Google’s Blue Origin rocket choice and Suncatcher’s unsolved cost problem

Google has picked Blue Origin to launch hardware for Project Suncatcher, its plan to run AI computing in orbit. A ride to space does not solve the.

Google has picked Blue Origin to launch hardware for Project Suncatcher, its plan to run AI computing in orbit. A ride to space does not solve the project’s main problem: whether putting data centres in space can ever be cheap enough to compete.

Key takeaways

  • Ars Technica reports that Google has turned to Blue Origin for a rocket to support Project Suncatcher, its research effort on space-based computing.
  • Project Suncatcher explores whether solar-powered satellites carrying Google’s AI chips could one day do some of the work that data centres on the ground do now.
  • Ars Technica describes Suncatcher as a long-term moonshot, meaning it is exploratory research rather than a product with a delivery date.
  • Securing a launch provider does not answer the deeper question of whether computing in orbit can become cheaper than building more data centres on the ground.

Google’s launch arrangement with Blue Origin

Ars Technica reports that Google has chosen Blue Origin, the private spaceflight company, to provide a rocket for Project Suncatcher. The source material gives no commercial terms, launch dates, payload details or vehicle. It is therefore not known how many launches are involved, what they will carry or when they will fly.

Even so, choosing a launch provider is a real step. Before this, Suncatcher existed mainly as research analysis and plans for early test hardware. A company that books a rocket has to finish building the satellites, pass them through safety reviews and integration with the launch vehicle, and fit them into the provider’s schedule. These are the practical stages where many space projects slip.

The choice also matters for the launch market. Blue Origin has spent years building heavy-lift capability and is trying to win commercial customers in a sector one competitor has dominated. A contract with one of the world’s largest technology companies, however small, gives Blue Origin a customer whose demand could in theory grow a great deal if space-based computing ever reached commercial scale. Whether this deal comes with any such commitment is not stated in the material available.

Project Suncatcher

Project Suncatcher is Google’s research programme on whether machine-learning workloads could run on satellites powered by sunlight. The idea is simple in outline. Google would place its tensor processing units (TPUs), the custom chips it designs for AI, on satellites with solar panels. The satellites would fly in close formation and send data between each other by laser.

The appeal comes from the pressures on computing on the ground. Training and running large AI models takes huge amounts of electricity. Ground-based data centres compete for grid capacity, land, cooling water and planning approval. In a suitable orbit, solar panels can get sunlight almost all the time, without night or cloud cover. Supporters argue that this could make orbit an attractive place to generate power for computing over the long run.

The engineering problems are serious. The chips have to survive radiation that wears down ordinary electronics. Waste heat cannot be carried away by air or water in a vacuum, so it has to be radiated into space, which needs large, heavy radiators. The satellites must also hold tight formations so their laser links can move data at the speeds AI workloads need. Hardware that breaks in orbit generally cannot be repaired, so replacing it means launching new satellites.

Ars Technica calls Suncatcher a long-term moonshot. That fits how Google has generally presented its most speculative research: as long-shot bets that may never become products, rather than near-term business lines.

Launch costs: the long-term problem

The problem the Ars Technica headline refers to can be read from the economics of the idea. Every kilogram of chips, solar panels, radiators and structure has to be lifted into orbit. On the ground, a data centre’s main costs are buildings, power and hardware. In orbit, the cost of launch is added on top, and it falls again each time hardware wears out and has to be replaced.

For space-based computing to compete with ground-based facilities, launch costs would need to fall far below today’s typical levels and stay there. That depends on factors Google does not control. Rockets would need to be fully and quickly reusable, launch rates would need to rise sharply, and competition among providers would need to push prices down. Booking a rocket from Blue Origin gets Google to orbit, but it does nothing on its own to make that trip cheap enough for large-scale use.

The source material does not show whether Google has updated its own estimates of when, or if, the economics might work. That remains an open question, and the answer will turn on how the launch industry develops over many years, not on any single contract.

Prototype hardware and testing

Experimental programmes like this usually start with a small number of demonstration satellites before any larger build-out. The aim of such missions is to learn, not to compute at scale. They check how chips behave under real radiation, whether the thermal designs work and whether satellites can hold formation and keep their optical links running.

The material available does not confirm what Google plans to put on the Blue Origin rocket or how it fits into any testing sequence. That is unknown. In general, data from early flights tends to decide whether such a project gets more funding or is quietly scaled back.

What the Blue Origin choice shows about orbital computing

Taken together, these points show that a speculative idea has moved into a more concrete phase without its central doubt being settled. Google picking Blue Origin shows the company is willing to spend money and engineering effort to put Suncatcher hardware into space. Calling the project a moonshot shows that nobody expects a quick payoff.

The episode also shows how closely AI and spaceflight are now linked. Demand for computing has grown fast enough that companies are looking at options that once seemed far-fetched, including putting processors in orbit. Whether those options are realistic depends mostly on the space industry: on cheaper rockets, more frequent launches and hardware that lasts in a harsh environment.

The physics of solar power in space looks favourable, and the engineering problems look hard but possibly solvable. The economics are the weak link, and they depend on launch prices that no single customer can set. A launch provider is one requirement. Making the idea economically viable is a separate, harder problem.

Sources and further reading

  • Ars Technica: space coverage reporting that Google has chosen Blue Origin to launch Project Suncatcher hardware.
  • Google research publications: background material describing the aims and technical approach of Project Suncatcher.
  • Blue Origin company materials: general information on the company’s launch vehicles and commercial services.
  • Industry analysis of launch economics: wider reporting on reusable rockets and the cost of reaching orbit.

Surfaced from the rss:arstechnica signal “space computing launch deal”. AI-assisted draft, editorially reviewed.

Visited 1 times, 1 visit(s) today
share this recipe:
Facebook
X
WhatsApp
Telegram
Email
Reddit