Towards a future space-based, highly scalable AI infrastructure system design
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915646383587328 |
|---|---|
| author | Arcas, Blaise Agüera y Beals, Travis Biggs, Maria Bloom, Jessica V. Fischbacher, Thomas Gromov, Konstantin Köster, Urs Pravahan, Rishiraj Manyika, James |
| author_facet | Arcas, Blaise Agüera y Beals, Travis Biggs, Maria Bloom, Jessica V. Fischbacher, Thomas Gromov, Konstantin Köster, Urs Pravahan, Rishiraj Manyika, James |
| contents | If AI is a foundational general-purpose technology, we should anticipate that demand for AI compute -- and energy -- will continue to grow. The Sun is by far the largest energy source in our solar system, and thus it warrants consideration how future AI infrastructure could most efficiently tap into that power. This work explores a scalable compute system for machine learning in space, using fleets of satellites equipped with solar arrays, inter-satellite links using free-space optics, and Google tensor processing unit (TPU) accelerator chips. To facilitate high-bandwidth, low-latency inter-satellite communication, the satellites would be flown in close proximity. We illustrate the basic approach to formation flight via a 81-satellite cluster of 1 km radius, and describe an approach for using high-precision ML-based models to control large-scale constellations. Trillium TPUs are radiation tested. They survive a total ionizing dose equivalent to a 5 year mission life without permanent failures, and are characterized for bit-flip errors. Launch costs are a critical part of overall system cost; a learning curve analysis suggests launch to low-Earth orbit (LEO) may reach $\lesssim$\$200/kg by the mid-2030s. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_19468 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Towards a future space-based, highly scalable AI infrastructure system design Arcas, Blaise Agüera y Beals, Travis Biggs, Maria Bloom, Jessica V. Fischbacher, Thomas Gromov, Konstantin Köster, Urs Pravahan, Rishiraj Manyika, James Distributed, Parallel, and Cluster Computing Emerging Technologies Machine Learning Space Physics B.m If AI is a foundational general-purpose technology, we should anticipate that demand for AI compute -- and energy -- will continue to grow. The Sun is by far the largest energy source in our solar system, and thus it warrants consideration how future AI infrastructure could most efficiently tap into that power. This work explores a scalable compute system for machine learning in space, using fleets of satellites equipped with solar arrays, inter-satellite links using free-space optics, and Google tensor processing unit (TPU) accelerator chips. To facilitate high-bandwidth, low-latency inter-satellite communication, the satellites would be flown in close proximity. We illustrate the basic approach to formation flight via a 81-satellite cluster of 1 km radius, and describe an approach for using high-precision ML-based models to control large-scale constellations. Trillium TPUs are radiation tested. They survive a total ionizing dose equivalent to a 5 year mission life without permanent failures, and are characterized for bit-flip errors. Launch costs are a critical part of overall system cost; a learning curve analysis suggests launch to low-Earth orbit (LEO) may reach $\lesssim$\$200/kg by the mid-2030s. |
| title | Towards a future space-based, highly scalable AI infrastructure system design |
| topic | Distributed, Parallel, and Cluster Computing Emerging Technologies Machine Learning Space Physics B.m |
| url | https://arxiv.org/abs/2511.19468 |