TSMC and Intel, which has not disclosed its recent sales performance in mainland China, did not immediately respond to requests for comment.
Intel’s chief financial officer David Zinsner said at a conference earlier this month that the company had seen more customers looking to its Gaudi chips as an alternative to processors that were in short supply.
A race by Chinese tech firms to build their own large language models (LLMs) has sparked robust demand for advanced chips used to train algorithms that support intelligent chatbots similar to Microsoft-backed OpenAI’s ChatGPT.
The move allowed Chinese companies to open their chatbots to the world’s largest internet population, setting the stage for businesses to compete for a slice of the burgeoning generative AI market.
However, amid growing demand for AI technology, the US has restricted China’s access to advanced chips, such as Nvidia’s A100 and H100 processors, which are widely used to train LLMs.
Intel, which generated 27 per cent of its total revenue last year in China, according to its annual report, introduced Gaudi2 after Nvidia launched its remodelled A800 and H800, which are also not subject to US export controls.
At the unveiling of Gaudi2 in China, Intel executive vice-president Sandra Rivera said the chip was designed to lower the barrier to entry and enhance mainland clients’ ability to “deploy AI through cloud and smart-edge technologies, helping build up China’s AI future”.
The news is published by EMEA Tribune & SCMP