AI

Google Selects Blue Origin’s New Glenn for Project Suncatcher: AI Data Centres Are Heading to Space

November 20, 2026
Google Project Suncatcher:
102
Views

Google is taking another step towards its ambitious plan to move artificial intelligence computing infrastructure beyond Earth. The company has partnered with Blue Origin to use its New Glenn rocket for the next major launch under Project Suncatcher, Google’s research initiative exploring solar-powered AI data centres in space.

Announced on November 19, 2026, the agreement marks an important milestone for the project. Blue Origin’s New Glenn rocket is expected to carry Google’s computing satellites into orbit in the first quarter of 2029.

Each satellite is planned to carry Google’s Tensor Processing Units (TPUs), with computing capacity roughly equivalent to a traditional data centre rack. The satellites will operate in sun-synchronous orbit, where they can receive sunlight for much of their orbital journey.

The idea is ambitious: instead of building every new AI data centre on Earth, Google wants to explore whether some computing workloads could eventually run in space using solar energy.

What Is Google’s Project Suncatcher?

Project Suncatcher is Google’s long-term research initiative to explore space-based computing infrastructure.

Artificial intelligence requires significant computing power. Training and running advanced AI models involves large numbers of specialised processors, electricity and cooling systems. As demand for AI services grows, technology companies are looking for ways to expand computing capacity without depending entirely on traditional data centres.

Google’s proposal is to place networks of small satellites equipped with its custom AI chips in orbit. These satellites would use solar panels to generate electricity and communicate with one another through high-speed optical links.

The project could eventually create a network of interconnected computing satellites that work together like a data centre in space.

Google first introduced the research concept in November 2025. Its technical proposal describes a system using solar-powered satellites, Google TPUs and optical links to connect the computing infrastructure.

You can read Google’s original Project Suncatcher research paper for a deeper look at the proposed architecture.

Why Has Google Selected Blue Origin’s New Glenn?

Google has selected Blue Origin’s New Glenn heavy-lift rocket for its first dedicated heavy-lift launch under Project Suncatcher.

New Glenn is designed to carry large payloads into orbit, making it suitable for missions that need to deploy multiple satellites or substantial space infrastructure.

The launch is currently targeted for the first quarter of 2029. It will place Google’s satellites into sun-synchronous orbit, with each satellite carrying computing capacity roughly equivalent to a data centre rack.

The agreement is significant because the project is moving beyond research discussions towards a dedicated launch plan.

Blue Origin’s launch vehicle also provides the payload capacity needed for Google’s longer-term experiments with space-based computing.

According to Blue Origin’s official announcement, the companies have entered into a launch services agreement for the mission.

Neither company has disclosed the financial value of the agreement or all the technical details of the satellite deployment.

How Will AI Data Centres Work in Space?

The concept behind Project Suncatcher is different from traditional satellite missions.

Most satellites collect information, provide communications or observe the Earth. Google’s proposed satellites would carry actual computing hardware capable of running machine-learning workloads.

Here is how the system is expected to work.

1. Solar Energy Will Power the Satellites

One of the biggest advantages of placing computing infrastructure in certain orbits is access to sunlight.

Solar panels on Earth are affected by night, weather and seasonal changes. Satellites in suitable orbits can receive sunlight for much longer periods.

Google’s research suggests that solar panels in space could receive up to eight times more solar energy per year than panels at mid-latitudes on Earth under the conditions considered in its study.

This could help provide electricity for AI processors without relying on traditional electricity grids.

2. Google TPUs Will Handle AI Computing

Each satellite is expected to carry Google’s Tensor Processing Units, or TPUs.

These are specialised processors designed to accelerate machine-learning workloads. They are already used across Google’s AI infrastructure.

Putting TPUs into orbit would allow the satellites to process information directly in space instead of sending all computing tasks back to ground-based data centres.

The company is investigating whether its AI chips can operate reliably under the demanding conditions of space.

3. Satellites Will Communicate Through Optical Links

A single satellite cannot provide the same computing capacity as a large data centre.

Google’s research therefore explores connecting multiple satellites through high-speed optical communication links.

These links would allow the satellites to exchange information and work together as a distributed computing network.

Maintaining reliable communication between satellites will be one of the most important technical challenges for the project.

4. Computing Workloads Could Run in Orbit

If the system works as intended, satellites could process AI workloads directly in space.

This could eventually provide additional computing capacity without requiring every new facility to be built on land.

However, the proposed network is still experimental. Google has not announced a commercial service or confirmed when a large operational orbital data centre could become available.

Why Is Google Exploring Space-Based AI Data Centres?

The growing demand for artificial intelligence is creating new challenges for data centre infrastructure.

Large AI facilities require electricity, land, cooling equipment and reliable connectivity. Technology companies are investing heavily in new data centres to support AI models, cloud services and other computing applications.

Project Suncatcher explores whether space could offer an alternative way to expand this infrastructure.

