New GPUs: Current Market Status

A couple of years ago, before AMD’s comeback in 2020, the general consensus was that NVIDIA GPUs were the faster and better cards.

But, with AMD’s RDNA (RX 5000) and RDNA 2 (RX 6000) GPUs, buying the right graphics card is no longer straightforward.

Things get even more complicated once you consider how many GPUs are out there, and they don’t stop coming.

So, if you want to find the right new GPU for your system while staying within budget, this is the guide you need.

Let’s check out the current GPU market status!

We already mentioned that NVIDIA has been on top of the GPU benchmarks for years. The GTX 980, GTX 1080 Ti, and RTX 2080 Super were miles away from AMD’s offerings.

With the RX 6000 Series AMD managed to catch up to the RTX 3090/Ti, but, with the RTX 4000 Series, NVIDIA is back on top.

The RTX 4090 is undoubtedly the fastest gaming GPU in the world.

However, that top-tier performance comes at a cost. Two of them.

The RTX 4090’s MSRP is $1,600, $100 more than the RTX 3090. Its closest competitor, the RX 7900 XTX, is $600 cheaper.

That’s not all, the RTX 4090 also has a 450W TDP (100W more than the 3090), so it is one of the most power-hungry GPUs we have ever seen!

Is this kind of price and power hike worth it? Let’s find out by looking at the performance!

Performance

The RTX 4090 is on top of the charts, no matter if its 1080p, 1440p, or 4K, by a long shot.

However, the performance gap is most impressive at 4K, which is the resolution that makes the most sense for such a powerful GPU.

In the 16-game tested, the RTX 4090 is 25% faster and 30% faster than the RX 7900 XTX and RTX 4080, respectively.

At lower resolutions, the performance gap with the RX 7900 XTX drops below 20%.

The RTX 4090 is definitely a beast of a card and destroys any other GPU, but is 30% performance worth it for $600 extra? It probably is if you don’t mind spending so much on a GPU and if you want the best of the best.

But, that’s not all!

The RTX 4090 can offer double the RX 7900 XTX ray tracing performance without any upscaling. With FSR/DLSS enabled on Quality mode, the 4090 is about 66% faster, as seen in this Cyberpunk 2077 RT benchmark.

NVIDIA’s advantage doesn’t end there either!

NVIDIA’s DLSS 3/Frame Generation updates make the RTX 4090 even faster than AMD’s flagship. DLSS 3 seemingly doubles FPS in games that support it while AMD still doesn’t have such a feature with FSR.

Specifications

Base Frequency	2230 MHz
Boost Frequency	2520 MHz
Memory Size	24GB GDDR6X
Memory Interface	384-bit
Memory Speed	21 Gbps
TDP	450W
Process Size	TSMC 4nm

RDNA 2/RX 6000 is AMD’s first GPU series in years that’s competitive with NVIDIA’s high-end segment. So competitive that the RX 6950 XT ended up being faster than the RTX 3090 in 1080p.

With RDNA 3, AMD fires again at NVIDIA, but this time around, it’s a response to the RTX 4080 rather than the RTX 4090.

The RX 7900 XTX is AMD’s fastest gaming GPU for $1,000. Expensive, but it is significantly cheaper compared to its competitor, the $1200+ RTX 4080.

Performance

Now, let’s get to what you’ve all been looking for. The performance of this GPU.

To help you understand the difference between the two flagship GPUs from team red and team green, we will use Hardware Unboxed’s chart.

Full HD, 1440p, or 4K, the performance difference between the RTX 4080 and RX 7900 XTX is practically the same.

Usually, NVIDIA’s cards perform better at higher resolution, but, surprisingly enough, the RX 7900 XTX is slightly faster than the 4080 at 4K.

But that’s not the whole story.

Looking at raw ray tracing performance, the RTX 4080 is about 27% faster than the RX 7900 XTX. That’s something to keep in mind.

Furthermore, we have to acknowledge that NVIDIA offers DLSS and Frame Generation, while AMD’s FSR 2.0 doesn’t offer anything similar to Frame Generation, yet.

