AI Trends

NVIDIA Interim CEO Cress Confirms Halt of H200 AI Chip Sales to China

Electronic Technology ApplicationUpdated Nov 18 14:11

AI summary

NVIDIA Interim CEO Cress confirmed on the November 17 earnings call that the company will stop selling H200 AI chips to China, formally closing the special supply channel to China that Jensen Huang had previously negotiated. On the same day, NVIDIA announced the establishment of a board-level special committee to oversee national security policy, export control compliance, and global trade risks. The U.S. Department of Commerce issued new restrictive rules on November 16 prohibiting the deployment of advanced chips in overseas data centers to provide computing power to restricted customers; investment bank Jefferies estimates that the rules could reduce NVIDIA's data center business revenue next year by $44 billion to $62 billion.

Why read this

SelectedAI score 82

The report outlines the timeline of the closure of the H200 supply channel to China and the new U.S. export restrictions, and provides investment banks' estimated ranges for the revenue impact.

Topics

Tags

Body · AI translation

On November 18, it was reported that Nvidia interim CEO Cress confirmed at the November 17 earnings call that the company will stop selling H200 AI chips to China, formally closing the special supply channel to China previously negotiated by Jensen Huang.

This earnings call was also Nvidia's first financial results communication meeting since Jensen Huang's resignation, and the supply of advanced chips to China became a focal issue in the market.

On the same day, Nvidia officially announced the establishment of a dedicated board-level committee responsible for overseeing national security policy, export control compliance, and global trade risks. The company's national security and trade compliance team will report simultaneously to this committee and the interim CEO on a dual-track basis, strengthening compliance management capabilities.

The H200 sales plan for China was originally finalized through negotiations between Jensen Huang and the Trump administration. This set of rules required that the chips must be transshipped through the United States, with the U.S. government collecting a 25% tariff.

On the timing front, the U.S. Department of Commerce issued new restrictive rules on November 16, prohibiting advanced chips from being deployed in overseas data centers to provide computing power to restricted customers through remote services.

Investment bank Jefferies estimates that this set of new rules could reduce Nvidia's data center business revenue next year by $44 billion to $62 billion. Cress responded that backlog orders for new products such as Vera Rubin are sufficient and will not impact overall revenue in the short term.

In the long term, there is also a possibility of losing computing power customers in multiple Southeast Asian countries. Before Nvidia stopped disclosing separate data for its Singapore subsidiary, this Singapore subsidiary's 2025 revenue reached $104 billion.

It should be noted that the overseas remote computing power channel has also been blocked by the new rules. Going forward, overseas data centers purchasing Nvidia chips must demonstrate that the service recipients do not include restricted customers.

Source

Electronic Technology Application