AI boom grows Nvidia’s secretive business

While Nvidia stands out on a global scale with its artificial intelligence chips, another business line that attracts less public attention continues to grow quietly. The company’s division focusing on data center network technologies has reached a remarkable level with its revenue and growth rate in recent years. However, this area is not as visible as chip production and lags behind the recognition on the gaming hardware side. Still, financial data reveals that this segment plays a critical role in Nvidia’s future. In particular, the increasing demand for artificial intelligence infrastructure has directly increased the need for network solutions.

According to Nvidia’s latest financial results, the company’s networking technology division generated $11 billion in revenue in the last quarter alone. This figure means an increase of 267 percent on an annual basis. In total, more than 31 billion dollars of revenue was generated for the whole year. This performance makes the unit the second largest source of revenue after Nvidia’s chip business. Although this growth is remarkable, it is noteworthy that it is not on the agenda as much as the company’s other business lines.

How is Nvidia growing so fast on the data center networking solutions side?

Behind this rapid rise is the increase in artificial intelligence-focused data centers. While NVLink technology developed by Nvidia provides high-speed communication between GPUs in data centers, solutions such as InfiniBand switches and Spectrum-X Ethernet platform optimize data flow. In addition, new generation solutions such as optical switch technologies also increase system performance. When all these components are evaluated together, a modern data center structure defined as an “artificial intelligence factory” emerges. This structure creates the necessary infrastructure for training large language models and other artificial intelligence systems.

The foundations of this business line are based on the Mellanox acquisition that took place in 2020. The acquisition of this Israel-based company for $7 billion enabled Nvidia to make a rapid leap forward in network technologies. Although the strategic value of this acquisition was not fully understood at that time, the results obtained today clearly demonstrate the impact of this move. In addition, offering GPU and network technologies together allows Nvidia to offer more integrated solutions to customers.

The company’s approach is not limited to just selling hardware. Nvidia aims to simplify data center installations by offering fully integrated systems. However, it prefers to deliver its products to the market through business partners rather than offering them directly to the end user. This model provides both scalability and access to a wider customer base in different sectors. On the other hand, this approach stands out as one of the important elements that distinguishes Nvidia from its competitors.

New technologies introduced at the Nvidia GTC conference also indicate that this growth will continue. The new chips introduced within the scope of the Rubin platform were developed to support artificial intelligence supercomputers. In addition, the Inference Context Memory Storage solution and more efficient Spectrum-X Ethernet Photonics switches were also announced. These developments show that data center networks are no longer just an auxiliary component, but directly become a core element of the computing infrastructure.