Nvidia to Release Lower-Cost Blackwell AI Chip for China After US Export Restrictions

Nvidia to Launch Affordable Blackwell AI Chip for China Amid US Export Restrictions

Nvidia is set to introduce a new AI chipset for the Chinese market at a significantly reduced price compared to its previously restricted H20 model. Sources indicate that this new GPU, based on Nvidia's Blackwell architecture, will be priced between $6,500 and $8,000, substantially lower than the $10,000-$12,000 range of the H20.

Key Features and Manufacturing Adjustments

The price drop reflects simpler manufacturing and less advanced specifications. The new chip is based on Nvidia's RTX Pro 6000D, a server-grade graphics processor. Unlike the H20, it will utilize conventional GDDR7 memory instead of high bandwidth memory (HBM), and won’t employ Taiwan Semiconductor Manufacturing Co’s (TSMC) advanced Chip-on-Wafer-on-Substrate (CoWoS) packaging.

This approach allows Nvidia to comply with US export controls while maintaining a foothold in China's lucrative $50 billion data center market. However, Nvidia has stated that final product design and US government approval are still pending.

Market Impact and Challenges

• China accounts for 13% of Nvidia’s sales, making it a critical market despite increasing restrictions.
• Before 2022, Nvidia held a 95% GPU market share in China, which has now fallen to around 50%.
• Competition from Huawei’s Ascend 910B chip is mounting, especially as US export curbs limit Nvidia’s offerings.

The latest US restrictions cap GPU memory bandwidth at approximately 1.7 to 1.8 terabytes per second, down from the H20's 4 terabytes per second capability. The new chip is expected to meet this limit using GDDR7 memory.

Future Developments

Beyond the initial June production, Nvidia is reportedly developing another Blackwell-architecture chip for China, expected to enter production around September. Details on this variant remain undisclosed.

This move marks Nvidia’s third attempt to produce GPUs specifically for the Chinese market in response to tightening US export controls, reflecting ongoing challenges in balancing compliance and market demand.

Considerations for Product Development Professionals

For those involved in product development, Nvidia’s strategy highlights the importance of adapting hardware design to regulatory environments without compromising core capabilities. Simplifying memory technology and packaging can reduce costs and manufacturing complexity but may also impact performance.

This case illustrates how export restrictions can drive innovation in product configuration and supply chain management, emphasizing the need for flexibility in product roadmaps.

Stay informed on AI hardware trends and compliance strategies by exploring relevant courses on Complete AI Training.