Nvidia has confirmed that there are no regulatory objections to Chinese tech giant ByteDance accessing a massive cluster of 36,000 Blackwell B200 GPUs located in Malaysia. This development, reported by the Wall Street Journal, highlights a strategic workaround where ByteDance can utilize advanced hardware for artificial intelligence research and development without directly violating current U.S. export controls. The GPU cluster, valued at approximately $2.5 billion, consists of 500 NVL72 GB200 rack-scale systems. While ByteDance seeks to expand its global AI capabilities, the hardware will be formally owned and operated by Aolani Cloud, a Malaysian cloud service provider. According to reports citing industry insiders, the servers were supplied by Aivres, a company specializing in Nvidia-based server construction. Although Aolani currently operates roughly $100 million worth of hardware, the proposed expansion represents a significant leap in scale. Initial payments for this arrangement have reportedly already been made. U.S. export regulations generally restrict the direct shipment of high-performance Nvidia chips to China. However, the rules primarily govern the physical location of the hardware rather than the location of the company accessing it via remote cloud services. As long as the infrastructure is built in compliance with U.S. export controls and located outside of restricted territories, the usage of the compute power by Chinese entities is permitted. Nvidia emphasizes that their compliance teams rigorously evaluate all cloud partners and original equipment manufacturers before shipping products. ByteDance is not on the U.S. Bureau of Industry and Security Entity List or the Military End Use list, meaning its access to this hardware does not automatically trigger regulatory red flags. The company has previously leased Hopper-based H100 servers in Malaysia from Aolani starting in February 2025. This existing relationship suggests the current Blackwell deployment may follow a similar model, with the Malaysian infrastructure serving as a bridge for ByteDance to access top-tier computing power. In addition to the Malaysia arrangement, ByteDance is reportedly exploring further deployments, including a potential cluster of over 7,000 B200 GPUs in Indonesia. Nvidia's stance remains that allowing foreign cloud operators to host this infrastructure generates substantial revenue and high-paying jobs in the United States. A company spokesperson noted that while export controls limited the Chinese market to foreign competitors, the policy intentionally enables the construction of cloud infrastructure globally using American hardware to maintain U.S. economic leadership. Despite some unease among U.S. lawmakers regarding the use of American AI accelerators by Chinese companies, the current framework legally supports this model. The arrangement underscores a complex global landscape where technology companies navigate regulatory boundaries to secure critical infrastructure for AI development, ensuring that American hardware remains integral to the worldwide AI ecosystem even as access is restricted within specific national borders.