Nvidia's Kyber rack-scale architecture, designed to support the 2027 Rubin Ultra chips, will be delayed to 2028 due to manufacturing difficulties with a key circuit board. This delay raises concerns about Nvidia's product roadmap and may provide competitors an opportunity in the AI space.
Nvidia's next-generation Kyber rack-scale architecture, intended to house Rubin Ultra chips, has been postponed to 2028. This decision arises from complications in manufacturing a critical printed circuit board (PCB), which connects various electronic modules within the system, leading to production challenges.
The delay compounds existing concerns regarding Nvidia's aggressive annual release schedule. With manufacturing limits becoming a significant issue, the company faces potential setbacks in its capability to deliver new products as scheduled.
Nvidia had considered an interim solution of combining two current-generation racks, but this plan was scrapped after pushback from cloud service providers. The design was deemed awkward and costly, further complicating Nvidia's expansion efforts.
As Nvidia struggles with these production issues, competitors like Advanced Micro Devices and Google could gain an edge in the high-end AI market. This situation presents a potential opening for rivals as they seek to capitalize on Nvidia's vulnerabilities.
Despite these delays, Nvidia's existing Rubin systems remain in full production, with shipments expected this fall to several cloud partners including Amazon Web Services, Microsoft Azure, and Google Cloud. Analysts forecast a 20% growth in Nvidia's data-center compute revenue, nonetheless.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
Nvidia's Kyber rack-scale architecture, designed to support the 2027 Rubin Ultra chips, will be delayed to 2028 due to manufacturing difficulties with a key circuit board. This delay raises concerns about Nvidia's product roadmap and may provide competitors an opportunity in the AI space.