8 articles tagged with #carbon-emissions. AI-curated summaries with sentiment analysis and key takeaways from 50+ sources.
A comprehensive study of 403 U.S. hyperscale data centers reveals they consumed 68-99 TWh of electricity between May 2024 and April 2025, generating 37-54 million metric tons of CO2 emissions. The findings show HDC carbon intensity is 48% higher than the national grid average, driven by rapid AI infrastructure expansion and heavy reliance on fossil fuels.
The United Nations is calling for AI companies to disclose their environmental costs by 2030, recognizing that unchecked AI growth threatens global resources. The directive emphasizes the need for greater transparency and sustainable practices to address the sector's escalating environmental footprint.
Researchers propose a federated learning framework that combines ARIMA, GARCH, LSTM-Attention, and XGBoost models to forecast global carbon emissions while preserving data privacy. The system enables collaborative forecasting across distributed clients without sharing raw data, achieving R² values averaging 0.73 across 14 test clients.
Researchers propose Carbon-Aware Governance Gates (CAGG), an architectural framework that integrates carbon budgeting and energy tracking into GenAI development workflows. The approach addresses the paradox where governance mechanisms designed to ensure responsible AI development inadvertently increase computational demands and environmental impact through repeated inference cycles and validation processes.
The EU is considering extending carbon emission charges to foreign airline flights, a policy move that could provoke international disputes and reshape global aviation economics. This expansion of the EU's emissions trading system (ETS) aims to reduce aviation's carbon footprint but risks retaliatory measures from other nations and potential disruptions to transatlantic and global air travel.
Researchers demonstrate that quantization and local inference techniques can reduce LLM energy consumption and carbon emissions by up to 45% without sacrificing performance. The findings address growing sustainability concerns surrounding generative AI deployment, offering practical optimization strategies for resource-constrained environments.
Researchers developed a deep learning framework using Continuous Wavelet Transform and CNNs for heat demand forecasting in district heating systems. The model achieved 36-43% reduction in forecasting errors compared to existing methods, reaching up to 95% accuracy in predicting day-ahead heat demand across multiple European cities.
Researchers developed a bi-level AI optimization framework using reinforcement learning to improve winter road maintenance operations on UK highway networks. The system strategically partitions road networks and optimizes vehicle routing while reducing travel times below two hours and minimizing carbon emissions.
