Towards a future space-based, highly scalable AI infrastructure system design

• URL: http://arxiv.org/abs/2511.19468v1
• Date: Sat, 22 Nov 2025 00:28:08 GMT
• Title: Towards a future space-based, highly scalable AI infrastructure system design
• Authors: Blaise Agüera y Arcas, Travis Beals, Maria Biggs, Jessica V. Bloom, Thomas Fischbacher, Konstantin Gromov, Urs Köster, Rishiraj Pravahan, James Manyika,
• Abstract summary: This work explores a scalable compute system for machine learning in space.<n>It uses fleets of satellites equipped with solar arrays, inter-satellite links using free-space optics, and Google tensor processing unit (TPU) accelerator chips.<n>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.
• Score: 1.6764108000740328
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: 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.

Related papers

• Lunar Silicon Cavity [3.4037613543160337]
  Moon's permanently shadowed regions (PSRs) are among the coldest places in the Solar System.<n>We propose to deploy a passive, ultrastable optical resonator in these regions that will enable laser systems with unprecedented phase-coherence.<n>Such a stable laser will form a basic quantum technology infrastructure in space to serve many applications.
  arXiv  Detail & Related papers  (2026-02-06T03:27:59Z)
• ASTREA: Introducing Agentic Intelligence for Orbital Thermal Autonomy [51.56484100374058]
  ASTREA is the first agentic system executed on flight-heritage hardware for autonomous spacecraft operations.<n>We integrate a resource-constrained Large Language Model (LLM) agent with a reinforcement learning controller in an asynchronous architecture tailored for space-qualified platforms.
  arXiv  Detail & Related papers  (2025-09-16T08:52:13Z)
• From Cell Towers to Satellites: A 2040 Blueprint for Urban-Grade Direct-to-Device Mobile Networks [0.0]
  We present the first end-to-end system architecture for a fully orbital telco.<n>We analyze spectral efficiency, beam capacity, and link budgets under dense urban conditions.<n>We propose a staged 15-year roadmap from today's fallback D2D systems to autonomous orbital overlays delivering 50-100 Mbps to handhelds in megacities.
  arXiv  Detail & Related papers  (2025-07-13T22:37:02Z)
• Stitching Satellites to the Edge: Pervasive and Efficient Federated LEO Satellite Learning [1.3121410433987561]
  This paper proposes a novel FL-SEC framework that empowers satellites to execute large-scale machine learning (ML) tasks onboard efficiently. Key components include personalized learning via divide-and-conquer, which identifies and eliminates redundant satellite images, and orbital model retraining, which generates an aggregated "orbital model" per orbit and retrains it before sending to the ground station. Our approach dramatically reduces FL convergence time by nearly 30 times, and satellite energy consumption down to as low as 1.38 watts, all while maintaining an exceptional accuracy of up to 96%.
  arXiv  Detail & Related papers  (2024-01-28T02:01:26Z)
• Orbital AI-based Autonomous Refuelling Solution [6.776059370975249]
  This paper documents the development of a proposed AI-based navigation algorithm intending to mature the use of on-board visible wavelength cameras as a main sensor for docking and on-orbit servicing (OOS) Multiple convolutional neural network backbone architectures are benchmarked on synthetically generated data of docking manoeuvres with the International Space Station (ISS) The integration of the solution with a physical prototype of the refuelling mechanism is validated in laboratory using a robotic arm to simulate a berthing procedure.
  arXiv  Detail & Related papers  (2023-09-20T21:25:52Z)
• A CubeSat platform for space based quantum key distribution [62.997667081978825]
  We report on the follow-up mission of SpooQy-1, a 3U CubeSat that successfully demonstrated the generation of polarization-entangled photons in orbit. The next iteration of the mission will showcase satellite-to-ground quantum key distribution based on a compact source of polarization-entangled photon-pairs. We briefly describe the design of the optical ground station that we are currently building in Singapore for receiving the quantum signal.
  arXiv  Detail & Related papers  (2022-04-23T06:28:43Z)
• Deep Learning for Real Time Satellite Pose Estimation on Low Power Edge TPU [58.720142291102135]
  In this paper we propose a pose estimation software exploiting neural network architectures. We show how low power machine learning accelerators could enable Artificial Intelligence exploitation in space.
  arXiv  Detail & Related papers  (2022-04-07T08:53:18Z)
• Deep Learning Aided Routing for Space-Air-Ground Integrated Networks Relying on Real Satellite, Flight, and Shipping Data [79.96177511319713]
  Current maritime communications mainly rely on satellites having meager transmission resources, hence suffering from poorer performance than modern terrestrial wireless networks. With the growth of transcontinental air traffic, the promising concept of aeronautical ad hoc networking relying on commercial passenger airplanes is potentially capable of enhancing satellite-based maritime communications via air-to-ground and multi-hop air-to-air links. We propose space-air-ground integrated networks (SAGINs) for supporting ubiquitous maritime communications, where the low-earth-orbit satellite constellations, passenger airplanes, terrestrial base stations, ships, respectively, serve as the space-, air-,
  arXiv  Detail & Related papers  (2021-10-28T14:12:10Z)
• Predicting Solar Flares with Remote Sensing and Machine Learning [0.0]
  Destruction of critical systems would lead to food shortages and an inability to respond to emergencies. A solution to this impending problem is proposed herein using satellites in solar orbit that continuously monitor the Sun. A system of systems approach will allow enough warning for safety measures to be put into place mitigating the risk of disaster.
  arXiv  Detail & Related papers  (2021-10-14T18:28:28Z)
• Entanglement demonstration on board a nano-satellite [76.84500123816078]
  Quantum networks for secure communication can be realised using large fleets of satellites distributing entangled photon-pairs between ground-based nodes. This paper describes a miniaturised, polarization entangled, photon-pair source operating on board a nano-satellite.
  arXiv  Detail & Related papers  (2020-06-25T14:11:56Z)
• Integrating LEO Satellite and UAV Relaying via Reinforcement Learning for Non-Terrestrial Networks [51.05735925326235]
  A mega-constellation of low-earth orbit (LEO) satellites has the potential to enable long-range communication with low latency. We study the problem of forwarding packets between two faraway ground terminals, through an LEO satellite selected from an orbiting constellation. To maximize the end-to-end data rate, the satellite association and HAP location should be optimized. We tackle this problem using deep reinforcement learning (DRL) with a novel action dimension reduction technique.
  arXiv  Detail & Related papers  (2020-05-26T05:39:27Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.