www.cnet.com
wccftech.com
| NOV 2021 | DEC 2021 | JAN 2022 | FEB 2022 | MAR 2022 | APR 2022 | MAY 2022 | JUN 2022 | JUL 2022 | AUG 2022 WK 1 | AUG 2022 WK 2 | Average | |
| AMD Share | 16.57% | 18.79% | 18.70% | 17.67% | 15.73% | 18.82% | 20.19% | 20.84% | 20.07% | 17.80% | 17.51% | 18.43% |
| NVIDIA Share | 83.43% | 81.21% | 81.30% | 82.33% | 84.27% | 81.18% | 79.81% | 79.16% | 79.93% | 82.20% | 82.49% | 81.57% |
| NAVI 6x | 26.38% | 27.40% | 29.37% | 28.93% | 29.60% | 37.05% | 35.07% | 35.81% | 37.43% | 35.82% | 35.57% | 32.58% |
| NAVI 5x | 18.64% | 18.81% | 18.65% | 17.26% | 15.31% | 17.55% | 17.76% | 17.63% | 17.63% | 16.48% | 16.33% | 17.46% |
| VEGA | 7.59% | 8.77% | 7.88% | 7.24% | 6.93% | 8.14% | 7.39% | 6.86% | 6.65% | 6.88% | 6.85% | 7.38% |
| POLARIS | 47.40% | 45.02% | 44.09% | 46.58% | 48.16% | 37.27% | 39.78% | 39.70% | 38.29% | 40.82% | 41.24% | 42.58% |
| AMPERE | 44.97% | 43.20% | 49.27% | 46.74% | 36.09% | 43.82% | 47.52% | 48.37% | 46.29% | 37.11% | 33.66% | 43.37% |
| TURING | 16.99% | 15.35% | 14.87% | 16.05% | 26.71% | 19.72% | 17.13% | 16.42% | 18.31% | 29.86% | 34.21% | 20.51% |
| TU 16x0 | 8.56% | 11.02% | 8.80% | 9.23% | 7.60% | 8.37% | 8.44% | 8.54% | 9.05% | 8.68% | 8.54% | 8.80% |
| PASCAL | 23.86% | 24.82% | 22.33% | 22.19% | 21.70% | 22.48% | 21.77% | 21.30% | 20.83% | 19.29% | 18.75% | 21.76% |
| MAXWELL | 5.62% | 5.61% | 4.74% | 5.79% | 7.91% | 5.62% | 5.14% | 5.37% | 5.52% | 5.06% | 4.84% | 5.56% |
This is being discussed in two threads, but when NVIDIA produces only discrete products (in categories JPR counts) and their shipments drop 25.7%, how can they suddenly drop only 19% when counting only discrete products (which were the only ones counted in the first place)NVIDIA Retained 80% Discrete GPU Market Share Versus AMD's 20% In Q2 2022 Despite Gaming Revenue Losses
NVIDIA seems to have retained its dominant position in the discrete GPU market against AMD despite having a brutal gaming revenue decline.
NOV 2021 DEC 2021 JAN 2022 FEB 2022 MAR 2022 APR 2022 MAY 2022 JUN 2022 JUL 2022 AUG 2022 WK 1 AUG 2022 WK 2 Average AMD Share 16.57% 18.79% 18.70% 17.67% 15.73% 18.82% 20.19% 20.84% 20.07% 17.80% 17.51% 18.43% NVIDIA Share 83.43% 81.21% 81.30% 82.33% 84.27% 81.18% 79.81% 79.16% 79.93% 82.20% 82.49% 81.57% NAVI 6x 26.38% 27.40% 29.37% 28.93% 29.60% 37.05% 35.07% 35.81% 37.43% 35.82% 35.57% 32.58% NAVI 5x 18.64% 18.81% 18.65% 17.26% 15.31% 17.55% 17.76% 17.63% 17.63% 16.48% 16.33% 17.46% VEGA 7.59% 8.77% 7.88% 7.24% 6.93% 8.14% 7.39% 6.86% 6.65% 6.88% 6.85% 7.38% POLARIS 47.40% 45.02% 44.09% 46.58% 48.16% 37.27% 39.78% 39.70% 38.29% 40.82% 41.24% 42.58% AMPERE 44.97% 43.20% 49.27% 46.74% 36.09% 43.82% 47.52% 48.37% 46.29% 37.11% 33.66% 43.37% TURING 16.99% 15.35% 14.87% 16.05% 26.71% 19.72% 17.13% 16.42% 18.31% 29.86% 34.21% 20.51% TU 16x0 8.56% 11.02% 8.80% 9.23% 7.60% 8.37% 8.44% 8.54% 9.05% 8.68% 8.54% 8.80% PASCAL 23.86% 24.82% 22.33% 22.19% 21.70% 22.48% 21.77% 21.30% 20.83% 19.29% 18.75% 21.76% MAXWELL 5.62% 5.61% 4.74% 5.79% 7.91% 5.62% 5.14% 5.37% 5.52% 5.06% 4.84% 5.56%
The other thread are responses to your original post on total gpu market share, so best keep the discussion there.
They count add-in boards, not discrete GPUs. For laptops that only includes MXM GPUs which is a miniscule fraction of the market
MXM hasn't existed for years in any capacity, surely they can't count it and call it counting notebook GPUs
And all but 3% of those gamers have 2060's or lower... because we know that 3% of Steam users have hardware equal to or more powerful than a PS5!But but but I heard otherwise...
NVIDIA Allegedly Increases Hopper H100 & Ampere A100 GPU Production Through TSMC Amidst China Chip Ban
NVIDIA seems to have accelerated the production of its Hopper H100 & Ampere A100 GPUs to supplement customers in China prior to the chip ban.
H100, A100 and HGX stations on top of the GPUs themselves. The ban also includes AMDs MI250 and possibly MI200.
I meant what "older or lower-end GPUs" would they be able to sell after the ban will become active? These should be made on 16nm at least as I understand which puts it on Pascal (Volta/Turing?) generation - does Nvidia even make these still?
I'd guess anything used in supercomputers need to be at least ready to quick resuming production as long as said computer is in active use. They're too expensive to go out of comission because you can't source replacement parts for failing hardware.
The whole GPU market is fucked from lack of competition, and I suppose this may never change. Even Intel is finding it impossible to enter given the decades of games and APIs made to run well only on NVIDIA and AMD hardware. Maybe Vulkan and DX12 will change this, but I'm sure IHV optimization will never go away, and who is going to optimize for a GPU that nobody has because nobody buys it because no games are optimized for it.
