It is a misunderstanding that Intel’s discrete graphics card sales exceed AMD’s
Previously, it was discovered in Jon Peddie Research’s GPU market data statistics that Intel had surpassed AMD to become the second-largest seller of discrete graphics cards in the fourth quarter of 2022, which seemed like a huge victory. However, some people have questioned this, as there haven’t been many Intel discrete graphics cards seen in the market. Although AMD has performed poorly, it seems incredible that Intel was able to achieve this in just a few months. At that time, Jon Peddie kindly reminded us that the shipment volume was based on Intel’s financial report data and other information, which may not be accurate.
According to a report by TomsHardware, Intel did not sell as many discrete graphics cards in the fourth quarter of 2022 as previously reported. A big reason for this is that data center GPUs like Ponte Vecchio were also included in the calculation, resulting in a significant increase in shipments and sales amounts. Intel’s market share was also revised from 9% to 6%, which is still a good performance for a company that has just entered the market. However, Intel’s average selling price (ASP) for graphics cards is lower than that of NVIDIA and AMD, and a large portion of its shipments are likely concentrated on low- to mid-range mobile discrete graphics cards.
In Intel’s official statistics, GPUs are included in the revenue of its Accelerated Computing and Graphics (AXG) division, which includes both consumer and data center usage. Intel is currently building the Aurora supercomputer for the US Department of Energy’s Argonne National Laboratory, and it is possible that a large portion of Ponte Vecchio was included in the shipments in the fourth quarter of 2022, thus boosting quarterly shipments. It is speculated that Intel did not intentionally mislead the industry, but rather due to the classification of its business units.
However, as Intel splits its Accelerated Computing and Graphics division into two parts, future statistics will be more convenient. Its consumer graphics team will join the Client Computing Group (CCG), while its accelerated computing team will join the Data Center and AI Business Unit (DCAI), and the GPUs used in consumer and data center applications will be divided into two different departmental statistics.