Satellites share wireless power to speed up ai model responses
Low-Latency MoE Inference in Space Data Centers: A Wireless-Power-Transfer Perspective
Distributed, Parallel, and Cluster ComputingArtificial IntelligenceNetworking and Internet Architecture
Summary
Satellites running parts of large AI models sometimes face power shortages because some parts get busier than others, while idle parts have extra solar energy. The authors suggest a way to wirelessly send solar power between satellites using lasers to balance power where it's needed most. This clever sharing reduces delays in AI responses from space by making sure active satellites have enough power. Their tests show this approach can make AI model replies up to 29% faster in groups of satellites.
What this means in practice
- •For satellite system engineers: Deploy wireless power sharing to reduce AI inference delays across satellite clusters managing distributed large language models.
- •For wireless power system designers: Design inter-satellite laser power transfer systems that optimize power distribution under realistic energy and communication constraints.
Authors
Haotian Zheng, Zhanwei Wang, Kaibin Huang
Abstract
Space data centers (SpaceDCs) are emerging as platforms for large language models (LLMs), supported by continuous solar-energy harvesting and high-speed interconnections. The mixture-of-experts (MoE) architecture supports distributed LLM inference by deploying experts across satellites. However, skewed expert activation creates uneven computation power demands despite uniform solar-power supply at the dawn-dusk sun-synchronous orbit (SSO). Consequently, active experts remain power-constrained while harvested power at idle experts is underused. These reduce inference efficiency and increase latency. To address this supply-demand mismatch, we propose the SpaceWPT framework, which uses inter-satellite laser-based wireless power transfer (WPT) to redistribute harvested power. The key design challenge is to balance the computation demands of power-donating and power-receiving satellites under WPT losses: transferring more power accelerates computation at the receiver but leaves less power for the donor. After deriving design principles that simplify WPT control, we characterize the resulting latency tradeoff through a tractable surrogate of average inference latency. Minimizing the surrogate yields a novel two-sided water-filling strategy that determines whether each active expert donates, retains, or receives power according to the water level and its WPT link efficiency. Simulations with Mixtral-8x7B and Phi-3.5-MoE show up to 29.25% lower inference latency than without WPT, with larger gains under more compact formations and higher transfer efficiencies.