4 ways AMD can still get RDNA 4 wrong

If you’re tuned into the PC hardware space, you’re probably eagerly waiting for the release of AMD’s new RDNA 4 graphics cards. Rumors have been swirling for sometime regarding price and performance, and it won’t be long till we find out how this series of Radeon cards will actually fair. With NVIDIA’s complete fumble of the RTX 50 series launch with their melting 12VHP connectors, horrible stock issues, and predatory price practices, surely this will be a home run for AMD, right? I wouldn’t be so sure. NVIDIA has really screwed things up, make no mistake, but AMD can still mess up the RDNA 4 launch if they’re not careful of these 4 things.

4

Struggle with stock

If they’re not available, who cares?

Availability issues have plagued almost every big hardware release in the last couple years, but if NVIDIA’s latest stock fumbles are anything to go by, just having AMD cards available will be enough for some to switch teams. The stock of the RTX 50 series has been laughable, with most of it being grossly overpriced or tied up in borderline predatory bundles. AMD is in a bit of a different boat, having had RDNA GPU production running for some time now. Delaying the launch *should* have bought them valuable time to have enough of the cards available for launch day and beyond.

The bottom line is: the RX 9070 and 9070 XT have to be available on launch, for MSRP, as solo graphics card purchases. It’s crazy we even have to outline such expectations, but that’s where we’re at in the GPU world. If AMD can meet those marks, they’ve done the bare minimum. Oh, and hopefully it doesn’t combust. That would help.

3

Price them inappropriately

AMD needs to be aggressive if they want this to go well

We’ve already seen enough evidence that beating NVIDIA on price by ~$100 isn’t going to be enough. Saving 20% is simply not a good enough incentive for people to buy a Radeon GPU, and this is shown in AMD’s market share falling year over year. Despite the poor value proposition that an RTX 5080 offers over its predecessor, it would take a hefty discount to convince one to switch from green to red. There has been quite a bit of discussion online surrounding what price makes sense for the 9070 and 9070 XT, and quite frankly, AMD has to go big. These GPUs need to feel like a huge bargain at their original MSRP; no changing prices post-release like they did in prior generations. Getting it right out of the gate is critical.

So what price does make sense for these GPUs? There have been rumors swirling for some time, putting the XT anywhere between $1000 and $550. It’s worth remembering that manufacturers are no strangers to changing the price at the 11th hour of a release. Reviewers will have hardware in hand, ready to release reviews, with official pricing nowhere to be found. I think the $550 mark for the 9070 XT, while exciting, might be too good to be true, and anywhere close to $1000 is far too much for a card that will almost certainly be outpaced by the RTX 5080. The truth is somewhere in between, and hopefully, for our sake, it’s on the lower end of that range.

NVIDIA’s cards, even at their MSRP, are not stunning value propositions versus the previous generation. The 5090 is far and away the best GPU, but everything underneath that is so similar, it’s a bit mind-boggling. With the release of the 5070 Ti being incredibly underwhelming, it feels like this is just another version of the 4080, and this is clear when you look through gaming benchmarks. There were very marginal gains over the 4070 Ti Super, and it’s almost exactly the same as the 4080, which was already hardly different to the 5080. What a mess.

AMD can have a huge impact with this generation of cards if they can release something competitive, at a more than competitive price. It’s impossible to ignore the potential for stock issues, but if the 9070 XT releases with an aggressive price and gamers can actually get their hands on a card, AMD just might be cooking with gas.

2

Ship an undercooked FSR 4.0

DLSS can’t be the undisputed king

Source: AMD

For AMD to be competitive with RDNA 4, FSR 4.0 needs to be good. They can’t afford ship cards that have poor upscaling technology, especially if (when) they’re going to be worse in ray tracing performance. Early peeks at FSR 4 look really promising, with most of the growing pains seemingly gone, but we won’t know for sure until we get a much larger sample size of games tested. Delaying the launch will also hopefully buy some time for developers to get better support integrated for new releases.

FSR 3 just isn’t comparable to the image quality and smoothness seen with DLSS. Even DLSS 3 was miles ahead of anything a Radeon card was capable of. We need to see more parity in the upscaling space for this launch to be successful for AMD.

1

Have borked drivers on launch

It wouldn’t be the first time that Radeon drivers didn’t get enough time in the oven before a GPU launch. Even for NVIDIA, it seems like they’re experiencing some small driver quirks of their own, but they can afford a small mistep with drivers. AMD absolutely cannot afford to have bad drivers at launch, especially considering the delay RDNA 4 underwent. I’d be okay if performance wasn’t as good as it could possibly be, but they need to avoid game-breaking bugs and crashes, such as those that plagued prior Radeon launches.

Can AMD knock this out of the park?

The ball is firmly in AMD’s court at this point. The more RTX 50 series cards get launched, the better things continue to look for them, but they need to execute well. Pricing will be incredibly important, but they also need to be able to get the cards into the hands of users on launch. They can’t afford to fight in the same game of attrition that NVIDIA is playing.