Green AI Carbon Optimizer: Carbon-Efficient Training Location Recommendation and Global AI Energy Demand Forecasting

Yuxin Chen; Hao Gao; Chujie Zou

cs.PFApr 6, 2026

Green AI Carbon Optimizer: Carbon-Efficient Training Location Recommendation and Global AI Energy Demand Forecasting

Authors: Yuxin Chen, Hao Gao, Chujie Zou

Organizations: University of Helsinki, Finland · Independent Researcher

Abstract

AI training and deployment consume substantial electricity, but carbon outcomes remain weakly integrated into routine model development decisions. This paper presents Green AI Carbon Optimizer with two primary contributions: (i) a carbon aware cloud region recommendation method for training workloads, and (ii) a power law forecasting pipeline for global AI energy demand. For location recommendation, we combine regional grid carbon intensity, renewable share, and data center Power Usage Effectiveness (PUE) into a unified scoring model across 100+ regions from major cloud providers. For a reference workload (8*A100, 100h), estimated emissions in our sampled regions range from 7.74kg to 272.00kg CO2. Selecting the best region instead of the worst corresponds to a 97.2% reduction relative to the worst case. Ablation shows that ranking by renewable share alone can select regions with higher CO2 emissions than rankings that include grid carbon intensity. For forecasting, we fit a power law relation between parameter count and training energy using 26 anchor models. We combine this fit with scenario assumptions on model growth, hardware efficiency, and training frequency, and evaluate sensitivity to inference ratio and ecosystem scaling. Across scenarios, projected 2030 demand ranges from 7TWh to 1,436TWh under the stated assumptions, highlighting the importance of deployment choices, model scaling discipline, and transparent energy reporting.

Figures & tables

Model	CV $R^{2}$	CV MAE	CV RMSE
Linear Reg. (Scaling Law)	0.764 $\pm$ 0.203	0.317	0.421
Random Forest	0.728 $\pm$ 0.134	0.393	0.489

TABLE I: Model comparison on log-space energy prediction (5-fold CV,

n=26

). Units: log 10 (kWh).

Region	Provider	$I_{r}$	PUE	CO 2 (kg)
Montréal	GCP	0.022	1.10	7.74
Montréal	AWS	0.022	1.15	8.10
Stockholm	AWS	0.040	1.10	14.08
São Paulo	GCP	0.090	1.15	33.12
Oregon	GCP	0.120	1.10	42.24
Singapore	AWS	0.410	1.35	177.12

TABLE II: Top and bottom regions for the reference workload (8

\times

A100, 100 h).

Fig. 1: Projected global AI energy demand under four scenarios (log-scale $y$ -axis). Parameters: $\rho_{\text{inf}}=50$ , $\mu_{\text{eco}}=10$ , $\overline{\mathrm{PUE}}=1.3$ .

Scenario	Energy (TWh)	CO 2 (MT)	% Global
Baseline	1,436	718	5.1%
Moderate	134	67	0.5%
High Efficiency	24	12	0.1%
Green AI	7	4	0.03%

TABLE III: Top-down forecast: projected 2030 AI energy (

\rho_{\text{inf}}=50

\mu_{\text{eco}}=10

\overline{\mathrm{PUE}}=1.3

Fig. 2: Sensitivity of 2030 global AI energy (TWh) to $\rho_{\text{inf}}$ and $\mu_{\text{eco}}$ at $\overline{\mathrm{PUE}}=1.3$ (Moderate scenario).

Model	Light	Medium	Heavy
Gemini 2.5 Pro	0.80	2.50	5.00
GPT-5	5.00	18.35	40.00
Claude 3.7 Sonnet	0.84	2.00	5.52
DeepSeek-R1	23.80	30.00	33.60

TABLE IV: Per-query energy by model and complexity (Wh). Values are compiled from published efficiency studies [ 18 ] and public model metadata [ 11 ] ; entries without direct public measurements (including GPT-5 and Gemini 2.5 Pro) are treated as order-of-magnitude extrapolated estimates.

Energy-Efficient ML

Energy-efficient machine learning reduces the energy and environmental costs of training and inference while preserving useful model performance.

See all 249 papers

AI Inference as Relocatable Electricity Demand: A Latency-Constrained Energy-Geography Framework
Apr 30, 2026Xubin Luo, Cheng YangEnergy-Efficient ML Cost-Aware Inference
School of Computer Science and Artificial Intelligence, Southwestern University of Finance and Economics, Chengdu, China
InFactPlanner: Planning Sustainable Geo-Distributed LLM Data Centers
Aug 13, 2026Nicoletta Tsiopani, Moysis Symeonides, George Pallis +1Energy-Efficient ML LLM Inference Serving
Department of Computer Science University of Cyprus
A Game-Theoretic Framework for Incentive-Compatible AI training Under Renewable-Energy Constraints
Sep 14, 2026Konstantinos Varsos, Ramin Khalili, Adamantia Stamou +2Mechanism Design Energy-Efficient ML
Athens University of Economics and Business, Greece · Independent researcher, Munich Germany

Apr 30, 2026J/K move · Enter open · S save

AI Inference as Relocatable Electricity Demand: A Latency-Constrained Energy-Geography Framework

Xubin Luo, Cheng Yang

School of Computer Science and Artificial Intelligence, Southwestern University of Finance and Economics, Chengdu, China

Energy-Efficient ML Cost-Aware Inference Distributed Inference

Open paper AskPDF arXiv

AI inference is becoming a persistent and geographically distributed source of electricity demand. Unlike many traditional electrical loads, inference workloads can sometimes be executed away from the user-facing service location, provided that latency, state locality, capacity, and regulatory constraints remain acceptable. This paper studies when such digital relocation of computation can be interpreted as latency-constrained relocation of electricity demand. We develop an energy-geography framework for geo-distributed AI inference. The framework models a three-layer architecture of clients, service nodes, and compute nodes, and formulates inference placement as a constrained optimization problem over electricity prices, marginal carbon intensity, power usage effectiveness, compute capacity, network latency, and migration frictions. The key object is the energy-latency frontier: the marginal cost and carbon benefit unlocked by relaxing inference latency budgets. The paper makes four contributions. First, it distinguishes physical electricity transmission from digital relocation of electricity-consuming computation. Second, it formulates a geo-distributed inference placement model with feasibility masks and migration frictions. Third, it introduces operational metrics, including relocatable inference demand, energy return on latency, carbon return on latency, and a relocation break-even condition. Fourth, it provides a transparent stylized simulation over representative global compute regions to show how heterogeneous latency tolerance separates workloads into local, regional, and energy-oriented execution layers. The results show that latency relaxation expands feasible geography, while migration frictions, egress costs, state locality, legal constraints, and capacity limits can sharply reduce realized benefits.

Abstract

Energy-Efficient ML

Explore similar work

AI Inference as Relocatable Electricity Demand: A Latency-Constrained Energy-Geography Framework

InFactPlanner: Planning Sustainable Geo-Distributed LLM Data Centers

A Game-Theoretic Framework for Incentive-Compatible AI training Under Renewable-Energy Constraints