Can the AMD RX 8000 series beat the Nvidia RTX 5000 series after release?

Picture of AMD
AMD's RDNA 3 die (Image via AMD)

AMD is gearing up for the launch of its AMD RX 8000 series, and we are already seeing the driver patches on Linux. Previously, Linux driver patches could only be found one or two months prior to the GPU launch, and they served as a good indicator of an imminent launch. However, it seems AMD is prioritizing the Linux driver patch much earlier so they can ensure stellar performance during launch.

We know the new GPU launch from AMD will happen sometime this year. But how good will these new AMD RX 8000 series GPUs be and can they beat Nvidia's upcoming RTX 5000 series? If the rumors are true, they won't, at least not the high-end cards. But when it comes to mid-range and budget value categories, there is a good chance they will.

This article will tell you how the AMD RX 8000 series is planning to beat the Nvidia RTX 5000 series.

Note: This article is subjective and solely reflects the writer's opinions.


AMD RX 8000 series: Expected specs and features

Before we compare the AMD RX 8000 series to the Nvidia RTX 5000 series, we need to understand what each GPU maker is bringing to the table. According to Kepler_L2, a prominent leaker, AMD will not have any high-end models in this GPU generation. Basically, AMD is not going to compete with Nvidia's high-end models like RTX 5080 or RTX 5090.

This is not the first time AMD has skipped the high-end GPU market. They did the same with the RX 5000 series (RDNA 1), where the most powerful GPU was the AMD RX 5700 XT. They also the same with the RX 400 and RX 500 series, so the news of AMD skipping the high-end once again is not coming out of left field.

Here's everything you need to know about the AMD RX 8000 series.

GPU architecture and performance target

RDNA 3 GPU die with chiplet design (Image via AMD)
RDNA 3 GPU die with chiplet design (Image via AMD)

The AMD RX 5000 series was based on the RDNA 1 architecture, the 6000 series on the RDNA 2 architecture, the 7000 series on the RDNA 3 architecture, and the AMD RX 8000 series will be based on the RDNA 4 architecture.

Multiple leakers, including Kopite7kimi, Olark29_, and Moore's Law is Dead, have confirmed that the RDNA 4 architecture powering the AMD RX 8000 series mainly consists of two GPU dies, Navi 44 and Navi 48. The Navi 48 die is the most powerful in this generation, while the Navi 44 die is budget-oriented.

According to leaker Moore's Law is Dead, the Navi 48 GPU die offers similar performance as the Navi 31 die that powers both AMD RX 7900 XT and RX 7900 XTX—AMD's current high-end GPUs. The full Navi 31 die powers the former, while a cutdown variant powers the latter.

So if the Navi 48 GPU die of the AMD RX 8000 series ends up performing somewhere between the AMD RX 7900 XT and RX 7900 XTX, it will be great news for gamers. That's because AMD is not targeting the high-end market, so the Navi 48 GPU will be cheaper while offering similar performance to the current high-end GPUs that launched for $899 to $999.

The Navi 44's expected performance range is around the AMD RX 7700 XT and Nvidia RTX 4060 Ti, but it is expected to come with a more accessible price tag. This performance target is further backed by leaker All_The_Watts.


Memory system

Large capacity of GDDR6 memory (Image via AMD)
Large capacity of GDDR6 memory (Image via AMD)

According to Moore's Law is Dead and other leakers, AMD is not going with the latest GDDR7 memory and instead sticking with the previous GDDR6 memory.

However, a new leak from Kepler_L2 suggests that the GDDR6 memory chips are not the faster 20Gbps variant but rather the 18Gbps one.


Ray tracing and upscaling

AMD's second generation of ray tracing (Image via AMD)
AMD's second generation of ray tracing (Image via AMD)

According to Kepler_L2, AMD is going all in on ray tracing and AI upscaling with the RDNA 4 architecture inside the AMD RX 8000 series GPUs. This series is going to have a BVH8 instruction set, which doubles the ray tracing throughput per clock cycle. The RDNA 2 and 3 powering the AMD RX 6000 and 7000 series only have a BVH4 instruction set.

Some reports also suggest the ray tracing cores will be decoupled from compute cores so they will have different clock speeds. Hence, expect a big boost in ray tracing performance, which could be enough to make it closer to the ray tracing game performance of the Nvidia RTX 5000 series when that launches.

AMD's CTO Mark Papermaster also confirmed that they are working on a newer upscaling technique based on AI, similar to Nvidia and Intel. So AMD's FSR could get a huge new update that pushes its visual quality closer to Nvidia's DLSS.


Nvidia RTX 5000 series: Expected specs and features

Nvidia Blackwell architecture (Image via Nvidia)
Nvidia Blackwell architecture (Image via Nvidia)

Unlike AMD, Nvidia seems to be looking to release two high-end GPUs, which will likely be followed by the launch of their mid-range and budget GPUs after a certain timeframe. Panzerlied, a known leaker from ChipHell forums, says that the RTX 5090 will bring the following improvements over the RTX 4090:

  • 50% shader core count
  • 52% memory bandwidth
  • 78% L3 cache size
  • 15% GPU clock frequency.

If you combine all these improvements with some math, the RTX 5090 may bring an overall performance improvement of 70% over the RTX 4090. This is huge news as the RTX 4090 offered around 60 - 70% performance improvement over the RTX 3090.

The RTX 5090 will also have the latest GDDR7 memory with a 512-bit memory bus. However, the mid-range and budget models of the RTX 5000 series are probably not getting such huge improvements considering Nvidia's previous records. They will probably not have the same 512-bit memory bus lane, but they will still retain the GDDR7 memory with a 384-bit and 256-bit memory bus.


Can the AMD RX 8000 series beat the Nvidia RTX 5000 series?

If the leaks are true, the Nvidia RTX 5090 will bring another 70% improvement over the RTX 4090, which already brought a 60 - 70% improvement over the RTX 3090. So can the AMD RX 8000 series beat the Nvidia RTX 5000 series? No, and according to the leaks, they are not even going to try.

It looks like AMD is not targeting high-end GPU market with this generation. However, they are planning to beat the budget Nvidia RTX 5000 series models with their Navi 48 and 44-based GPUs. The Navi 48 GPU die promises to deliver a performance between AMD RX 7900 XT and 7900 XTX. So it will probably deliver a similar performance to that of the RTX 4080 Super, and if AMD manages to provide this level of performance at a decent price, then that's still a win.

They were quite successful with this strategy before with the RX 400 series and later with the RX 5000 series, and the same may happen again. It also looks like AMD is planning to beat Nvidia in the mid-range and low-end segment in rasterization and ray tracing performance, but it remains to be seen by how much.

Quick Links

App download animated image Get the free App now