Google has partnered with Jeff Bezos-founded Blue Origin to launch satellites carrying its artificial intelligence (AI) computing hardware into space in 2029. The mission is part of Project Suncatcher, Google’s ambitious research initiative to explore whether solar-powered satellites can eventually operate as interconnected data centres in orbit.
Under the launch services agreement announced on November 19, 2026, Blue Origin’s New Glenn rocket will carry Google’s computing satellites into sun-synchronous orbit during the first quarter of 2029. Each satellite is expected to have computing capacity roughly equivalent to a conventional data centre rack, powered by Google’s custom Tensor Processing Units (TPUs). The project could eventually offer a new way to expand AI computing infrastructure without relying entirely on electricity, land and cooling systems on Earth.
Key Takeaways
- Google has selected Blue Origin’s New Glenn rocket for a dedicated Project Suncatcher launch.
- The mission is targeted for the first quarter of 2029.
- Satellites will carry Google’s custom Tensor Processing Units (TPUs).
- Each satellite will have computing capacity roughly comparable to a data centre rack.
- Project Suncatcher aims to explore solar-powered AI computing infrastructure in orbit.
- Google has already launched an initial prototype satellite with Planet Labs and SpaceX.
- Cooling, radiation protection, satellite communication and launch costs remain major engineering challenges.
Google Selects Blue Origin’s New Glenn Rocket for 2029 Mission
The partnership marks an important milestone in Google’s effort to investigate whether large-scale AI computing can eventually move beyond traditional data centres.
Google confirmed that Blue Origin’s New Glenn heavy-lift rocket will conduct the first dedicated heavy-lift launch for Project Suncatcher. The mission is scheduled for the first quarter of 2029.
According to the companies, the rocket will deploy multiple computing satellites into sun-synchronous orbit. This orbital arrangement allows satellites to receive sunlight regularly, making it suitable for experiments involving solar-powered infrastructure.
Google’s Senior Director for Paradigms of Intelligence, Travis Beals, described the partnership as an important step in the company’s long-term plan to expand AI infrastructure using solar energy available in orbit.
Blue Origin also highlighted New Glenn’s heavy-lift capabilities and large payload capacity as important advantages for the mission.
The agreement does not mean Google is ready to operate commercial data centres in space. Instead, it represents a research milestone designed to test the technology and infrastructure required for that possibility.
What Is Google’s Project Suncatcher?
Project Suncatcher is Google’s long-term research initiative exploring how machine-learning computing could operate using satellites equipped with Google’s AI processors.
The basic idea is to place computing hardware in orbit and power it using solar energy. Rather than building every new data centre on land, Google wants to investigate whether groups of satellites can work together to process AI workloads.
Traditional data centres require large amounts of electricity, physical space, cooling infrastructure and network connectivity. As demand for AI services grows, these requirements are becoming increasingly important for technology companies.
Google believes space could offer an alternative source of energy for future computing infrastructure.
Satellites in suitable orbits can receive sunlight for much longer periods than solar panels on the ground. Google has previously estimated that solar panels in low Earth orbit could generate up to eight times more energy than comparable panels on Earth under certain conditions.
The company wants to investigate whether that advantage can make space-based computing technically and economically practical.
How the Project Could Work
The proposed system would involve several interconnected satellites carrying AI computing hardware.
- Solar power: Satellites would use solar panels to generate electricity.
- AI processors: Each satellite would carry Google’s Tensor Processing Units.
- Onboard computing: The processors would execute machine-learning workloads directly in orbit.
- Satellite networking: Laser communication systems would connect satellites and allow them to exchange data.
- Thermal management: Radiators and heat pipes would help remove heat generated by the processors.
The long-term ambition is to create a network of satellites capable of working together rather than operating as isolated computing machines.
Google’s AI Chips Will Power the Space Satellites
A major component of Project Suncatcher is Google’s Tensor Processing Unit technology.
TPUs are specialised processors designed to accelerate machine-learning workloads. Google uses them across its AI infrastructure for tasks such as training and running machine-learning models.
Unlike general-purpose processors, TPUs are designed to handle the mathematical operations involved in AI computing efficiently.
For Project Suncatcher, Google is exploring whether these processors can continue operating reliably under the unusual conditions of space.
The company has already conducted tests involving its Trillium-generation TPUs. These tests examined how the hardware responds to radiation exposure and other environmental stresses.
In September 2026, Google said initial proton-beam testing indicated that its Trillium TPUs could withstand radiation exposure exceeding the total ionising dose expected during a five-year space mission.
That result was an encouraging early finding, but laboratory testing cannot replace long-term operation in orbit.
The company therefore needs further experiments to understand how the hardware performs under real-world space conditions.
Google Has Already Tested Its First AI Satellite in Orbit
Project Suncatcher has moved beyond theoretical research.
Google and Planet Labs launched the programme’s first prototype satellite on October 1, 2026, aboard a SpaceX rocket.
On October 30, Google announced that the satellite had successfully powered up its onboard hardware and completed an important initial AI computing test.
The company reported that it had run its first machine-learning workloads using an onboard TPU. The system also generated a response from Google’s Gemma 4 open model after receiving a prompt from a ground station.
Google described the achievement as its first AI response from a data-centre-class TPU operating in orbit.
The initial satellite is designed to help engineers understand whether AI processors can survive launch conditions and operate under radiation and temperature extremes.
The next major research milestone is to test two interconnected satellites, with Google previously identifying 2027 as a target for that stage of the programme.
The Blue Origin launch in 2029 would then provide a dedicated heavy-lift opportunity to deploy a larger group of computing satellites.
Why Google Wants to Build Data Centres in Space
The rapid expansion of generative AI has increased demand for computing power across the technology industry.
