Page 364 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

Scalability and Adaptability: The platform supports integration of heterogeneous devices from
multiple vendors. Modular architecture and flexible APIs enable easy adaptation and expansion
as service demands evolve, ensuring the system's long-term viability.

Deployed across China Telecom’s national network, managing 5.13 million 4G/5G sectors
(including 2 million 5G), 3,800 equipment rooms, and 46,000 servers. Achieving over 900
million kWh in annual energy savings and reducing CO₂ emissions by more than 500,000 tons.
This project demonstrates a scalable, replicable model for achieving large-scale, sustainable
digital infrastructure. It highlights the transformative role of AI-powered systems in enhancing
operational efficiency, reducing energy consumption, and supporting digital innovation within
the ICT sector.
Partners

The project involves collaboration with key partners across the ICT ecosystem, including leading
equipment providers such as Huawei, ZTE [1], CICT Mobile, and Ericsson [2], which contribute
to standardizing and developing energy-saving functionalities, as well as supporting device
maintenance and deployment. Integrated IDC service providers, including China Telecom
Tianyi IoT and China Telecom Tianyi Cloud, contribute expertise in data center design,
construction, and commercial services. Telecom operators such as China Telecom and China
Unicom [8] oversee large-scale information infrastructure and serve as primary clients for AI-
driven energy efficiency initiatives.

2�2 Benefits of the use case
1. The project significantly improves energy efficiency and green operations of information
infrastructure through the application of AI, big data, and other digital technologies.
It ensures the green and high-quality development of next-generation infrastructure
while reducing annual greenhouse gas emissions by over 500,000 tons, contributing
meaningfully to global climate change mitigation and environmental sustainability.
2. The project has established one of the world’s largest cross-disciplinary AI-powered
energy-saving platforms, enabling large-scale, safe, and efficient energy optimization
across multi-vendor, multi-domain information infrastructure. It reshapes green operation
and maintenance processes, reducing deployment and operational costs of energy-
saving systems by 90%, and empowering enterprises in their green transformation, cost
reduction, efficiency improvement, and sustainable development.
3. Through massive-scale green digital transformation and AI-driven upgrades of information
infrastructure, the project drives upstream and downstream industrial upgrades, including
for equipment manufacturers and service providers, promoting a green and sustainable
transition across the global information and communications technology (ICT) sector.

2�3 Future Work

From 2025 to 2027, this project will introduce large language models to reconstruct the AI
energy-saving brain, enhance knowledge cognition and intelligent decision-making capabilities,
realize self-intelligence and energy savings, and continuously expand the scale of platform
applications while improving energy efficiency. Aligned with ITU’s mission for sustainable ICT
development(e.g., Recommendation ITU-T L.1470 on AI for energy efficiency in networks), it
is expected to cover more than 99% of China Telecom's base stations, 95% of class AB IDC
business server rooms, and more than 50% of cloud hosts by 2025, saving more than 1.5
billion kWh of electricity per year, which is equivalent to reducing carbon dioxide emissions
by 850,000 tons. Globally, it will disseminate methodologies to establish AI-driven green

328