China Claims 2nm AI Chip Development, Mass Production Uncertain

Chinese AI chip startup Dishan Technology announced it has entered the verification stage for a prototype of its 2nm process-based AI Graphics Processing Unit (GPU) semiconductor. Founded in 2021, the company is led by experts selected as 'Fellows,' the highest grade by the Institute of Electrical and Electronics Engineers (IEEE), and aims to build China's self-sufficient semiconductor system and reduce reliance on foreign technology. Dishan Technology's 2nm AI chip development is designed to be compatible with NVIDIA's AI software ecosystem, which is expected to facilitate transitions for existing customers using NVIDIA products.

However, the semiconductor has not yet entered the trial production phase, and local reports suggest it will take another one to two years for commercialization.

Securing a foundry capable of mass-producing 2nm AI chips within China is highlighted as the biggest challenge. China's top foundry, SMIC, has its most advanced process technology limited to the 7nm level, presenting high technical barriers for 2nm mass production. This is evident even with Huawei's latest AI chips being produced using SMIC's 7nm process.

Even if technical challenges are overcome, it is impossible for Dishan Technology to contract production with advanced overseas foundries such as TSMC or Samsung Electronics due to US technology export restrictions. Consequently, Dishan Technology is expected to inevitably face significant difficulties due to US sanctions on its path to reaching the mass production stage for AI semiconductors.