Companies such as Google, Microsoft, Amazon and other cloud providers are investing heavily in data centres to support AI models and cloud services.
However, building these facilities requires access to reliable electricity, land, water and cooling infrastructure.
Some regions are already facing pressure from the electricity requirements of large data centre projects.
Google’s space-computing research explores whether orbital infrastructure could eventually reduce some of these constraints.
Potential Advantages
1. Access to solar energy
Satellites in suitable orbits can receive sunlight for extended periods. This could allow solar panels to provide a consistent source of electricity for computing systems.
2. Reduced dependence on terrestrial power grids
An orbital data centre would not need to connect to an electricity grid on Earth. Its power system would operate independently using solar energy and onboard storage.
3. Alternative locations for AI infrastructure
Space could eventually provide another environment for expanding computing capacity as demand for AI services increases.
4. New satellite infrastructure opportunities
If the technology becomes practical, companies could develop specialised satellites designed not only for communications and observation but also for computing.
However, these potential benefits remain research objectives. They do not yet demonstrate that orbital data centres will be cheaper or more efficient than facilities on Earth.
The Biggest Challenge: Cooling AI Chips in Space
One of the most difficult engineering problems facing Project Suncatcher is heat management.
AI processors generate substantial heat when running demanding workloads. Conventional data centres use fans, liquid cooling systems and other technologies to prevent processors from overheating.
Space presents a different challenge because there is no surrounding air to carry heat away through convection.
In a vacuum, heat must be transferred and released through other methods, including radiators.
Google has been investigating heat pipes and radiator systems to manage the thermal output of its computing hardware.
The company has also tested cooling technologies inside thermal vacuum chambers designed to reproduce some of the conditions found in space.
These systems will need to work reliably over long periods without the maintenance and repair options available to terrestrial data centres.
A cooling failure could damage expensive computing hardware and compromise the entire satellite mission.
Radiation and Satellite Communication Pose Additional Risks
Cooling is only one of several challenges.
Electronics operating in space face exposure to radiation from the Sun and cosmic sources. High-energy particles can damage electronic components and cause errors in computing systems.
Google’s proton-beam experiments are intended to understand how its TPUs respond to this exposure.
The company also needs to develop reliable communication systems between satellites.
Future versions of Project Suncatcher could involve groups of satellites working together on large AI workloads. This requires them to exchange significant amounts of data with very low latency and high reliability.
Google has been exploring laser-based communication systems for this purpose.
The company has explained that maintaining high-bandwidth laser connections between moving satellites requires extremely precise positioning and tracking.
This is important because AI computing systems often need to exchange large volumes of information between processors. If satellites cannot communicate efficiently, the overall performance of an orbital computing network could be limited.
Why the Blue Origin Partnership Matters
The agreement gives Blue Origin an important role in Google’s long-term space infrastructure research.
New Glenn is Blue Origin’s heavy-lift launch vehicle, designed to transport large payloads into orbit.
For Google, selecting a dedicated heavy-lift launch provider offers a way to deploy multiple satellites as part of a coordinated mission.
The partnership also demonstrates how demand for AI infrastructure could create new opportunities for the commercial space industry.
Launch providers may increasingly compete for contracts involving data centre satellites, communication networks and other advanced computing infrastructure.
Google’s earlier prototype mission used SpaceX, while the upcoming dedicated heavy-lift mission will use Blue Origin.
The choice of different partners reflects Google’s broader work with satellite and launch providers rather than a simple replacement of one company by another.
For Blue Origin, the agreement represents a significant future customer mission as it seeks to expand New Glenn’s commercial launch business.
Google Is Not the Only Company Exploring Space-Based Data Centres
The idea of moving computing infrastructure into orbit is attracting interest across the technology and space industries.
SpaceX, startups such as Starcloud and other companies are exploring different approaches to space-based computing.
The interest is driven partly by the growing electricity requirements of AI systems and the possibility of using solar energy in orbit.
However, companies are pursuing different technical approaches, and most projects remain at the research or early development stage.
Google’s advantage is its experience in AI hardware, cloud infrastructure and machine-learning software. Its research also provides an opportunity to test whether existing data centre-class processors can function in space.
Nevertheless, the industry still needs to solve difficult problems involving launch economics, radiation protection, thermal management, networking and satellite maintenance.
The Bigger Picture
Google’s partnership with Blue Origin reflects a broader shift in how technology companies are thinking about the future of AI infrastructure.
Until recently, the expansion of AI computing was largely associated with building more powerful processors and constructing larger data centres on Earth. Project Suncatcher introduces another possibility: placing computing hardware in orbit and using solar energy to power it.
If successful, this approach could eventually create a new category of infrastructure connecting artificial intelligence, cloud computing and commercial space operations.
However, the economic case remains uncertain. Launching satellites, replacing damaged hardware and operating complex communication systems could make orbital computing expensive. Even if solar power is abundant, the total cost of transporting and maintaining computing equipment in space must be considered.
Looking Ahead
Google’s next milestones will involve testing its prototype satellites, improving radiation protection and developing reliable cooling and laser communication systems. The planned 2027 interconnected-satellite experiment will be particularly important because it could demonstrate whether multiple orbital computing systems can exchange data and operate together. The dedicated Blue Origin New Glenn launch in the first quarter of 2029 is expected to advance the programme further by deploying additional computing satellites.
For now, Project Suncatcher remains a long-term research effort rather than a commercial replacement for terrestrial data centres. If Google can overcome the technical and financial challenges, space-based computing could eventually become another option for expanding AI infrastructure. The 2029 mission will be an important test of that ambition, but the development of commercially viable orbital data centres is likely to require years of additional engineering and investment.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.



