AgentDecarbonizer: Carbon-Aware Execution for AI Agents

Leyi Yan; Shuangning Li; Sihang Liu

cs.LGAug 20, 2026

AgentDecarbonizer: Carbon-Aware Execution for AI Agents

Authors: Leyi Yan, Shuangning Li, Sihang Liu

Organizations: University of Waterloo, Canada · University of Chicago, USA

Abstract

AI agents extend large language models from single prompt-response interactions to long-running, goaldirected workflows that issue many model calls, invoke tools, and interact with external environments. These workflows enable tasks such as software repair, data analysis, and experiment management, but their repeated model invocations can incur substantial carbon emissions. This paper characterizes the carbon emissions of OpenClaw agent workloads using WildClawBench, and shows that emissions depend on token consumption, context cache reuse, and the carbon intensity of the grid. Our characterization identifies deadline flexibility as an opportunity for carbon-aware execution: agent tasks can wait for lower-carbon-intensity periods or shift to lower-carbon grids. However, doing so requires handling uncertain execution time for temporal shifting and cached context recomputation during spatial shifting. We present AgentDecarbonizer, a carbon optimizer for AI agents that runs alongside OpenClaw. Given a task prompt and user-specified deadline, AgentDecarbonizer conservatively estimates task duration and selects deadline-feasible execution schedules, while accounting for cache recomputation overhead during spatial shifting. Evaluated on WildClawBench workloads with 60 agent tasks across four grids, AgentDecarbonizer reduces carbon emissions by up to 57.9 % compared with a carbon-agnostic baseline and by up to 37.5 % compared with a baseline that selects the carbon-optimal grid at task start time.

Figures & tables

Figure 1. Example of an autonomous AI agent workflow on the OpenClaw platform [ 63 ] .

Category	# Tasks	Example tasks
Productivity Flow	10	Paper digest, PDF analysis
Code Intelligence	12	Website development, code debugging
Social Interaction	6	Chat extraction, calendar scheduling
Search & Retrieval	11	Data collection, fuzzy search
Creative Synthesis	11	Slides generation, posters generation
Safety Alignment	10	Skill injection filtering, password leak detection

Table 2. WildClawBench tasks and categories [ 13 ] .

Figure 7. AgentDecarbonizer overview.

Figure 8. Execution time estimator workflow.

Symbol	Description
$\mathcal{R}$	Candidate execution regions.
$\mathcal{H}$	Deadline-feasible candidate intervals.
$H$	Number of deadline-feasible candidate intervals.
$K$	Number of ordered execution intervals induced by $\widetilde{T}$ .
$r_{k}$	Region assigned to execution interval $k$ .
$t_{k}$	Candidate interval assigned to execution interval $k$ .

Table 3. Notation in the AgentDecarbonizer placement objective in Equation 6 .

Panel of Figure 9 — Figure 9. (a) Overall carbon savings, (b) grid usage, and (c) carbon intensity distribution. Model is GPT-5.4.

Figure 10. Example of agent execution optimized by AgentDecarbonizer. Start time Feb 4, 2026 at 16:00 UTC.

Panel of Figure 11 — Figure 11. (a) Execution time, (b) token usage, and (c) carbon emissions of Gemini 3.1 Pro vs GPT-5.4.

Figure 12. Emissions of single-grid only scenarios compared to AgentDecarbonizer. The model API is GPT-5.4.

Appendix figures & tables1 asset

Supplementary material from the paper’s appendix.

Appendix

Category	Typical steps/range	Reference
Single-context text transformation	1	IFEval [ 97 ]
Structured strict-format output	4	ToolLLM [ 68 ]
External retrieval/browser/API workflow	7.3	Mind2Web [ 12 ]
Multi-turn communication/coordination	8.93–15.39	MultiWOZ [ 10 ]
Repository/code/debug workflow	14.71	SWE-agent [ 87 ]
Multimodal visual/computer-use workflow	15	OSWorld [ 82 ]

Appendix

Table 4. Phase categories and number of steps per phase.

Carbon-Aware Computing

See all 20 papers

Carbon-Aware Routing for Function Calling in Edge-Cloud LLM Systems
Sep 11, 2026Aikaterini Maria Panteleaki, Varatheepan Paramanayakam, Spyros Tragoudas +1LLM Routing Carbon-Aware Computing
School of Electrical, Computer and Biomedical Engineering Southern Illinois University Carbondale
GAR: Carbon-Aware Routing for LLM Inference via Constrained Optimization
May 12, 2026Disha Sheshanarayana, Rajat Subhra Pal, Manjira Sinha +1LLM Inference Efficiency Constrained Optimization
Manipal University Jaipur · TCS Research
AI-CARE: Carbon-Aware Reporting Evaluation Metric for AI Models
Feb 17, 2026KC Santosh, Srikanth Baride, Rodrigue RizkCarbon-Aware Computing Energy-Efficient ML
USD Artificial Intelligence Research, Department of Computer Science University of South Dakota Vermillion, SD 57069, USA

Sep 11, 2026J/K move · Enter open · S save

Carbon-Aware Routing for Function Calling in Edge-Cloud LLM Systems

Aikaterini Maria Panteleaki, Varatheepan Paramanayakam, Spyros Tragoudas, Iraklis Anagnostopoulos

School of Electrical, Computer and Biomedical Engineering Southern Illinois University Carbondale

LLM Routing Carbon-Aware Computing Edge-Cloud Computing

Open paper AskPDF arXiv

Large Language Models (LLMs) with function-calling capabilities are becoming critical for modern agentic AI systems. Nevertheless, current deployments typically route inferences to powerful cloud-based models, incurring significant energy use and carbon emissions. We address this sustainability challenge with a carbon-aware routing framework that distributes function-calling queries across a three-tier edge-cloud architecture, combining edge and cloud LLMs on heterogeneous hardware. At its core, a lightweight k-NN predictor operating in a unified semantic-lexical embedding space estimates query-specific accuracy, delay, and power consumption on each edge tier. These predictions are then combined with real-time grid carbon intensity to route every query to the lowest-emission tier capable of executing it successfully. Evaluated on state-of-the-art function-calling benchmarks and LLM families, our framework matches cloud-level accuracy while reducing operational carbon emissions by

4\times

on average.

Abstract

Carbon-Aware Computing

Explore similar work

Carbon-Aware Routing for Function Calling in Edge-Cloud LLM Systems

GAR: Carbon-Aware Routing for LLM Inference via Constrained Optimization

AI-CARE: Carbon-Aware Reporting Evaluation Metric for AI Models