Although AMD recently announced FSR 3, which should offer double the performance of FSR 2.0.

Specifications

Base Frequency	1855 MHz
Boost Frequency	2500 MHz
Memory Size	24GB GDDR6
Memory Interface	384-bit
Memory Speed	20 Gbps
TDP	355W
Process Size	TSMC 5nm

The RTX 4000 Series took everyone by surprise with its ridiculously high pricing. The RTX 4080 was especially surprising with its jaw-dropping $1200 price tag.

NVIDIA got accustomed to the RTX 3000 scalper pricing and raised the price by 70% (in relation to the RTX 3080).

Even considering NVIDIA’s big jump in performance and the introduction of DLSS 3, which effectively doubles the FPS in games, $1200 for what used to be a $700 GPU is still too much.

Either way, it’s an impressive GPU, so let’s look at some benchmarks.

Performance

The RTX 4080 is expensive on its own, but it’s even harder to justify considering the RX 7900 XTX costs $200 less.

In this Cyberpunk 2077 benchmark, the AMD flagship is faster, but overall, the two GPUs are on par.

Furthermore, the RTX 4070 Ti is only about 25% slower than the RTX 4080 while costing $400 less.

And, of course, the 4080 is about 27% faster than the RX 7900 XTX in ray tracing (check out the benchmark in the RX 7900 XTX segment).

Specifications

Base Frequency	2205 MHz
Boost Frequency	2505 MHz
Memory Size	16GB GDDR6X
Memory Interface	256-bit
Memory Speed	22.4 Gbps
TDP	320W
Process Size	TSMC 4nm

$1000 is a lot for a graphics card, so the RX 7900 XTX is not accessible to most people. Unfortunately, AMD’s cheaper option isn’t all that cheaper.

The RX 7900 XT is just $100 cheaper than its bigger brother, so its MSRP is $900. Considering the performance difference between the two, paying that $ 100 extra is well worth it.

It’s especially difficult to consider AMD when NVIDIA’s contender, the RTX 4070 Ti is $800.

However, it seems that AMD is slowly pushing for a cheaper price since we often see the RX 7900 XT discounted to around $800.

Performance

The RX 7900 XT and RTX 4070 Ti go head to head and deliver almost the exact same in-game performance.

You can see that in this extensive test by Hardware Unboxed.

Even when including games with ray tracing enabled, the two GPUs are still neck and neck.

But, ultimately, AMD loses out since their GPU is $100 more. If AMD officially reduces the price of the RX 7900 XT then it will be a serious contender.

Specifications

Base Frequency	1500 MHz
Boost Frequency	2025 MHz
Memory Size	20GB GDDR6
Memory Interface	320-bit
Memory Speed	20 Gbps
TDP	300W
Process Size	TSMC 5nm

The RTX 4070 Ti, unfortunately, is the best value GPU in the RTX 4000 Series. It’s unfortunate because it’s still priced ridiculously high at $800. The price is especially a problem since the RTX 4070 doesn’t even exist, there is no alternative.

However, if you’re certain that you need an RTX 4000 GPU, the RTX 4070 Ti makes a lot more sense than the RTX 4080 or RTX 4090.

Performance

We already covered the performance of the RTX 4070 Ti against the RX 7900 XT. The conclusion is that both GPUs offer the exact same level of performance.

Which one of the two you should get depends on two things: price and how much you appreciate ray tracing.

If you prefer ray tracing over raw rasterization performance, the RTX 4070 Ti is the obvious choice since it is about 7-8% faster. With upscaling it can be up to 25% faster than the RX 7900 XT.

Specifications

Base Frequency	2310 MHz
Boost Frequency	2610 MHz
Memory Size	12GB GDDR6X
Memory Interface	192-bit
Memory Speed	21 Gbps
TDP	285W
Process Size	TSMC 4nm

We looked at the fastest, most expensive, and high-end GPUs of both team red and team green.

But now, let’s look at the more sensible offers, or more specifically, the best value high-end GPU, the RTX 3070.

For just $500, the RTX 3070 can bring near-RTX 3080 performance while drawing considerably less power. Don’t believe it? Here are the performance numbers.

