Nvidia's next-generation AI server rack, the Kyber NVL144, has reportedly been pushed back by more than a year — from 2027 into 2028 — and Wall Street noticed.
The delay was first flagged by research firm SemiAnalysis, according to marketscreener.com and other outlets. The Kyber rack is the system designed to house Nvidia's upcoming Rubin Ultra chips. TheNextWeb reports that the holdup comes down to a single hard-to-build circuit board that has proven difficult to manufacture. CNBC-TV18 similarly attributes the roughly year-long slip to manufacturing issues affecting the Rubin Ultra rollout.
The market reaction was swift. Per Stocktwits, Nvidia (NVDA) shares swung in premarket trading on the report that its next-gen AI rack had been delayed by a year. Meanwhile, TechStock² notes that AMD's stock rebounded, weighing on the broader "AI premium" as investors reassessed the competitive landscape.
That competitive angle is the heart of the story. Neowin frames the 12-month setback as an opportunity for rivals AMD and Google to catch up, and GuruFocus reports the delay raises broader concerns for the AI market. When the dominant supplier stumbles, competitors racing to build their own AI hardware get breathing room to win customers who can't afford to wait.
It's worth noting the reporting rests largely on a single research firm's analysis, and Nvidia has not been quoted confirming a new timeline in these items. Details on exactly which board is to blame, or how firmly 2028 is set, remain thin.
Why it matters: Nvidia's chips power much of the current AI boom, so even a rumored one-year delay in its flagship rack can move billions in market value and hand rivals a rare chance to close the gap.