Thread 723544563

AMDs next GPU will dethrone Nvidia!

2%

might unironically happen because nvidia doesnt give a shit about video games anymore. then again amd always fumbles the ball

>>723544618
more fumble with Lisa Su's tiddies if you catch my drift.

>>723544590
45%*

>>723544673
wow japan, like 10 people have gaming pcs there in total

>>723544563 (OP)
Well considering the 9070xt is trading blows with 5080(double the price btw) I believe it. Nvidia has stagnated

>>723544736
Nothing ever happens. Until it does, retard.

>>723544736
goalpost dorifto...!! sugoi!!!

>>723544563 (OP)
>no FSR 4 for RDNA 2 and 3
Fuck AMD and fuck /v/ for shilling their garbage GPUs.

>>723544780
>amd tostally has more than 2% marketshare if you ignore 99% of the world! youre shifting goalposts!

>>723544783
it needs specialized hardware faggot. might as well complain about gtx 1080ti not being able to ray trace.

>>723544908
>tostally
SAAR, DO NOT REDEEM THE BLOODY AMD BENCHOD
ONLY BUY NVIDIA SAAR

>>723544946
You can use DLSS 4 even on a RTX 2050.

>>723544956
>he made a typo? this is the chance of a lifetime, ill call him a jeet to win this argument!
lmao didnt radeon even have a street shitter in charge just a while ago?

>>723545040
yes, but not on a 1080ti, because it lacks the specialized hardware. similarly rdna1 and 2 just lack the hardware for fsr4.

>>723544757
now thats just stupid and im an amd advocate
the 9070xt has about the same punching power as the 5070TI for 150$ cheaper on the tradeoff of giving up power on path tracing and the more robust support of nvidia
The 5080 is an overpriced piece of ewaste but it eats the 9070xt
inb4 cherry picked graphs
get real anon, propaganda helps nobody

GPU Retail Sales Amazon US July '25

Total Units Sold: 41,800
AMD: 7,750 units (18.54% market share)
Intel: 1,100 units (2.63% market share)
NVIDIA: 32,950 units (78.83% market share)

Total Revenue: $20,996,254.50
AMD: $4,284,973.50 (20.41% revenue share)
Intel: $223,115.00 (1.06% revenue share)
NVIDIA: $13,488,166.00 (64.24% revenue share)


The 9070/60xt are doing ok

>>723544618
No, the best innovators all still work for Nvidia. It's true they don't care about gaming performance anymore but they still can fund technologies that keep them thoroughly entrenched and ahead of the competition without having to spend even 1% of revenue on r&d. The market is just to big for fast matrix math for AMD to ever really matter ever again. Unless Nvidia pulls a 2012-2020 Intel or some shit.

>>723546581
didnt a ton of nvidias senior engineers retire because the share price rose so much they became multimillionaires? nvidias drivers have been shit for a while which shows they arent infallible.

and it will cost like, 50 bucks less than nvidia while also performing 30% worse

>>723547038
to be fair even when they sold strictly superior cards at a lower price nvidia still ate their lunch. rember 7970 ghz vs gtx 680.

Uhh...

>>723546954
Many of their employees are wealthy but it's the opposite of the usual "rest and vest" problem you hear about in tech companies. They offer increasingly large stock compensation with long vest times as soon as employees first deal matures, and Jensen famously runs his employees until the ground so the deal is a vice grip that's worth the torture for many of their employees. Basically they're all terrified to leave because they want to get rich and he bends them over a barrel for it.
This article talks about it more https://www.tomshardware.com/tech-industry/big-tech/nvidias-stock-surge-mints-new-employee-millionaires-but-many-cant-yet-enjoy-their-wealth

>>723548540
seems about right, they cost the same, and while amd has 13% better raster, nvidia has better software support and path tracing.

>>723550178
>and path tracing.
you're not getting realistic gaming performance on sub-90 GPU with path tracing

Got me into thinking, why don't they finally split graphics card into two slots, one for the vRAM and the other processing power? You could finally decide how much vRAM you want to have installed, and it would reduce electronic waste.

>>723550212
ive tried path tracing in a couple of games with my 3080 and its been almost passable, im sure a 2 gens newer card can do it adequately even if you have to set dlss to a bit lower quality

>>723550525
The reason I've heard before is that the VRAM on a GPU is integrated into the board to enable the speeds and latencies that are needed for graphic workloads. It just doesn't work the same with a regular slot. Same reason the ryzen AI chips are sold with soldered RAM.
Not sure if it's true or just marketing lies to fuck with consumers though.

>>723550525
latency mostly

>>723544563 (OP)
it will. In power usage like the 6900XTX

>>723550898
if i ever dump my rtx 3080 / 5800x3d build ill probaby get an APU build, i dont play modern games enough to really justify a dedicated gpu nowadays that APUs are as fast as a gtx 1060.

>>723550985
a 8700G gets half the fps of a 6500XT

>>723544563 (OP)
It won't but I don't want 600W $3000 monster anyway.

unironically there's a high chance they'll make a monster shader die that can beat them at the flagship due to their chiplet design and udna making it not a gamer only waste of money