Performance

The two AMD contenders against the RTX 3070 are the RX 6800 and RX 6700 XT. But, whichever way you look at it, the RTX 3070 comes out on top.

In the Cyberpunk 2077 @1080p/High (left image) benchmark, the RTX 3070 is 15% faster than the RX 6700 XT and less than 2% slower than the RX 6800. In GTA V, on the other hand, the RX 6700 XT comes out 6% faster than the 3070.

However, even with this back and forth, with a larger game library, the RTX 3070 is around 10% faster on average vs. RX 6700 XT.

Ray tracing further widens the gap between the two GPUs.

Specifications

Base Frequency	1500 MHz
Boost Frequency	1725 MHz
Memory Size	8GB GDDR6X
Memory Interface	256-bit
Memory Speed	14 Gbps
TDP	220W
Process Size	Samsung 8nm

Now we’re finally getting into the budget-friendly territory. The RTX 3070 is a great buy, but there’s no contender for the RTX 3060 Ti’s impeccable price-per-performance value.

Hardware Unboxed claims that the RTX 3060 Ti has a $3.70 cost per frame, beating the RTX 3070’s $4.20 cost per frame.

Here’s the chart for further info:

A pretty good deal for just $400 (MSRP).

Performance

Performance-wise, this $400 GPU outperforms NVIDIA’s previous generation flagship, like the RTX 2080 Super, and even closes in on the RX 6700 XT and RTX 2080 Ti.

So, if you’re looking to max out games at 1080P, the RTX 3060 Ti is the go-to buy. With DLSS, you could even enable Ray Tracing or go 1440P while staying above 60 FPS.

Specifications

Base Frequency	1410 MHz
Boost Frequency	1665 MHz
Memory Size	8GB GDDR6
Memory Interface	256-bit
Memory Speed	14 Gbps
TDP	200W
Process Size	Samsung 8nm

The budget segment of the current GPU market status isn’t doing too good. With disappointing offers like the RX 6500 XT, RTX 3050, and RX 6600 XT, either because they’re priced too high or perform too slow, it’s hard to justify a budget GPU.

The RTX 3050 is marginally better as a budget GPU than AMD’s offering at $250 MSRP, but it still isn’t as good as we would like. Below, you’ll see why.

Performance

Again, we’ll use one of Hardware Unboxed’s charts as an insight into performance.

With a 12-game average, the RTX 3050 is only 5% faster than the GTX 1660 Super and considerably faster than the $200 RX 6500 XT.

While this might seem impressive, consider that the GTX 1660 Super is two and a half years older and $30 cheaper (at MSRP).

You could argue that the RTX 3050 has other benefits like ray tracing capabilities and DLSS, but we simply don’t think that’s enough to justify the $250 price tag.

Unfortunately, there is no other contender at this price point, so it is the best budget pick for now.

Specifications

Base Frequency	1552 MHz
Boost Frequency	1777 MHz
Memory Size	8GB GDDR6
Memory Interface	128-bit
Memory Speed	14 Gbps
TDP	130W
Process Size	Samsung 8nm

Newly Released GPUs

Upcoming GPUs

There are some great picks from this list, but the current GPU market status isn’t all that great. So, if you’re looking for better value out for your money and are willing to wait for future generations, you’re in luck.

Soon, we should see AMD’s new RDNA 3-based RX 7000 GPUs released around October 2022.

RDNA 3 will supposedly bring substantial performance improvements over the current generation and perform much better at ray tracing.

The RTX 4000 series, codenamed Ada Lovelace, should also release around that same time and deliver a substantial increase in performance. However, that performance may come at the cost of a power draw.

Future GPUs to look out for:

• NVIDIA RTX 4080 – The best high-end performance for a sensible price
• NVIDIA RTX 4070 – The sweet spot for high-end performance
• AMD RX 7800 XT – Potentially the better alternative to the RTX 4080
• AMD RX 7800 – Contender for the best overall GPU
• NVIDIA RTX 4060 Ti – Potentially taking over the best price-per-performance spot

Make sure to come back and check out the guide when these new GPUs do release!