After hitting a landmark $4 trillion valuation, Nvidia is sharpening its focus on China’s artificial intelligence market, even as US export restrictions tighten and local rivals gain ground. CEO Jensen Huang is expected to visit China in September, a trip that aligns with the rollout of a new AI chip designed specifically for Chinese customers.

The visit carries both strategic and symbolic weight, a signal to Chinese partners and officials that Nvidia isn’t backing away, despite the growing political friction and regulatory headwinds.

Jensen Huang’s planned visit to China includes high-level meetings with senior officials such as Premier Li Qiang and Vice-Premier He Lifeng, part of a broader effort to reaffirm Nvidia’s long-term presence in the country.

At the same time, Huang hasn’t held back in criticizing US export controls, arguing they’ve done more harm than good, fast-tracking China’s homegrown AI industry while cutting Nvidia’s dominance in the region.

Once commanding 95% of the Chinese AI chip market, the company now holds just 50%, a drop Huang attributes directly to policy missteps in Washington.

China specific AI chip

The upcoming chip is a reworked variant of Nvidia’s Blackwell RTX Pro 6000, redesigned to meet stricter US export rules.

Key features like high-bandwidth memory (HBM) and NVLink have been removed, meaning the chip won’t deliver the same top-tier performance as Nvidia’s global flagship models.

Still, it retains compatibility with Nvidia’s proprietary CUDA software, a critical selling point for many Chinese developers who are reluctant to shift to unfamiliar platforms despite rising local alternatives.

Nvidia is pressing pause on sales until September as it looks for clear signals from US regulators that the new chip won’t face sudden restrictions post-launch.

The company is still reeling from a $5.5 billion writedown after its earlier H20 model was abruptly blacklisted, and it’s determined to avoid a repeat.

The chip’s final specs are still being shaped in ongoing talks with Washington, a sign of just how fluid and fraught the landscape remains when it comes to US-China tech policy.

The new chip is likely to carry a lower price tag than the earlier H20 model, which sold for between $10,000 and $12,000.

Current estimates place the upcoming version in the $6,500 to $8,000 range, a drop that reflects its simpler design and the removal of high-end features.

Jensen Huang and his balancing act

Although Nvidia’s chips now lag behind local rivals like Huawei’s Ascend 910B in some technical areas, major Chinese tech players including Alibaba, ByteDance, and Tencent are still actively testing Nvidia’s hardware.

The reason? Transitioning away from Nvidia’s deeply entrenched software ecosystem is costly and technically challenging.

Despite rising anxiety over-reliance on US technology, especially in an unpredictable policy environment, many Chinese firms remain reluctant to make a full switch.

China is still Nvidia’s fourth-largest market, bringing in $17.1 billion last fiscal year, about 13% of the company’s total revenue.

But staying competitive there isn’t getting any easier. Nvidia now faces a balancing act: stockpiling enough inventory to keep supply chains steady, while bracing for possible regulatory shocks.

At the same time, it’s contending with a new wave of capable domestic competitors that are rapidly closing the gap.

The post Jensen Huang’s China gambit: Nvidia bets on AI growth amid geopolitical crossfire appeared first on Invezz