But it is clear AMD has prioritised production of consoles and CPUs over dGPUs. You saw that by some of the numbers that got leaked out about 7NM wafer allocations during the pandemic. AMD dGPUs are easy to get in Western Europe and the US,but in many parts of the world it is much easier to find Nvidia products.
You can also see it this generation - there is more concentration on cutting production costs. AMD has access to the same TSMC 4N 5NM process as Nvidia but is using it for APUs:
The All-New Radeon 780M Is The Fastest iGPU! Ryzen 9 7940HS Hands On First Look
The all New AMD Ryzen 9 7940HS RDNA3 APU is crazy fast and in this video we test out the built in radeon 780M RDNA3 Based iGPU! The New AMD Ryzen RDNA3 APUs ...www.youtube.com
This is an IGP which clocks upto 3GHZ in a TDP constrained environment. Navi 31 doesn't even have a big GCD and its made on bog standard TSMC 5NM(Nvidia uses TSMC 4N 5NM). Nvidia went for a 600+ MM2 top die dGPU,and AMD went for a mixed process node dGPU closer to 500MM2.Navi 33 looks like a slightly die shrunk Navi 23 made on an economy TSMC 6NM process. They are not using GDDR6X or stacked cache either.
The best we can hope for is that in the mainstream area they produce decent enough dGPUs to sort of keep Nvidia in check.
That phoenix apu is powerful. The Asus ROG Ally is gonna be a beast! AMDs approach works for them - if you think about it, what does AMD sell - they sell silicon and they've found that for their market they can make more money selling high volume small silicon chips, ao naturally they are not as focused as selling large silicon chips such as those used in desktop GPUs
Last edited: