Positioning & Architecture
NVIDIA GeForce RTX 5080 was the second GPU launch in January, right after the RTX 5090. These were followed by several other cards and AMD has since launched their RDNA 4-based Radeon RX 9070 and RX 9060 cards as well. This review marks our 12th(!)
review of a GeForce RTX 5080, so we've definitely achieved good coverage of what's available on the market. MSI has been busy behind the scenes and released a new version of their "Expert" graphics cards that they introduced last year. The new RTX 5080 Expert comes with improved cooling performance, but is still true to the design goal of offering something that is clean, stylish, and elegant, without the RGB overload that we've seen on some other recent releases.
The Blackwell architecture introduces several architectural improvements under the hood, like giving all shaders the ability to run FP32 or INT32 instructions, on Ada only half the cores had that ability. The Tensor Cores are now accessible from the shaders through a new Microsoft DirectX API, and they now support FP4 and INT4 instructions which run at lower precision, but much faster with less memory usage. There's numerous additional architecture improvements, we talked about all of them on the first pages of this review.
The GeForce RTX 5080 increases the number of GPU cores to 10,752, up from 10,240 on the RTX 4080 Super (+5%) and 9,728 on the RTX 4080 non-Super (+11%). Other unit counts are increased, too, the ROPs remain at 112. The memory capacity remains at 16 GB with a 256-bit bus, but NVIDIA has upgraded to the brand-new GDDR7 graphics memory chips.
From a fabrication perspective nothing has changed though—Blackwell is fabricated on the same 5-nanometer "NVIDIA 4N" TSMC node as last generation's Ada. NVIDIA claims this is a "4 nanometer process," but during Ada it was confirmed that NVIDIA 4N is actually not TSMC N4 (note the order of N and 4), but 5-nanometer. At the end of the day the actual number doesn't matter much, what's important is that NVIDIA is using the same process node.
The MSI Expert OC is a factory-overclocked custom design running at a rated boost clock of 2715 MHz, which is a medium-sized increase of 98 MHz over the reference clock speeds, or +3.7%. In addition to the factory overclock, compared to the Founders Edition, you do get a much more powerful and bigger cooler.
Performance
We measured a 4% performance uplift of the MSI Expert vs the NVIDIA Founders Edition, which puts it roughly in the middle of the OC'd models, which achieve between +2% and +6%. Generally, the RTX 5080 is a very decent card for 4K, but the gap towards last generation's RTX 4080 isn't exactly huge. We got +16% over the RTX 4080 Super, +18% vs the non-Super RTX 4080. These gen-over-gen gains are definitely MUCH less than expected and not nearly as much as what we saw on the RTX 5090, which beat the RTX 4090 by 35%. Compared to the GeForce RTX 3080, the performance increase is 75%, which means NVIDIA missed the "twice the performance every second generation" rule. Last-generation's flagship, the RTX 4090 is 11% faster than the RTX 5080 and the new RTX 5090 flagship is 46% faster, but twice as expensive.
GeForce RTX 5080 is still faster than AMD Radeon RX 7900 XTX, Team Red's best GPU, by 18% in a pure raster scenario, much more in RT. AMD's new Radeon RX 9070 XT is 25% slower at 4K, but the gap in ray tracing is now smaller. AMD has confirmed that they are not going to release more higher-end models than the RX 9070 XT, which means they don't pose a threat to the RTX 5080, which might explain why we're not getting bigger performance improvements—there's no incentive for NVIDIA to do so.
Ray Tracing & Neural Rendering
NVIDIA is betting on ray tracing and Blackwell comes with several hardware improvements here. Interestingly, the MSI RTX 5080 Expert runs only 12% faster at RT than RTX 4080 Super—remember, we got +16% in without RT. It looks like this is partly due to the game selection. The games that show the biggest gains in our non-RT test suite do not support RT. While RDNA 4 has definitely improved RT performance big time, the RX 9070 XT, AMD's fastest card in ray tracing is still 25% slower than the RTX 5080, at much lower pricing though. NVIDIA is also introducing several new optimization techniques that game developers can adopt. The most interesting one is Neural Rendering, which is exposed through a Microsoft DirectX API (Cooperative Vectors). This ensures that the feature is universally available for all GPU vendors to implement, so game developers should be highly motivated to pick it up.
