Driven by the artificial intelligence boom and the large-scale expansion of global data centers, the global data center Ethernet switch market hit a record high of US$15.4 billion in the first quarter of 2026, a year-on-year increase of 39.8%. Against the background of this rapid growth, NVIDIA has won the top spot in the data center Ethernet switch market revenue for the first time, marking its acceleration in dominating the entire AI infrastructure stack in addition to GPUs and CPUs.
According to the latest data from research firm IDC, NVIDIA's data center Ethernet switch business achieved a year-on-year growth of 192.7% in the first quarter of 2026, with quarterly revenue reaching US$2.1 billion and market share climbing to 21.5%. The core product driving its rapid revenue jump is the Spectrum‑X platform for large-scale GPU clusters. This platform deeply integrates GPUs with network infrastructure through software and hardware co-design, targeting the so-called "AI factory" level data center network needs.
IDC pointed out that NVIDIA’s Spectrum‑X platform has been widely adopted by hyperscale cloud vendors and enterprise customers and plays a key role in the construction of AI factory network infrastructure, thus reshaping the vendor landscape of the data center network industry. Spectrum‑X provides end-to-end Ethernet networking solutions, including components such as BlueField DPU and NVIDIA LinkX cables, focusing on large-scale GPU cluster scenarios for AI training and inference.
From a regional perspective, the Americas had the strongest growth in the data center Ethernet switch market in the first quarter of this year, with a year-on-year growth of 49.7%; Europe, the Middle East and Africa (EMEA) followed closely, with a year-on-year growth of 32.2%; the Asia-Pacific region (APAC) recorded a year-on-year growth of 25.9%. Among them, hyperscale cloud data centers and enterprise data centers alone contributed approximately US$10 billion in revenue, a year-on-year increase of 61%, fully reflecting the concentrated explosion of AI-related infrastructure investment.
In terms of rate structure, high-speed Ethernet products continue to be the main force in the market. In the first quarter of 2026, 800G switches have accounted for 35.8% of data center Ethernet switch revenue, while 200G and 400G switches together account for 34.1%. These three types of high-speed switches together contribute about 70% of global data center Ethernet revenue, showing that the industry is accelerating the evolution to higher bandwidth and lower latency network architecture to meet the stringent throughput and latency requirements of AI training and inference.
Nvidia has long been the absolute leader in the GPU market, especially in the field of AI accelerator cards and high-end consumer graphics cards, and is expanding its territory in the computing field with products such as Vera CPU. Now coupled with its No. 1 ranking in the data center Ethernet switch market, NVIDIA is transforming from a single "accelerator supplier" to a full-stack AI infrastructure provider covering GPUs, CPUs, and high-performance networks, further consolidating its voice in the data center in the AI era.
The industry expects that as large-scale AI model training and the "AI factory" concept continue to be implemented globally, the demand for 400G, 800G and even higher-speed network equipment will continue to rise in the next few years. With the AI-optimized Spectrum‑X platform and its end-to-end network solution built around GPU clusters, NVIDIA is expected to further expand its lead in this new round of infrastructure competition and become a core player in the "full stack" of data centers from computing to interconnection.