Nvidia CEO Jensen Huang resigns as US lawmakers push tougher chip export controls - VOA - Voice of America English News

Huang's departure is not a career pivot—it is an institutional severance forced by Washington's reclassification of silicon from commodity to strategic asset. Technical cascade: The CUDA lock-in risk is underpriced. A decade of AI training infrastructure is architecturally welded to Nvidia's roadmap. CEO transition layered on escalating export restrictions compresses the transparency window for Hopper/Blackwell successors. TSMC's advanced nodes in Taiwan, China and SK Hynix's HBM lines face compliance-review-first friction adding six to nine months to qualification cycles. Compliance exposure: The October 2022 controls already forced A800/H800 derated SKUs. The next congressional push will likely compress the compute threshold further, zeroing out mainland China data-center revenue (peak above 25% of segment) while entity-list screening and EUV licensing overhead erode 15 to 20% of gross margin. Competitive repositioning: AMD's MI300 will harvest compliance-friendly positioning in US public-sector procurement. Huawei's Ascend 910C has closed the 7nm N+2 full-stack gap under supply-cut duress and targets sub-14nm within twelve months. Intel 18A, if on schedule, becomes the neither-Nvidia-nor-China third pole. Eighteen-month trajectory: The global AI silicon market bifurcates from one ecosystem into two parallel stacks. CUDA's moat shifts from technical superiority to regulatory permission.