JPR Figures Show Nvidia Losing Market Share
Jon Peddie Research has released its Q1 2011 estimates, and the figures don't favor Nvidia. The research firm tracked strong gains for Intel and AMD (in terms of volume). AMD's total share of the GPU market rose to 24.8 percent, up from 21.5 percent a year ago (a gain of 24.8 percent total). Intel came in at 54.4 percent, 9.7 percent higher than its Q1 2010 total of 49.6 percent. Both companies' gains were at the expense of Team Green—Nvidia's share of the market fell from 28 percent to 20 percent—a significant decline.
This loss of share partly reflects NV's decision to exit the integrated chipset market, but it's also an ominous sign of things to come. Sandy Bridge and Llano, once fully launched, will threaten NV's price points at the low end of the market. Intel's Ivy Bridge could mount a fresh challenge against NV's low-end bastions by the end of the year, while AMD's Llano is rumored to be significantly faster than Sandy Bridge's GPU.
For now, these threats are little more than shadows; Nvidia reported excellent quarterly results back in February. Measured according to product mixture or overall performance, Team Green is still playing a strong hand. Long term, however, we expect integrated GPUs from both CPU manufacturers will sap the company's earnings from the bottom up.
This isn't just a threate to NV's income. At the moment, the company has a spot at the table in any sort of API or graphics standard development conference; its products power a huge section of the entire market. Marginalization would substantially hamper NV's ability to make its voice heard—AMD and Intel would become the de facto heavy weights, with the vast majority of graphics split between them. Poor positioning is part of what killed 3dfx; it's not a risk NV will want to take.
Nvidia may fight back by offering faster GPUs at cheaper prices, but the margins at the low-end of the market are quite thin. Long-term, this approach may likely prove unsustainable. Indeed, AMD will likely shed its own low-end SKUs below a certain price point, though likely not for at least another product generation. Even if it doesn't, AMD can offer OEMs the prospect of linking Llano's GPU
to a discrete product as it first unveiled with the 780G chipset. Nvidia has no such option.
The good news is that company executives clearly saw the problem coming. Nvidia has spent years working with the HPC community and explicitly designed certain features of its latest GPUs to meet the needs of such customers. Tegra, like Tesla, is a specialized product meant to appeal to a very different group of users than the core gamers and OEMs that have constituted NV's sales base to date. Nvidia's Quadro business shouldn't be discounted, either. NV's workstation solutions are a substantial source of revenue that won't be easily cannibalized by the availability of high-performance integrated graphics products.
This loss of share partly reflects NV's decision to exit the integrated chipset market, but it's also an ominous sign of things to come. Sandy Bridge and Llano, once fully launched, will threaten NV's price points at the low end of the market. Intel's Ivy Bridge could mount a fresh challenge against NV's low-end bastions by the end of the year, while AMD's Llano is rumored to be significantly faster than Sandy Bridge's GPU.
Note: JPR tracks sales by volume, not by revenue. AMD reported $413 million in GPU sales in its last quarter, compared to $886 million from NV.
For now, these threats are little more than shadows; Nvidia reported excellent quarterly results back in February. Measured according to product mixture or overall performance, Team Green is still playing a strong hand. Long term, however, we expect integrated GPUs from both CPU manufacturers will sap the company's earnings from the bottom up.
This isn't just a threate to NV's income. At the moment, the company has a spot at the table in any sort of API or graphics standard development conference; its products power a huge section of the entire market. Marginalization would substantially hamper NV's ability to make its voice heard—AMD and Intel would become the de facto heavy weights, with the vast majority of graphics split between them. Poor positioning is part of what killed 3dfx; it's not a risk NV will want to take.
Nvidia may fight back by offering faster GPUs at cheaper prices, but the margins at the low-end of the market are quite thin. Long-term, this approach may likely prove unsustainable. Indeed, AMD will likely shed its own low-end SKUs below a certain price point, though likely not for at least another product generation. Even if it doesn't, AMD can offer OEMs the prospect of linking Llano's GPU
to a discrete product as it first unveiled with the 780G chipset. Nvidia has no such option.
The good news is that company executives clearly saw the problem coming. Nvidia has spent years working with the HPC community and explicitly designed certain features of its latest GPUs to meet the needs of such customers. Tegra, like Tesla, is a specialized product meant to appeal to a very different group of users than the core gamers and OEMs that have constituted NV's sales base to date. Nvidia's Quadro business shouldn't be discounted, either. NV's workstation solutions are a substantial source of revenue that won't be easily cannibalized by the availability of high-performance integrated graphics products.
