Nvidia Confirms Mass Production of Vera Rubin Platform Without Delays
Photo: Tom's Hardware
Quick answer
Nvidia has confirmed the mass production of its Vera Rubin AI platform, dismissing rumors of delays. The company maintains its roadmap, though timelines for large-scale solutions like NVL144 may shift due to production…
Nvidia CEO Jensen Huang dismissed reports of delays in the release of the new AI platform Vera Rubin. According to him, production has already begun, with supply volumes expected to be "gigantic." This statement was made at an event in Japan and aims to reassure investors anticipating a surge in GPU and Rubin-based system sales over the coming quarters.
Despite confirmation of plans for Vera Rubin, questions remain about the timeline for more complex configurations. The Kyber NVL144 system, which is expected to combine 144 Rubin Ultra GPUs using NVLink 7 copper connections, may face delays until 2028 due to challenges in producing multilayer PCBs. Alternative solutions, such as NVL72x2, have been canceled due to low customer interest.
Nvidia confirmed that its roadmap remains intact but did not specify whether original timelines for all products are still on track. The company is also exploring optical connections for system scaling, though their readiness for mass adoption remains uncertain. Amid growing competition from AMD and Google, Nvidia must maintain its leadership in the high-performance AI solutions segment.
Common questions
- What is Nvidia's Vera Rubin platform?
- Vera Rubin is Nvidia's next-generation AI platform, featuring advanced GPUs and processors for high-performance computing. It is designed for scalable AI and supercomputing solutions.
- Why were rumors about Vera Rubin delays circulating?
- Rumors stemmed from concerns over potential production challenges for complex Rubin Ultra-based systems, including NVL144. However, Nvidia has refuted these claims, confirming mass production has already begun.
- Which competitors could challenge Nvidia in the AI systems segment?
- AMD and Google are developing competing scalable solutions, such as AMD Mega Pod and Google TPU 8i/8t, which may offer systems with hundreds or thousands of accelerators in a single domain.
Dzen feed: /feed/dzen.xml · RSS: /feed.xml