Nvidia Excludes China From Revenue Forecast Due To $10.5 Billion Sales Loss

Nvidia CEO Jensen Huang has revealed that the company will exclude the Chinese market from its revenue forecast due to imposed restrictions on chip sales to China. The export restrictions are expected to cost Nvidia $2.5 billion in sales in Q1 2025 and an expected $8 billion in Q2 2025. Huang stated that Nvidia is not counting on the U.S. to ease the restrictions, and any easing would be a bonus to the company. The Trump administration has recently tightened restrictions on selling Nvidia H20 AI chips to China, fearing that Beijing will use them to boost its military and AI capabilities.

Huang has been critical of the April Trump tariff on chip exports, which prevented the sale of their most advanced AI chips to China. He believes that the goals of the export controls are not being achieved and that they are doing more harm to the U.S. than to Chinese businesses. Huang revealed last month at a conference that the U.S. curbs on chip exports were a failure.

Nvidia will no longer include China in its future forecasts. 12.5% of Nvidia's revenue comes from China, but the company is currently evaluating its limited options for the Chinese market. The chip maker added that until it settles on a new product design and receives approval from the Federal government, they are effectively closed from China’s data center market.

According to Michael Ashley Schulman, CIO of Running Point Capital, by zero-basing China, Nvidia removes the volatile variable neither Wall Street nor the Commerce Department can reliably handicap. He added that any Chinese sales would be an upside surprise to the company.

Nvidia has maintained its position as a major AI player globally by announcing a potential project to build the first cloud computing platform in Europe. The platform will focus on Industrial artificial Intelligence applications and use Nvidia’s Blackwell architecture to power the AI infrastructure. The AI factory will be based in Germany and feature 10,000 GPUs, including Nvidia DGX B200 systems and RTX Pro Servers.