NVIDIA and AMD Launch New AI Chips in China Amid US Export Rules

The global semiconductor industry is facing a pivotal moment as leading chipmakers NVIDIA and AMD prepare to introduce new AI-focused GPUs tailored for the Chinese market. This strategic move is a direct response to the U.S. government's stringent export restrictions on advanced semiconductor technology aimed at China.

NVIDIA plans to launch a stripped-down AI GPU, codenamed “B20,” while AMD is targeting AI workloads with its new Radeon AI PRO R9700 workstation GPU. Both products are expected to hit the Chinese market starting July 2025, according to Taiwanese tech publication Digitimes, which cited supply chain sources.

This development comes amid significant financial impacts for NVIDIA. The company disclosed a $4.5 billion charge in the first quarter due to licensing requirements that limit its ability to sell the high-end H20 AI chip in China. Additionally, $2.5 billion worth of H20 chips could not be shipped during the quarter because of these restrictions.

NVIDIA forecasts that licensing requirements will result in an $8 billion revenue hit in the second quarter of 2025. To mitigate this, the company is developing a budget AI chip based on its Blackwell architecture, expected to be priced between $6,500 and $8,000, compared to the $10,000 to $12,000 price range of the H20 GPUs.

The introduction of these new GPUs tailored for China highlights how geopolitical factors are reshaping the AI hardware landscape. Companies must innovate not only technologically but also strategically to navigate complex regulatory environments.

Implications for AI Development and Market Dynamics

These new AI chips will enable Chinese companies to continue advancing their AI workloads despite export limitations. However, the performance and capabilities of these stripped-down or budget GPUs may differ from their unrestricted counterparts, potentially influencing AI research and commercial applications in China.

For global AI developers and businesses, this situation underscores the importance of understanding supply chain and regulatory risks. It also highlights the need for flexible AI infrastructure strategies that can adapt to shifting geopolitical landscapes.

In essence, NVIDIA and AMD’s approach to selling modified AI GPUs in China is a case study in balancing innovation with compliance. It’s a reminder that in the fast-evolving AI hardware market, success depends as much on navigating policy as on pushing technological boundaries.