VANDENBERG SPACE FORCE BASE — Google will launch its first experimental satellite for Project Suncatcher on October 1. The spacecraft, named MVP, is approximately the size of a refrigerator and contains four of Google’s custom Tensor Processing Unit (TPU) AI accelerators.
The goal of the MVP mission is to measure how well TPUs handle the physical stress of spaceflight, radiation, and thermal extremes. The satellite will be launched from Vandenberg Space Force Base near Santa Barbara, California. MVP contains four specialized computer chips called tensor processing units, which have the computing power of one server in a data center. The satellite’s solar panels supply approximately one kilowatt of power.
The spacecraft bus for the MVP satellite was built by Planet Labs. James Mason is the chief space officer at Planet Labs. James Mason said: "Back then, it seemed further off. That was really before large language models took off and AI demand really started going exponential."
Project Suncatcher aims to design orbital AI data centers as an alternative to terrestrial data centers. Former Google CEO Eric Schmidt has floated the idea of orbital data centers as an alternative to data centers on Earth. Blaise Agüera y Arcas is a Google vice president and AI researcher who leads a team studying intelligence.
He conceived Project Suncatcher after attending a gathering of entrepreneurs and AI researchers where the main topic was the escalating energy needs of AI systems. In May 2025, Agüera y Arcas pitched Project Suncatcher to Google CEO Sundar Pichai and co-founder Sergey Brin. "There are long-standing ideas in science fiction about using stars for computation," he said.
Google opted to integrate its chips into a satellite that Planet Labs had already built for an early test to accelerate its efforts. The initial plan for Project Suncatcher was to launch two custom satellites in 2027. Travis Beals is the Senior Director of Product Management for Project Suncatcher.
"Exploring space as a viable location for scalable AI compute won’t happen all at once," Beals said. "It takes methodical engineering, starting with proving our hardware can handle the physical and unpredictable realities of operating in orbit."
"This first launch is about seeing what works, identifying points of failure, and applying those findings to future missions," he said. Google executive Travis Beals stated the satellite is designed to operate for roughly one year. The MVP satellite will operate for just a few months.
The satellite will operate for a year but is set to circle Earth for up to six years. The chips in the MVP satellite can only run for about 15 minutes before they need to be turned off to cool down.
Technicians performed a vibration test on the MVP satellite in a San Francisco lab to simulate launch conditions. Google’s MVP satellite underwent vibration testing simulating rocket launch conditions, with TPU chips experiencing up to 50–100 g-force loads. These loads exceeded typical launch stresses.
The MVP satellite passed the vibration test with screws remaining tight and chips intact. Google is planning to test a cooling system involving heat pipes and radiators while the chips are in orbit. Suncatcher features a layered cooling system to manage heat in space, as conventional fans cannot operate there. Google’s thermal management system for space-based TPUs includes heat pipes and radiators. It requires 1.3 square meters of radiator area to dissipate heat. This design was validated in a thermal vacuum chamber simulating space conditions.
James Manyika is Google’s senior vice president for research. "We don’t expect, to be perfectly frank, that we’ll have anything usefully operational in the next few years," Manyika said. "Remember how Google was researching for like 15 years before anything showed up? I think this is going to look like that." Google plans to put two satellites into orbit next year.
Google wants to eventually have networks of orbiting satellites that communicate over high-speed laser connections. Google estimates the cost of orbital data centers will become roughly equivalent to terrestrial ones in the mid-2030s. Pichai suggested that orbital computing could become mainstream within a decade.
Brandon Lucia is a professor of electrical and computer engineering at Carnegie Mellon University. "If you do this on a large scale, there are additional engineering problems," he said. "That is uncharted waters."
Why It Matters
Project Suncatcher represents a shift toward orbital computing as a potential solution to the energy demands of artificial intelligence. By testing custom TPU chips in space, Google seeks to validate hardware resilience against radiation and thermal extremes. The mission builds on prior concepts from former executives and responds to the exponential growth in AI computational needs.
The collaboration with Planet Labs leverages existing satellite manufacturing expertise to accelerate development. While immediate operational utility is not expected, the project establishes a foundation for future orbital data centers. Cost projections suggest parity with terrestrial facilities may arrive in the mid-2030s, marking a long-term strategic investment in space-based infrastructure.
Timeline
Blaise Agüera y Arcas conceived Project Suncatcher after attending a gathering of entrepreneurs and AI researchers where the main topic was the escalating energy needs of AI systems. On February 1, 2025, Google began testing AI chips at Crocker Nuclear Laboratory in Davis, California. On that same date, Google used a cyclotron at Crocker Nuclear Laboratory to blast chips with a radiation dose equivalent to five years in space. Also on February 1, 2025, Google found that restarting the chips could usually reset bit flips caused by radiation.
On May 1, 2025, Blaise Agüera y Arcas pitched Project Suncatcher to Google CEO Sundar Pichai and co-founder Sergey Brin. During that meeting, Brin said, "OK, so this is a good idea. Let’s talk about how we’re doing it." On that same date, the MVP satellite passed the vibration test with screws remaining tight and chips intact.
What's New
Later reporting detailed the engineering philosophy behind the mission. Travis Beals said, "It takes methodical engineering, starting with proving our hardware can handle the physical and unpredictable realities of operating in orbit." He added, "This first launch is about seeing what works, identifying points of failure, and applying those findings to future missions."
James Manyika provided perspective on the timeline, stating, "Remember how Google was researching for like 15 years before anything showed up? I think this is going to look like that." He also noted, "Exploring space as a viable location for scalable AI compute won’t happen all at once." Technical details emerged regarding pre-launch testing. Google’s MVP satellite underwent vibration testing simulating rocket launch conditions, with TPU chips experiencing up to 50–100 g-force loads, exceeding typical launch stresses, and passing tests with all components intact.
Background information showed that Planet Labs, known for its satellite imaging constellation, has previously launched over 200 satellites since 2016, including the SkySat series for high-resolution Earth observation, which demonstrates its capability to build and operate complex orbital platforms. Brandon Lucia offered an external view, saying, "If you do this on a large scale, there are additional engineering problems."
How Sources Differ
Sources differ on the timeline for launching two satellites.
forum Comments (0)
No comments yet. Be the first to comment.