TikTok’s parent company ByteDance Ltd. reportedly has a plan to get around tough U.S. restrictions on the export of advanced computer chips to China.

The company is said to be planning to spend a whopping $7 billion on Nvidia Corp.’s most powerful graphics processing units to fuel the development of cutting edge artificial intelligence models. It doesn’t seek to buy any chips, but rather just rent access to them via data centers located outside of mainland China.

ByteDance’s plans were reported by The Information, which cites a number of anonymous sources familiar with the matter. The sources said ByteDance founder Zhang Yiming is personally negotiating with data center operators across Southeast Asia and the Middle East, trying to secure access to Nvidia’s next-generation Blackwell GPUs, which are expected to become widely available later this year.

In conversations with those chip suppliers, Zhang has reportedly indicated that his company’s AI investments will dwarf the combined spending of all of its rivals, including the likes of Alibaba Cloud, Tencent Holdings Ltd., Baidu Inc. and Huawei Technologies Co. Ltd.

ByteDance needs a workaround because Chinese companies are prohibited from buying advanced processors from western companies due to national security fears. The U.S. is convinced that China will use the chips to develop more sophisticated weapons systems and so it has taken numerous steps to stop Chinese firms from getting their hands on them.

That has forced Chinese technology giants to resort to renting access to chips instead. ByteDance is already believed to be using data centers located outside of China to utilize Nvidia’s previous-generation Hopper AI GPUs, which are not allowed to be exported to its home nation.

However, the U.S. government may yet scupper ByteDance’s plans. Last year, the Wall Street Journal reported that U.S. officials are working on yet more rules that would restrict the ability of Chinese companies to rent chips from cloud providers in the U.S. In addition, the U.S. is developing regulations that would position companies such as Google LLC and Microsoft Corp. as “gatekeepers” to cutting-edge AI chips.

Under the proposed rules, those companies would need to report key information on their customers to the U.S. government and ensure Chinese firms are not allowed to access them. In exchange, they would be allowed to offer AI capabilities via global data centers without any licenses. Other cloud providers would have to compete for licenses to obtain a limited number of high-end chips in each country.

While some Chinese firms are engaged in a game of cat and mouse with the U.S. government, looking to bypass its chip export restrictions, others have resorted to doing more with less. For instance, the Chinese AI startup DeepSeek recently announced a new, open-source large language model that it says can compete with OpenAI’s GPT-4o, despite only being trained with Nvidia’s downgraded H800 chips, which are allowed to be sold in China.

DeepSeek reportedly trained its DeepSeek-R1 model at a total cost of just $5.6 million, far less than its rivals have spent to develop comparable models.

Image: SiliconANGLE/Meta AI