Nvidia’s Kyber NVL144 reportedly pushed back more than a year, Asian suppliers drop

What happened

Nvidia’s next-generation AI server rack, the Kyber NVL144, has been delayed until 2028. The postponement exceeds a year and comes amid circuit board manufacturing problems. An improved version, the Rubin Ultra, has been canceled entirely. The delays triggered notable stock drops among Asian suppliers tied to the hardware’s production.

Why it matters

This setback stalls Nvidia’s planned upgrade path for AI data centers, slowing the rollout of more powerful and efficient server racks. The Kyber NVL144 was expected to advance Nvidia’s lead in high-performance AI infrastructure, so the delay hands a potential opportunity to competitors like AMD and Google to gain traction. For cloud operators and enterprises betting on Nvidia’s next-gen hardware, development and deployment timelines will stretch, affecting upgrade cycles and project budgeting.

Suppliers in Asia losing market value signals strained supply-chain confidence and might cause tighter component availability, pushing costs and complicating production schedules for Nvidia and others relying on similar manufacturing. The cancellation of the Rubin Ultra heightens the risk Nvidia’s hardware roadmap faces pressure from manufacturing challenges that could slow innovation delivery.

What to watch next

Track how AMD and Google respond. Both companies have been developing their own AI infrastructure initiatives and may accelerate efforts to capture market share amid Nvidia’s disruption. Watch for Nvidia’s public comments on the delay’s cause and mitigation strategies. Also, monitor supplier earnings and semiconductor market movements for further signs of stress or recovery in the supply chain. For customers, keep an eye on revised Nvidia product launch timelines and budgeting adjustments for AI infrastructure upgrades.

AI Quick Briefs Editorial Desk