The outside world is paying attention to NVIDIA's server supply status. Quanta's Vice President: Rubin series will be shipped as early as August.

As NVIDIA’s Vera Rubin announced mass production at CES, the industry is watching the actual shipment speed of the supply chain. Recently, Quanta’s Executive Vice President and Yunda Technology General Manager Yang Qiling revealed: Rubin-related products are expected to be delivered as early as August, and Quanta’s AI server revenue has already accounted for about 80% of the total server revenue this year.

Quanta Vice President: Rubin deliveries in August still based on GB series

As NVIDIA’s new generation Rubin platform gradually comes to light, the market is highly concerned about the actual shipment progress of the related supply chain. In response, Yang Qiling, Executive Vice President of Quanta and General Manager of Yunda Technology, pointed out that Rubin-related products are expected to begin delivery to customers as early as August this year. However, since the platform is still in the early stages of deployment, the short-term impact on revenue structure remains relatively limited.

Yang Qiling further explained that earlier market speculation suggested Rubin might adopt a completely new Hyper architecture. However, based on current information, most Rubin servers are still built on existing GB series (such as GB200, GB300) designs. This continuity in architecture actually helps the supply chain to connect quickly, making subsequent mass production and deployment more stable.

AI server structure shift, shipment focus moving to high integration level

From the overall AI server shipment pattern, Yang Qiling pointed out that the market is clearly shifting from the traditional L10 assembly mode to the more integrated L11, with some large projects even advancing to L12 level, mainly delivered in rack form.

Regarding network switches, Yang Qiling also highlighted the different thinking between Quanta and industry peers. He stated that switches are not just simple OEM projects for Quanta but an indispensable part of the overall AI cluster architecture. Through experience in extending from rack to cluster level integration, Quanta has gained deeper insight into data center networking logic. In this model, each rack’s switch can become a tangible and expandable business opportunity, which in the long run helps enhance the overall system value.

Strong shipment momentum, Quanta simultaneously expands manpower and capacity

Looking ahead to this year’s operations, Quanta expects AI server shipments to continue high-speed growth, with a projected annual increase of triple digits. To meet order demands, the company is continuously increasing manpower investment. Currently, the server-related team has nearly 3,000 employees, with plans to expand by another 500 to 800 this year.

In terms of capacity, Quanta is also actively pushing expansion plans, aiming to at least double overall capacity by the end of the year. Yang Qiling revealed that over the past two to three years, the company has almost weekly reviewed factory expansion progress and execution details. The core principle is simple: as long as customer demand is clear, manpower and capacity must be ensured to keep pace. He also pointed out that currently, AI server revenue accounts for about 80% of total server revenue, and the overall performance this year remains promising.

This article, “Industry focuses on NVIDIA server supply status, Quanta Vice President: Rubin series shipments as early as August,” first appeared on Chain News ABMedia.