Nvidia CEO Jensen Huang has publicly acknowledged that the company has "largely conceded" China's AI chip market to Huawei. This isn't a temporary setback or a quarterly blip. It's a strategic retreat from the world's second-largest economy, driven by US export controls and homegrown Chinese competition.

Let's be clear about what this means: Nvidia dominated AI chips globally. Their GPUs—particularly the H100 and A100 series—are the gold standard for training large AI models. Every major AI lab in the US runs on Nvidia. Google, Meta, OpenAI, Anthropic—they all use Nvidia chips. That market dominance made Nvidia one of the most valuable companies in the world.

But in China, that dominance is over. According to CNBC, Huang made the admission during an earnings call, essentially saying the company has accepted that the Chinese market is now Huawei's to lose.

The immediate cause is US export controls. Starting in 2022, the US government began restricting the export of advanced chips to China, specifically targeting AI accelerators. The goal was to slow down Chinese AI development by cutting off access to cutting-edge hardware. Nvidia tried to work around this by creating "nerfed" versions of their chips—slower models that technically complied with the restrictions but still offered decent performance.

China responded by doing what countries do when faced with export controls: they built their own. Huawei, which was already developing chips after being cut off from US technology, accelerated its AI chip programs. The result is the Ascend series, which Huawei claims can compete with Nvidia's offerings—at least for the Chinese market.

Now, are Huawei's chips as good as Nvidia's bleeding-edge hardware? Probably not. The US still has a significant technological lead in semiconductor manufacturing and design. But here's the thing: they don't have to be as good. They just have to be for China's needs. And they are.