China is learning to build without Nvidia

Hello, this is Evelyn, writing to you from Beijing. Welcome to the latest edition of China Connection — a short summary of what I see and hear from local businesses.

China’s goal of technological independence is rapidly becoming a reality as companies focus on business questions that go deeper than geopolitics. What does that mean for Nvidia?

Robovan startup Zelostech plans to use multiple chip suppliers from China and elsewhere, within the next year or two, instead of relying solely on Nvidia of its self-driving plans, the company told CNBC.

The biggest factor is cost, said Shi Yunjian, director of finance and investment. Using chips made in China, for example, would cost much less than the two Nvidia Orin chipsets used in each car, he said.

That’s a big deal because scale turns into a competitive advantage. The more autonomous vehicles can use, the more performance data they can get and the easier it is to convince regulators that the technology is ready for widespread use.

Zelostech says it already has more than 25,000 vehicles in operation in more than 20 countries, with plans to grow rapidly. These do not carry people, and most are smaller than a mail truck. Most of them work in China, especially in package delivery companies.

By comparison, Alphabets-backed Waymo has just under 4,000 vehicles on the road, while Chinese rivals Baidu, WeRide and Pony.ai have yet to roll out a similar fleet.

Zelostech is not alone in pursuing alternatives to Nvidia.

Waymo uses custom-made chips, and it’s a stick for Chinese electric cars BYD last week he joined Nio again Xpeng in disclosing its semiconductors for driver assistance systems.

This year, Nio said it plans a fivefold increase in computing power costs. When I asked if that included Nvidia, CEO William Li said the company no longer buys chips but rents computing power powered by various processors.

The Xpeng car developed in cooperation with Volkswagen also uses the “Turing chip” of the Chinese company, while the German company cooperates with China. Horizon Robotics developing driver assistance systems in China – without Nvidia.

Nvidia’s driver-assistance chips are not subject to the same US export restrictions that apply to more advanced semiconductors used to train and run AI models.

But even after Nvidia CEO Jensen Huang joined US President Donald Trump on his trip to Beijing in May, it’s clear that China is not willing to let more Nvidia chips in.

The change extends beyond cars. Chinese AI developers have largely optimized their models to run on domestic hardware, instead of Nvidia’s widely used CUDA ecosystem.

The latest MiniMax and Kimi models, as well as DeepSeek’s V4, are compatible with local Chinese semiconductors.

“We believe the pivot to domestic chips will accelerate beyond 2026E-28E,” Goldman Sachs analysts said in a May 5 report. They pointed out that DeepSeek V4 works with eight Chinese-made chips, including those from Huawei and AlibabaT-head chip unit.

Huawei last week also revealed that it was using a new scientific method to make its chips, and plans to include them in future products. It is the latest sign of the Chinese telecom giant’s comeback, after years of US restrictions.

Kevin Xu, founder of hedge fund Interconnected Capital, expects Chinese companies to continue to demand Nvidia chips in the next three to five years.

But he says Beijing has an incentive to reduce that dependence sooner rather than later. Chips made in China can only improve if companies use them in real-world situations, generating the feedback needed to make the technology useful for businesses.

“When Nvidia chips come into the ecosystem, that’s when you have a significant reduction in relationships,” he said.

Nvidia’s revenue from mainland China and Hong Kong is declining, however, as the company doubles down in Taiwan with plans to spend up to $150 billion a year there.

That investment will likely reverse Taiwan’s initial plan to limit AI data centers and nuclear power, said Chris Cottone, president of TriOrient Investments and chairman of AmCham Taiwan’s separate commodities committee. He also expects more local businesses to use AI.

Nvidia’s growing presence in Asia – but not in China – is driving the chip maker to find other ways to maintain its technological leadership.

An American company is trying to keep its foot in China’s world of “physical AI” by partnering with Chinese humanoid startup Unitree on a research robot that is being sold around the world. Huang is also under scrutiny as he reportedly joined the board of Tsinghua University, a company and school that did not respond to requests for comment.

It is a sign that the tide is turning. China’s technological ambitions are no longer defined by access to Nvidia, but by companies that can build without it.

China is ‘losing an opportunity’ by not being at the Shangri-La Dialogue: German defense official

German defense chief General Carsten Breuer’s comments come after Chinese defense minister Dong Jun skipped the summit for the second year in a row, and Beijing sent a low-level delegation.

China’s industrial profits jump 24.7% in April, fastest gain in two years despite headwinds

In the first four months of the year, industrial profits increased by 18.2%, up from 15.5% growth in the first quarter. The production of computer and electronic equipment, the largest sector in terms of profit, saw revenues more than double the previous year.

European companies are doubling down on manufacturing in China despite EU threats

Many European companies are keeping or expanding their stores in China in order to remain competitive in the world, according to a study released on Wednesday by the European Union Chamber of Commerce in China.

May 31 – June 2: Brazilian Foreign Minister Mauro Vieira will visit China

June 1 – 3: UK Foreign Secretary Yvette Cooper will visit China for the 11th bilateral strategic dialogue

June 3: RatingDog China Services PMI (May)

June 9: Trade data for May