China’s AI laboratory DeepSeek, whose AI types wreaked havoc on Wall Street last week, may use Chinese GPUs for future models, suggests a report. As investors began to wonder whether billions of dollars in capital expenditures were necessary for AI growth, DeepSeek’s ability to apparently reduce costs significantly over American types while using GPUs from chip big NVIDIA Corporation upended areas. Instead of NVIDIA’s offered language, Compute Unified Device Architecture ( CUDA ), the model development firm’s ability to use low-level programming language called Parallel Thread Execution ( PTX ) on NVIDIA’s GPUs is at the heart of the company’s model development.
According to a report, DeepSeek Was Rely On PTX Expertise to Get Higher Performance Out Of Domestic Chinese GPUs.
The Semiconductor Manufacturing International Corporation ( SMIC ), China’s largest foundry, has frequently been the target of US sanctions, which have prevented ASML, a Dutch company, from acquiring the most modern chipmaking equipment. These tools are essential for producing the most cutting-edge chips, with some reports suggesting that SMIC is only allowed to produce chips using the more recent 7-nanometer process technology family.
The company’s ability to use a programming language called Parallel Thread Execution or PTX has been a crucial factor in DeepSeek’s ability to match achievement of some Eastern AI models while using fewer and less performant NVIDIA GPUs. There are fewer designers who are capable of using PTX, which gives engineers more control over a device’s fundamental operations than other languages.

Due to US GPU restrictions, DeepSeek’s choice to use the notoriously difficult PTX resulted from its ability to secure either minimal amounts of GPUs or only those with lower performance. According to a statement from the Hong Kong push, the company’s expertise in using PTX may allow it to rely on private Chinese GPUs for future versions.
According to the report, PTX will allow DeepSeek’s engineers to extract more performance from local Chinese GPUs because it enables programmers to firmly control the operation of a GPU. The 5-nanometer production method network is used to create the Hopper H800 GPUs by NVIDIA, according to DeepSeek. They were established in 2023 and had at the time employed the most recent production process technology in the sector.
DeepSeek may use comparable programs with Taiwanese Graphics by understanding NVIDIA’s PTX application and how owners control a device’s features, according to the report. Any Chinese GPUs that are built using older production process systems will likely be significantly less performant than NVIDIA’s most recent Blackwell GPUs.