Some potential advantages include:

Access to solar energy: Satellites in suitable orbits can receive sunlight for extended periods, reducing dependence on electricity generated on Earth.

Reduced land requirements: Space-based infrastructure would not require large plots of land for traditional data centre buildings.

Additional computing capacity: A network of satellites could eventually provide another location for running certain AI workloads.

New infrastructure possibilities: Space-based computing could support future applications that benefit from processing data closer to satellites and other space systems.

These benefits are potential advantages rather than proven commercial outcomes. The cost and complexity of operating computing hardware in orbit remain major questions.

What Are the Biggest Challenges for Project Suncatcher?

Moving AI infrastructure into space is not as simple as launching computers into orbit.

Google must address several technical and financial challenges before the concept can become practical.

1. Cooling Hardware in Space

AI processors generate significant heat when operating.

On Earth, data centres use air conditioning, liquid cooling and other systems to maintain safe operating temperatures. Space does not provide the same cooling environment.

Satellites need carefully designed thermal systems to remove heat from their electronics. This is particularly important when several powerful processors operate together.

2. Radiation Protection

Space exposes electronic components to radiation that can affect their operation.

Radiation can cause memory errors, damage electronic circuits and reduce hardware reliability.

Google’s research includes testing its TPUs to understand how they respond to radiation exposure and whether they can operate reliably over extended periods.

3. Launch and Maintenance Costs

Building satellites and sending them into orbit is expensive.

Even if solar energy is more readily available in space, the cost of launching, operating and eventually replacing the hardware could make the overall system difficult to compete with terrestrial data centres.

Google’s research paper discusses the importance of significantly reducing launch costs for the concept to become economically practical.

4. Communication Between Satellites

A distributed computing network needs reliable communication.

The satellites must exchange data quickly enough to support machine-learning workloads. This requires accurate positioning, reliable optical links and effective coordination between spacecraft.

Any communication delays or failures could affect the performance of the entire network.

Project Suncatcher vs Traditional AI Data Centres

The two approaches have different infrastructure requirements.

FeatureTraditional AI Data CentresProject Suncatcher
LocationOn EarthIn space
Power sourceElectricity grids and other energy sourcesSolar panels
CoolingAir and liquid cooling systemsSpace-based thermal management
Computing hardwareGPUs and AI acceleratorsGoogle TPUs
ConnectivityFibre and data centre networksInter-satellite optical links
MaintenanceTechnicians can access facilitiesRequires specialised space operations
Current statusCommercially operationalExperimental research

The comparison shows why Project Suncatcher is interesting, but also why it is difficult to implement.

Traditional data centres already have established supply chains, maintenance processes and network infrastructure. Space-based computing would need to overcome substantial engineering and cost challenges before becoming a practical alternative.

When Will Google’s AI Satellites Launch?

Google and Blue Origin are targeting the first quarter of 2029 for the dedicated New Glenn mission.

This is an important step in the programme, but it does not mean that a full commercial space-based data centre will be operational in 2029.

The mission is part of a longer research programme. Google is working towards testing the technology, demonstrating reliable satellite computing and understanding the challenges involved in scaling the system.

The company has also worked with satellite partners on earlier prototype efforts. These tests are intended to help researchers understand how computing hardware behaves during launch and in the space environment.

A successful 2029 mission would provide another important opportunity to test the concept in orbit.

What Project Suncatcher Means for the Future of AI

Google’s partnership with Blue Origin highlights how the AI industry is exploring new ways to meet growing computing demand.

Instead of relying only on larger data centres on Earth, companies are investigating alternative infrastructure models, including space-based computing.

Project Suncatcher is still a research initiative, and many technical and financial questions remain unanswered. However, the planned New Glenn launch gives Google a concrete milestone in its long-term effort.

If the technology becomes practical, it could introduce a new approach to powering and operating AI infrastructure.

For now, the biggest achievement is not the arrival of commercial AI data centres in space. It is the progress towards testing whether powerful computing systems can operate reliably beyond Earth.

The first quarter of 2029 could become an important milestone in that journey.

Final Thoughts

Project Suncatcher shows that the future of AI infrastructure may involve more than building larger data centres on Earth.

By combining Google’s TPUs, solar-powered satellites and Blue Origin’s New Glenn rocket, the project aims to explore a new way of delivering computing power.

There are still major challenges involving cooling, radiation, communication and launch costs. But the partnership represents a significant step in turning Google’s space-based AI research into a practical experiment.

The idea of running AI data centres in space is no longer just a concept on paper. Google is now working towards putting more of the technology into orbit.

Article Categories:
AI

Leave a Reply

Your email address will not be published. Required fields are marked *

The maximum upload file size: 3 GB. You can upload: image, audio, video, document, spreadsheet, interactive, text, archive, code, other. Links to YouTube, Facebook, Twitter and other services inserted in the comment text will be automatically embedded. Drop file here