RDNA 5 Confirmed for 2027 as Nvidia Faces Pressure from AMD Flagship GPUs

RDNA 5 chiplet architecture aligns 2027 Xbox launch with next generation desktop and laptop GPU releases

Hardware by Masaru Hoshino on Feb 18, 2026

Quoting from AMD's quarter 4 earnings call, Lisa Sue said, “Development of Microsoft's nextG Xbox featuring an AMD semicustom SOC is progressing well to support a launch in 2027. RDNA5 shares chiplets between this and desktop cards.”

If Magnus and Lisa Sue confirmed this, and if the next-gen Xbox is ready for 2027, then the dedicated RDNA5 GPUs will be ready for 2027 as well. Things can slip. No one knows if it will be late or not. But this seems to fully confirm that RDNA5 is not just in Xbox, but probably in laptops and desktops as well, coming out next year.

RDNA 5, Confirmed for 2027, Nvidia Faces Pressure, AMD Flagship GPUs, NoobFeed

That raises a serious question. Should Nvidia care if AMD launches a flagship RDNA5 GPU? The exact compute unit count can change, but imagine a 154 compute unit RDNA5 that, in tests and simulations seen behind the scenes, has 10% better IPC and rasterization than RDNA4, which is already pretty fast per CU.

On top of that, it could have over double the ray tracing capabilities per CU. If we do the math on that, we are looking at something potentially 2 to 3 times faster than what AMD has now with RDNA4. If that happens, it would not just be competitive; it would be a destroyer.

Would Nvidia Respond to a Faster RDNA5?

Should Nvidia care if AMD launches a flagship that is way faster than the 5090, and they do not have anything ready for half a year? Even if gaming is more of a back-burner focus compared to AI, brand recognition, brand loyalty, and brand positivity still matter.

NVIDIA cannot let AMD completely run away with the market. Every time AI gets discussed, the AI bubble question comes up, and Nvidia has to consider that gaming remains a reliable market for them. For that reason alone, there has to be some type of response, even if it is low volume compared to what they usually put out.

We have seen Nvidia’s behavior before. There have been times when there was no supply for the 3060. Then the 6600XT came out, and suddenly the 3060 became available. NVIDIA always seems to have a launch ready when the competition enters the market.

If we read between the lines, it feels unlikely that Nvidia would have nothing prepared until 2028. If AMD is officially saying RDNA5 is launching next year, it suggests Nvidia is not 100% confident they can wait until 2028 without answering.

At a minimum, big RDNA5 could be at least 30% faster than a 5090, if not 50%, unless the design has issues. NVIDIA would not want to see a performance loss like that. Even if they push a 5090 to its limits, there is no 5090Ti yet. It is already around 600w.

Where would they push it next, 1000w? At most, maybe 20% more performance could be squeezed out, and that might still leave them 10% to 20% behind. In that scenario, would it make more sense to paper launch at least a 6090 and have it show up on charts at outlets like Hardware Unboxed? We will see. But unless AMD massively fails expectations with RDNA5 for some reason, it likely would not be enough for Nvidia to answer with just a minor refresh.

RDNA 5, Confirmed for 2027, Nvidia Faces Pressure, AMD Flagship GPUs, NoobFeed

The PS6 Timing Question

Another interesting question is whether Sony would be making a mistake by not launching the PS6 next year if the new Xbox arrives in 2027. However, it may not be as simple as reacting to what Xbox does. It might simply align with Sony’s usual cadence. We do not necessarily think Sony needs to think about what Xbox is doing as much as they used to. It is not quite the same dynamic as Nvidia versus Radeon.

There are documents stating production of a PS6 handheld and the PS6 itself in mid 2027, pointing toward a late 2027 launch. That timing perfectly coincides with when manufacturing should be underway for many RDNA5 chiplets for desktop and other products. Sony, at least as of a year ago, was planning to manufacture and then launch the console in late 2027.

The question becomes whether Sony should delay at all. If they delay until late 2028, would they keep the same design they already locked in? Or would they switch to 2nm and something like RDNA5.5 or RDNA6? That seems unlikely. Even if supply is lower at launch, it might make more sense to launch sooner rather than later.

There is also the PS6 handheld die, reportedly only 135mm². That suggests it could be a very cost-effective device. It could make more sense to launch that instead of continuing to rely on the PS5 base model.

If Microsoft comes out with something like Xbox Magnus, even if it is $1200, and Sony responds by saying they have something weaker coming out a year later, that could be a tactical mistake. Even if the PS6 ends up stronger, launching much later could reduce the hype factor.

If we see a Switch2 in 2025 and then new Xbox hardware in 2027, waiting too long could mean less excitement around Sony’s release. Launching closer to competing hardware keeps the momentum and attention high. Based on manufacturing timelines pointing to mid 2027, it still seems likely that the PS6 is targeting a late 2027 release.

Competitive Pressure Will Shape 2027

With confirmation that the next-gen Xbox is progressing for 2027 and RDNA5 is tied to both console and desktop chiplets, 2027 is shaping up to be a pivotal year. If AMD delivers a flagship GPU that is 30% to 50% faster than Nvidia’s top offering, Nvidia will almost certainly need a visible response to protect brand perception and market position.

At the same time, Sony's plan for launching the device will probably line up with the RDNA5 production schedule. Multiple big hardware releases could happen at the same time in 2027, either because they are planned to happen at the same time or because they need to compete. In a very short amount of time, the GPU and console market might change a lot if that happens.

Also, check our other hardware articles: