Following US restrictions, Tencent and small chip makers are vying for Nvidia's China crown

According to four people familiar with the discussions, Chinese chip designers such as Tencent Holdings (OTC:TCEHY) are aggressively selling their AI processors as alternatives to Nvidia's (NASDAQ:NVDA), expecting that U.S. export restrictions will prompt clients to switch.

Nvidia, based in California, controls up to 90% of China's $7 billion market for chips used to handle massive amounts of data in order to develop artificial intelligence (AI) software.

However, the intensification of US strategic technology regulations in October has empowered even smaller names, such as state-backed Hygon Information Technology and startup Iluvatar CoreX, to take on the US goliath.

Huawei Technologies is largely regarded as making the most progress, with its Ascend 910B being compared in terms of computing power but not total performance to Nvidia's A100.

However, Tencent and other smaller AI players are speeding up chip product launches and increasing marketing visits, hoping that even if US laws affect just the most advanced chips, clients will still avoid Nvidia, according to the sources.

Tencent, China's largest social media and gaming company that also offers cloud services to third parties, has been pushing services that employ the AI inference processor Zixiao it built with deep learning startup Enflame, claiming performance equivalent to some Nvidia chips, according to two of the people.

According to one of the sources, Tencent is positioning its Zixiao v1 model as a less expensive alternative to Nvidia's A10, which is used for image and speech recognition AI applications. According to the source, it is also promoting a future v2Pro model designed for AI training as capable of replacing Nvidia's now-blocked L40S.

Tencent uses Zixiao chips internally and does not sell them to third-party customers. It rents computing capacity to clients through its cloud services, which provide the option of Nvidia or AMD (NASDAQ:AMD) CPUs.

Tencent stated that it had no intentions to expand Zixiao beyond the current version.

"Tencent created Zixiao to supplement our cloud products and solutions while adhering to laws and regulations." Clients can only access it through Tencent Cloud's enterprise services, according to a company spokeswoman.

Others promote direct sales. According to two of the people, Tencent-backed Enflame, which has an AI training accelerator chip called Yunsui, and Iluvatar CoreX, which develops the Tiangai graphics processing unit (GPU), have been marketing impending enhancements of their wares as substitutes for Nvidia's sophisticated A100 chip.

According to a third source, Hygon is presenting its freshly released GPU, Shensuan No. 2, as being developed from the start to be compatible with Nvidia's chips computing platform CUDA, which means Nvidia users can switch chips with minimal design changes.

Last month, Intellifusion launched the Deepedge10 chip, which will compete with Nvidia's upcoming H20 processor, which is designed to comply with the latest export restrictions.

According to a fourth source, Intellifusion rushed its release to capitalise on Nvidia's dilemma, licencing the chip to clients even though it has yet to be mass produced.

Enflame, Iluvatar CoreX, Hygon Information, and Intellifusion did not respond to calls for comment right away. Nvidia did not respond to requests for comment.

Technology companies in the world's second-largest economy have begun to look for Nvidia alternatives. Tencent has stated that US tariffs have prompted them to seek domestic supply of AI training chips. According to sources, internet search pioneer Baidu (NASDAQ:BIDU) placed a large order for Huawei chips last month.

"Our Chinese competitors include a slew of startups... AI startups are in the hundreds. Huawei is a formidable rival. Other US competitors, Intel (NASDAQ:INTC), and AMD, are all fierce competitors. "There is no shortage of competition," said Nvidia CEO Jensen Huang last week in Singapore.

If Chinese chip designers win orders, they may still struggle for production capacity due to restrictions imposed by the United States on foundries such as TSMC from engaging with Chinese manufacturers, according to Isaiah Research Vice President Lucy Chen.

"The majority of China's advanced manufacturing processes and advanced packaging capacity will almost certainly be prioritised for use by Huawei." "These emerging companies must plan how to overcome the constraints imposed by US restrictions and production limitations," she said.

Nonetheless, the restrictions have created a market opportunity as tech firms pursue a strategy of stocking various types of AI chips than simply Nvidia's, with the sustainability of their AI strategy being a priority rather than performance, according to Nori Chiou, investment director at White Oak Capital.

"The (United States') original goal was to slow down China's AI capabilities but, in fact, related action has boosted China's self-development capability," he added.

"Many Chinese cloud giants are working on building their AI ecosystems without U.S. chips due to these restrictions."