Download App

[CCTV Finance: Buying Computing Power is Now as Convenient as Buying Electricity] This year, my country's daily word usage has exceeded 140 trillion, a more than 1,000-fold increase in two years. Many small and medium-sized enterprises (SMEs) need to purchase computing power. To make buying computing power as convenient as buying electricity, the southern region is exploring a new model of computing-electricity collaborative trading. In Shenzhen, an artificial intelligence company consumes approximately 1 billion words daily. The company recently completed its first collaborative computing-electricity trading transaction: publishing computing power needs, timeframes, and budgets, and automatically matching idle computing power at a more advantageous price through an operating platform. "Through the new trading model, the cost is reduced by about 25% compared to the original. Demanders can place orders, and suppliers can bid for them on the platform." In this transaction, the computing power seller was a data center in Guizhou. This idle computing power was listed on the platform and the trading was completed at a lower price than the market price. Currently, this trading platform connects to computing power resources from multiple data centers in five southern provinces. The platform uses an artificial intelligence model to predict electricity prices to guide scheduling. When electricity prices are low, computing power tasks are prioritized; when electricity prices are high, non-urgent tasks are suspended or migrated. For example, if a user submits an AI training task, the system might find that electricity prices are lower in Guizhou, so it might reroute the task to Guizhou. Li Minhong, Marketing Director of China Southern Power Grid, said, "By connecting more buyers and sellers, we can make buying computing power as simple as online shopping: order on demand, pay as you use, and have transparent billing. Buying computing power is as simple as buying electricity."