Bytedance reportedly accesses NVIDIA's latest AI chips outside China

Bytedance, the Chinese parent company of TikTok, is reportedly partnering with a Singapore-based firm to acquire NVIDIA's advanced B200 AI chips for use in Malaysia, bypassing US export restrictions. This arrangement aims to support AI research and development outside China. The project involves significant investment and adheres to export regulations.

Bytedance has found a workaround to US export controls on advanced AI hardware by collaborating with Aolani Cloud, a Singapore-based company, to build computing systems in Malaysia. According to a report from The Wall Street Journal, this setup will provide access to approximately 36,000 B200 chips, NVIDIA's most powerful processor to date. The initiative is expected to cost over $2.5 billion, with the chips intended for AI research and development outside of China.

The B200 chip, designed in California, falls under US export restrictions that prevent its direct sale to China. Aolani Cloud will purchase the components from NVIDIA and operate the systems exclusively in Malaysia, allowing Bytedance indirect access. An NVIDIA spokesperson stated, "By design, the export rules allow clouds to be built and operated outside controlled countries." The company emphasized that all cloud partners undergo review before receiving products.

A representative from Aolani Cloud told Reuters that the firm complies with all applicable export control regulations. Bytedance will be one of multiple customers, with Aolani planning to offer cloud-computing services to companies across Asia and globally. Currently, Aolani operates with about $100 million in hardware, and Bytedance's planned $2.5 billion investment highlights the scale of this expansion.

Separately, the US recently permitted Bytedance to purchase NVIDIA's H200 chips, but these face a 25 percent tariff and a Know-Your-Customer requirement to prevent access by China's military. NVIDIA has not yet agreed to these terms. This development follows broader trends where Chinese firms seek alternatives to restricted technologies.

Related Articles

Business executives from Nvidia and SK hynix shaking hands in a high-tech boardroom with AI factory elements in the background.
Image generated by AI

Nvidia and SK Hynix sign multiyear deal for AI factories

Reported by AI Image generated by AI

Nvidia and SK Hynix announced on June 8 that they have entered a multi-year technology partnership to develop next-generation memory technologies for AI factories.

Start-up Tensordyne has secured letters of intent worth more than 200 million dollars for its semiconductors. Fifteen data center operators are interested in the chips, which are said to consume far less energy than Nvidia products.

Reported by AI

China's Zhengzhou core node has doubled its chips to 60,000 from 30,000 since early February trials, becoming the nation's most powerful scientific intelligent computing infrastructure, CCTV reported.

Google has introduced two new Tensor Processing Units, the TPU 8t for training and TPU 8i for inference, targeting what the company calls the agentic era of AI. These eighth-generation chips follow the Ironwood TPU from 2025 and promise faster, more efficient AI development. The hardware aims to cut training times for large models from months to weeks.

This website uses cookies

We use cookies for analytics to improve our site. Read our privacy policy for more information.
Decline