Green Computing in the Cloud Era: Highlights from the 2023 Inclusion·Bund Conference

Since the beginning of the year, the large‑model industry has exploded, with over 100 domestic models released, leading to severe GPU shortages and long reservation queues. This rapid growth has driven unprecedented demand for intelligent compute power and a sharp rise in energy consumption.

At the "Green Computing in the Cloud‑Native Era" forum held on September 7, 2023, guests from government, industry, academia, and research shared perspectives on how the industry can respond to these challenges.

Shanghai's Communications Administration highlighted the "Computing Power Pujiang Action Plan," proposing three key recommendations: accelerate R&D of green computing technologies, improve data‑center resource utilization and standardization, and deepen industry collaboration to build a green computing ecosystem.

The China SME Association, together with Shanghai partners, released ESG disclosure guidelines and rating standards to promote low‑carbon transformation for small and medium enterprises, encouraging them to adopt green computing practices and strive for carbon neutrality.

Ant Group emphasized an "end‑to‑end green compute" approach, focusing on greening the compute stage of applications to break efficiency bottlenecks and achieve collaborative green development, a view widely supported by conference participants.

End‑to‑end green computing spans the entire value chain—from power generation and compute infrastructure (including intelligent‑compute centers, hardware vendors, and cloud providers) to application‑level compute. The "White Paper on Green Computing for Application Layers" released by the China Academy of Information and Communications Technology identified application‑level green computing as the area with the greatest potential for efficiency gains.

On the hardware side, China Mobile showcased its new intelligent‑compute center built around five core pillars: new interconnect, new compute efficiency, new storage, new platform, and new energy‑saving technologies, aiming to construct an open, collaborative, and sustainable compute foundation.

AI techniques are also being applied to optimize data‑center energy use. Researchers from Tsinghua University and Intel presented two key technologies: high‑efficiency intelligent computing systems and AI‑driven quality‑improvement platforms that enable data‑driven decision making for carbon‑neutral computing.

Professor Jin Hai from Huazhong University of Science and Technology highlighted the low utilization rate of generated data (only 0.2% is effectively used) and proposed three near‑data processing modes—storage‑up, compute‑down, and distributed expansion—to overcome the "storage wall" and improve compute efficiency.

Ant Group's CTO He Zhengyu described the company's green, high‑efficiency intelligent compute system, which includes low‑carbon compute selection, global resource scheduling, workload optimization, and green monitoring, saving 4.1% of machine‑resource consumption in 2022 while supporting massive PV compute workloads.

Industry leaders from Alibaba Cloud IDC, Inspur, and PICC Technology discussed carbon‑aware metrics, green data‑center designs, and hardware innovations, emphasizing the need for standardized measurement and cross‑industry collaboration.

Overall, the conference underscored that, amid a persistent compute‑resource shortage, coordinated hardware‑software efforts and green‑focused standards are essential to break bottlenecks and drive sustainable, high‑performance computing.