temperatures and radiation, as well as protection from clutter and orbital debris. The Transporter-18 flight will be the second potentially historic launch in a single day for SpaceX, which is slated to carry astronauts to the International Space Station from Florida's Space Coast for NASA on Thursday, beginning a six-month mission.

TechCrunch reported that Satlyt, founded by former Google and SpaceX product manager Rama Afullo, has raised an $8 million seed round. The company, with headquarters in Sunnyvale, California, and Nairobi, is building software for satellites that it hopes will let many satellites share the work of large computing jobs in space. Afullo previously worked in Google's cloud computing business before a brief stint at SpaceX's Starlink communications network in 2024. He said he pitched the idea internally at both companies and was turned down.

Satlyt is not planning to build its own spacecraft, unlike SpaceX, Google, Starcloud, and Cowboy Space Company. Afullo's team is building software to operate AI models on satellites, an approach he compared to VMware and Snowflake. “The folks like SpaceX who are doing orbital data centers—if they are the iPhone, we'll build Android as a horizontally integrated, open ecosystem,” he said. In other words, Satlyt wants to offer software that works across many companies' satellites.

Satlyt has already flown its software on two demonstration missions. Its initial focus is operational efficiency for spacecraft, which typically depend on flight controllers on the ground. Sending data back to Earth, known as downlink, is expensive and often slow, so using AI to make decisions on the spacecraft can produce savings when resolving anomalies or processing sensor readings onboard instead of on the ground. Earlier this year, Satlyt deployed Google DeepMind's Gemma AI model onboard a spacecraft operated by Momentus. The model cut the size of a transmission about onboard software errors by more than 60%, an efficiency Afullo said can save hundreds of thousands of dollars per satellite each year.

This week's launch will put Satlyt's software onboard a spacecraft built by TakeMe2Space, an Indian startup that builds computing hardware for satellites. The mission involves three customers: NASA, which is paying the company to test protocols for cloud computing in space; Stellerian, a startup focused on space surveillance, or tracking objects in orbit, that wants to test image processing workloads; and TakeMe2Space itself, to show the satellite can host other companies' software.

Few high-powered GPUs are in orbit today, and Satlyt's near-term opportunity lies in the kinds of workloads it is already pursuing. Its next project is to demonstrate a shared computing system, or cloud, spanning two different satellites, which it expects to attempt next year. If successful, the company says it could build a true compute cloud in orbit as a third-party provider to spacecraft builders. “The way to think about it is that we turn your satellite into a revenue-generating managed service,” Afullo said.

Satlyt's seed round was led by Houston-based Non Sibi Ventures. Partner Bernard Harris, who served as a NASA astronaut for more than 20 years, helped ground the firm's conviction in the space-based business, TechCrunch reported. Non Sibi partner Kent Lucas said one reason the firm invested was that Satlyt does not depend on the most ambitious vision of space data centers, which faces uncertain economics and a launch bottleneck. “We don't need data centers in space for Rama to be wildly successful, right?” Lucas said. “It can just be driven by the number of satellites going up.” Afullo hopes Satlyt will be live on 20% of satellites by the end of the decade, when he expects virtually every spacecraft builder to put GPUs and similar advanced processors in their satellites. “If you're putting up a satellite without putting up a GPU on it, at the very least, you're doing yourself a disservice, right?” he said.