LLMSpace: Carbon Footprint Modeling for Large Language Model Inference on LEO Satellites
First framework to quantify lifecycle carbon of LLM inference in space, addressing a gap for sustainable AI deployment.
LLMSpace models the carbon footprint of LLM inference on LEO satellites, revealing trade-offs between carbon emissions, latency, and hardware design. It shows that space-based inference can reduce terrestrial energy use but incurs significant embodied carbon from launches and hardware.
Large language models (LLMs) impose rapidly growing energy demands, creating an emerging energy and carbon crisis driven by large-scale inference. Solar-powered, AI-enabled low Earth orbit (LEO) satellites have been proposed to mitigate terrestrial electricity consumption, but their lifecycle carbon footprint remains poorly understood due to launch emissions, satellite manufacturing, and radiation-hardened hardware requirements. This paper presents \textit{LLMSpace}, the first carbon modeling framework for LLM inference on AI-enabled LEO satellites. LLMSpace jointly models operational and embodied carbon, peripheral subsystems, radiation-hardened accelerators and memories, and LLM-specific workload characteristics such as prefill-decode behavior and token generation. Using realistic satellite and GPU configurations, LLMSpace reveals key trade-offs among carbon footprint, inference latency, hardware design, and operational lifetime for sustainable space-based LLM inference. Source code: https://github.com/UnchartedRLab/LLMSpace.