Nvidia’s Huang ousted by board in move that roils AI economy - The Japan Times

Jensen Huang's abrupt removal marks a structural fracture: Nvidia shifts from founder-driven architecture to institutional governance, and the consequences cascade through the entire AI stack. Technically, the CUDA roadmap—historically calibrated by Huang's cross-functional intuition—will decelerate. Expect a 6-12 month delivery gap on HBM4 packaging and chiplet interconnect milestones in 2026-2027, creating a direct opening for AMD's CDNA4 and Broadcom's custom ASICs to capture hyperscaler share. On compliance, Huang functioned as an informal technical diplomat bridging US export-control negotiations and TSMC's CoWoS capacity allocation in Taiwan, China. His absence strips Nvidia of its most critical human buffer, pushing operational costs up an estimated 15-20% and exposing supply-chain fragility. Competitively, AMD will weaponize delivery certainty rather than price, locking in wobbled hyperscalers. Google TPU and AWS Trainium gain internal political momentum as in-house alternatives. The 12-24 month tail: the AI silicon market shifts from monopoly to tri-polar fragmentation. CUDA's moat develops its first structural crack, accelerating open-source ecosystem migration. The era of buy Nvidia or die is over.