Jensen Huang表示，他预期中国当局最终会允许进口来自美国的人工智慧晶片，并认为中国市场会随著时间推移而打开。这番话是在他随同Donald Trump赴中国参加峰会后几天发表的；他说自己没有直接与中方官员讨论向中国客户出售H200晶片，但承认两国官员的讨论中确实提到了这一议题。

Nvidia在中国的曝险自2022年美国实施管制后已大幅缩小，但Trump在12月同意放行H200对华出货，显示对限制中国AI发展的措施有所松动。尽管Huang曾把中国视为Nvidia的一个500亿美元机会，Nvidia今年早些时候仍维持中国AI晶片销售为零的预测；公司将于周三公布业绩，投资者将关注对中国出货前景的更新。

Huang也表示，尽管美国努力扩大本土晶片制造能力，Taiwan仍会是半导体制造中心，因为需求遍及各方且非常强劲。他与Michael Dell一致认为，记忆体晶片仍是最大瓶颈；AI需求因agent-driven software的采用而激增，供应商面临压力，而Huang估计，追赶这波建置需求至少还要10年。

Jensen Huang said he expects Chinese authorities to eventually allow imports of artificial intelligence chips from the US, and he believes the Chinese market will open over time. He made the remarks days after accompanying Donald Trump to China for a summit; he said he did not directly discuss selling H200 chips to Chinese officials, though he acknowledged the issue did come up in talks between officials on both sides.

Nvidia’s exposure to China has shrunk sharply since US controls were imposed in 2022, but Trump’s December decision to permit H200 shipments to China signaled some easing of restrictions on China’s AI development. Although Huang has described China as a $50 billion opportunity for Nvidia, the company earlier this year still kept its forecast for zero AI-chip sales in China; investors will look for an update on Wednesday when the company reports earnings.

Huang also said Taiwan will remain a semiconductor manufacturing center even as the US tries to expand domestic chip production, because demand is strong across the board. He and Michael Dell agreed that memory chips remain the biggest bottleneck; AI demand is surging on the adoption of agent-driven software, suppliers are under pressure, and Huang estimates it will take at least 10 years to catch up with the buildout.