Industry stub

JPR: NVidia Grows GPU Shipments 21% in 3Q15, AMD 16%

Posted on November 11, 2015

Jon Peddie Research today released its independent GPU shipment analysis, its highlights including an averaged 9% growth in GPU shipping volume, despite a year-over-year desktop GPU shipment decline of 13%.

JPR has released these reports before. The research firm includes Intel's IGPs and AMD's APUs whenever mentioning “GPU shipments” or “GPU sales,” unless otherwise noted in their documentation. This fact is important to understand why some of the metrics appear the way they do.

The firm listed the following as its “quick highlights” of the research endeavor:

 

  • AMD's overall unit shipments increased 15.87% quarter-to-quarter, Intel's total shipments increased 5.08% from last quarter, and Nvidia's increased 21.39%.
  • The attach rate of GPUs (includes integrated and discrete GPUs) to PCs for the quarter was 138% which was up 1.24% from last quarter and 29.95% of PCs had discrete GPUs, which is up 3.51%.
  • The overall PC market increased 7.55% quarter-to-quarter, and decreased -8.94% year-to-year.
  • Desktop graphics add-in boards (AIBs) that use discrete GPUs increased 27.59% from last quarter.

According to JPR, third-quarter volume normally sees improvement over the traditionally-down second-quarter. Holiday sales are forthcoming, of course, so we should expect to see a shift in GPU sales as Black Friday approaches. The report calls-out PC gaming as a “bright spot in the market,” shining a beam on gaming's favor toward mid-range and high-end GPUs.

APU sales increased 15% in desktops, reportedly, and AMD discrete GPUs increased a reported 33.33% over Q2. Overall graphics shipments, including a 17.6% quarterly growth in notebook sales, build a shipment volume increase of 15.9% for AMD.

NVidia's dGPU shipments grew by 26.35% quarterly, according to JPR, and saw a notebook volume growth of 14.1%. NVidia's combined shipment volume fronts a growth of 21.4% over Q2.

Note that JPR sometimes reports data which conflicts with nVidia's own metrics. This is partly a result of the way JPR defines a “GPU” and aggregates its data.

- Steve "Lelldorianx" Burke.