AsianFin -- Two Chinese tech titans on Thursday were reported to adopt their in-house artificial intelligence (AI) chips to help ease their demand for Nvidia Corporation’s products amid trade tensions between China and U.S.
AI Generated Image
Alibaba Group and Baidu Corporation have been using their self-designed chips to train their AI models, partly replacing Nvidia chips, The Information cited people with knowledge of the matter, noting that neither of their chips are comparable with Nvidia’s most advanced products.
Alibaba has reportedly started using its own chips, the Zhenwu perse processing unit, for training its smaller models since early this year, and the chip is good enough as a rival of Nvidia’s H20, , the most powerful AI chip Nvidia is created to comply with the Biden-era export controls and allowed to sell in China.
Alibaba is also working on a more advanced version of Zhenwu, which is set to slightly outperform Nvidia’s A100, the data center graphics processing unit (GPU) for AI rolled out five year ago, Per the report. That was seen as a sign of how far behind the company’s chips are.
Baidu was said to be experiment with training new versions of its Ernie AI model using its own Kunlun P800 chip. The sources didn’t reveal how the chip performs comparing with Nvidia’s H20. it said the Kunlun chip, designed specially for large language models (LLMs), can handle both inference and training tasks. Baidu was said to begin using the chip in for the post-training of older models after it open-sourced these models in June.
A spokesperson for Nivida acknolwedged compeition from Chinese companies and vowed support for develpers with its products. “The competition has undeniably arrived. Customers will choose the best technology stack for running the world’s most popular commercial applications and open-source models. We’ll continue to work to earn the trust and support of mainstream developers everywhere,” said the spokesperson.
The report on Thursday is one of the latest sign that Chinese companies are working on alternatives for Nividia’s H20. News about Alibaba’s new AI chip went viral nearly two weeks ago.
Alibaba, a long-time big customer of U.S. AI chip leader Nvidia Corporation, has developed a new AI chip "that is more versatile than its older chips", and the company and other chip designers could begin "filling the void left after Nvidia ran into regulatory barriers to selling its products in China," the Wall Street Journal reported on August 29.
Unlike previous cloud-computing chips mainly designed for specific applications, Alibaba’s new chip, now in testing, appeared as a general-purpose one since it was reported to serve a broader range of AI interference task. The chip is designed for inference, instead of training, as China’s biggest weakness is training AI models, the report noted.
Alibaba’s new chip was also said to be designed to work with the Nvidia software ecosystem so it will be compatible with the Nvidia platform to ease engineers and developers’ adoption. The report suggested the chip is domestically built. That was an essential step for China toward more localized production, though local players relying on Chinese chip factories like Alibaba are now facing the challenge to get enough supply due to manufacturers’ difficulties in increasing capacity.
Alibaba didn’t respond to the report, while CNBC on August 29 echoed, citing anonymous sources that Alibaba is developing an in-house chip for AI applications. The semiconductor is said to be specially designed for inferencing, which is the actual running of AI applications, rather than training which is when huge amounts of data is used to develop a large model, but Alibaba will continue to use chips from other vendors including Nvidia.
