Jensen Huang Courts China: Nvidia Finds a Way Back Into the World’s Biggest AI Market

BEIJING – Jensen Huang, the charismatic CEO and co-founder of Nvidia, has made his third trip to China in just six months. Fresh off celebrating Nvidia’s new $4 trillion market cap, Huang’s latest visit to Beijing comes at a pivotal moment: Nvidia is expected to resume sales of its H20 AI chip to China, following a three-month pause due to U.S. export controls.

Wearing his signature black leather jacket and speaking under the blazing Beijing sun, Huang showed optimism, humor, and strategic intent in his interactions with the local press.

“Only in China can we do this out in the sun!” he joked, before realizing the press conference was actually scheduled indoors.

A Diplomatic Tech Balancing Act

According to Huang, the green light for H20 chip sales may be part of a broader exchange between the U.S. and China, where rare earth minerals from Beijing meet chip concessions from Washington. Though he downplayed his role in the negotiations, Huang admitted he had briefed former U.S. President Donald Trump during a recent meeting.

“It’s not my decision. It’s my job to inform the President about what I know — the technology industry, AI, and global development.”

Huang emphasized Nvidia’s full compliance with international regulations and said the company is prepared to adapt to changing policies and tariffs.

Rebuilding Business Ties in China

During his trip, Huang met with Chinese Vice Premier He Lifeng and Xiaomi CEO Lei Jun, calling the latter “a brilliant business person.” The two discussed AI’s role in large language models (LLMs), autonomous driving, and robotics — sectors where Xiaomi and Nvidia already collaborate via automotive chips.

Despite the restrictions, Huang made clear his ambition:

“I hope to bring even more advanced chips than the H20 to China.”

While Nvidia’s China revenue has taken a massive hit — with a projected $10 billion loss over two quarters due to export limits — Huang remains optimistic that future AI-capable chips will find their way back into the market legally and efficiently.

Nvidia vs Huawei: A Race of Ecosystems

When asked about rival Huawei, Huang didn’t hold back praise:

“Anyone who underestimates Huawei or China’s manufacturing ability is deeply naïve.”

Huawei, which operates independently from Nvidia’s ecosystem, has built its own chip (Ascend), AI framework (CANN), and AI cloud (CloudMatrix). Although some developers find it challenging to switch away from Nvidia's CUDA platform, Huang believes it's just a matter of time before Huawei becomes more competitive.

“They’ve only been at it a few years. The fact they’re already in the game speaks volumes.”

China’s Open-Source AI Models: A New Powerhouse

In a keynote at a major supply chain expo earlier in the day, Huang praised Chinese AI models such as DeepSeek, Alibaba’s Qwen, and Moonshot’s Kimi, calling them “excellent.”

He especially admired their open-source approach:

“99% of companies around the world downloaded DeepSeek R1 for healthcare, robotics, imaging, and more.”

DeepSeek, developed under tight U.S. chip restrictions, shocked the global AI community earlier this year by releasing a model that significantly undercut OpenAI on development and operating costs — reportedly made possible by pre-stockpiling Nvidia chips.

Looking Ahead

As the press conference wrapped up, a reporter asked whether Huang would return to China again in 2025. Huang, smiling, replied:

“I hope so. You have to invite me.”