China: Chinese tech firms advance inference computing technologies to meet AI demands.
· flipboard.com · popularity 19/100 · 0 similar stories today · AI & Data Science (Artificial Intelligence)
Read the original at flipboard.com →
Summary
As demand for inference computing rises, Chinese companies are fighting through bottlenecks by upgrading technologies. Inference computing is the process of running a trained artificial intelligence or machine learning model on new, unseen data to generate real-time predictions or content. Lanxin Computing in south China's Shenzhen City has developed chips that can run large models on all-in-one AI computers. Xie Shengluo, co-founder, Lanxin Computing: "In inference computing, the key factor now is token-generation efficiency. With the rapid growth of AI agents, demand is surging across government services, healthcare, finance and other sectors. Our products are currently in short supply." Electricity supply for chips is also facing pressure because conventional silicon-based solutions are increasingly challenged by energy losses and heat management. At the Shenzhen Pinghu Laboratory, researchers are working on a gallium nitride crystal film that substantially improves both efficiency and power-delivery capacity when used to charge GPUs. Wan Yuxi, director, Shenzhen Pinghu Laboratory: "It means by 2035, we estimate that AIDCs (AI computing data centers) in China alone could require as much electricity as the output of five Three Gorges hydropower stations. After a technology upgrade, the energy losses alone could be reduced by an amount equivalent to the electricity generated by half a Three Gorges station." Wan says that the efficiency of silicon-based gallium nitride devices has risen from 85 percent to about 95. [Restrictions: No access Chinese mainland]
Found in: Machine Learning. Rating basis: single outlet; outlet tier 2 (unlisted); 1 feeds; feed position 8.