SemiAnalysis reported that Nvidia's Kyber NVL144 next-gen AI rack system could be delayed until 2028 due to troubles with manufacturing a crucial circuit board. This delay could help give AMD and Alphabet a foothold into the high end of the AI infrastructure market. Nvidia has stated its roadmap remains intact, although questions linger. AMD's improved ROCm platform and use of open-source AI frameworks like OpenAI's Triton have helped close the software ecosystem gap with Nvidia, while its chiplet designs and acquisition of Mext position it for high-end server solutions. Alphabet's Tensor Processing Units (TPUs) are highly regarded, and its next generation will have chips optimized for both training and inference; a Kyber delay could make its cost-efficient TPU offering more attractive. Neither AMD nor Alphabet needs to displace Nvidia to be winners, as capturing a piece of the high-end market would be a big boost.