VRAM
While the RTX 5090 went overboard with 32 GB VRAM, the RTX 5080 comes with 16 GB, which is a reasonable choice for this segment. Of course 24 GB would be nicer, also to achieve parity with AMD. On the other hand, this advantage is only psychological, because the 24 GB RX 7900 XTX still can't beat the RTX 5080, and the RX 9070 XT uses 16 GB. Building a RTX 5080 with 24 GB would require not only additional memory chips, the memory bus in the PCB layout would have to be increased as well and the memory controller inside the GPU, all of which would significantly increase the manufacturing cost of the whole card. Another approach is on the horizon though. Memory manufacturers are expected to release 3 GB memory chips shortly, enabling a 50% increase in VRAM sizes without altering the chip or PCB design. It is rumored that NVIDIA will adopt this approach for their next RTX 50 Super cards, to enhance VRAM capacity.
DLSS 4 Upscaling & Frame Generation
NVIDIA made a big marketing push to tell everyone how awesome DLSS 4 is, and they are not wrong. Firstly, DLSS 4 Multi-Frame-Generation. While DLSS 3 doubled the framerates by generating a single new frame, DLSS 4 can now triple or quadruple the frame count. In our testing this worked very well and delivered the expected FPS rates. When using FG, gaming latency does NOT scale linearly with FPS, but given a base FPS of like 40 or 50, DLSS x4 works great to achieve the smoothness of over 150 FPS, with similar latency than you started out with. Image quality is good, if you know what to look for you can see some halos around the player, but that's nothing you'd notice in actual gameplay.
Just to clarify, even multi-frame-generation will not turn 15 base FPS into 60 "playable" FPS. Sure, the movements will be very smooth, just like 60, but the game would react to your inputs at a rate of 15 FPS, which is noticeably sluggish. That's why the "base" FPS value is so important, as mentioned before, 40+ works very well, but it also depends on the type of game. Some slower titles, not shooters, will run great, even with a base FPS of 30.
Want lower latency? Then activate DLSS 4 Upscaling, which lowers the render resolution and scales up the native frame. In the past there were a lot of debates whether DLSS upscaling image quality is good enough, some people even claimed "better than native"—I strongly disagree with that—I'm one of the people who are allergic to DLSS 3 upscaling, even at "quality." With Blackwell, NVIDIA is introducing a "Transformer" upscaling model for DLSS, which is a major improvement over the previous "CNN" model. I tested Transformer and I'm in love. The image quality is so good, "Quality" looks like native, sometimes better. There is no more flickering or low-res smeared out textures on the horizon. Thin wires are crystal clear, even at sub-4K resolution! You really have to see it for yourself to appreciate it, it's almost like magic. The best thing? DLSS Transformer is available not only on GeForce 50-series, but uses Tensor Cores on all GeForce RTX cards! NVIDIA has constantly been working on DLSS Transformer support, and it's now available in virtually all modern titles, also thanks to their brilliant override mechanism. While AMD's FSR 4 has definitely improved quality a lot, Transformer still looks a little bit better, and FSR 4 support isn't as common as NVIDIA DLSS 4.
Physical Design, Heat & Noise
MSI did a great job coming up with a design theme that looks stunning and high-quality. Build quality is impeccable, both the backplate and the cooler shroud are made from thick metal with a nice matte texture—the perfect surface finish in my opinion. The "gold" color theme works well, too, everything reminds of the NVIDIA Founders Edition, with a bigger form factor of course, which allows better cooling performance. MSI's airflow concept has the front fan suck air in, it then moves sideways through the card and gets exhausted out the back. On the RTX 40 Expert this approach this didn't exactly impress us with its cooling performance—MSI definitely improved things for this new card. With 69°C and 33.7 dBA, the card runs at good temperatures and is reasonably quiet. However, it doesn't rank among the top for quietness or cooling, as some other RTX 5080 models from MSI do.
Our apples-to-apples cooler comparison test reveals that the MSI Expert cooler is considerably stronger than the NVIDIA Founders Edition, but weaker than competing custom designs. Based on these numbers, I believe MSI achieved a good balance between temperatures and fan noise.
PCI-Express 5.0
NVIDIA's GeForce Blackwell graphics cards are the first high-end consumer models to support PCI-Express 5.0. This increases the available PCIe bandwidth to the GPU, yielding a small performance benefit. Of course PCIe Gen 5 is backwards compatible with older versions, so you'll be able to run the RTX 5080 even in an older computer.
Just like we've done over the years, we took a detailed look at
PCI-Express scaling in a separate article today. Testing includes x8 Gen 5, for instances when an SSD is eating some lanes. The popular x16 4.0 was tested, which is common on many older CPUs and entry-level motherboards. Finally, some additional combinations were run, down to PCIe x16 1.1. The results confirm that unless you are on an ancient machine, PCIe bandwidth won't be a problem at all.
Power Consumption
Gaming power consumption of the MSI Expert is quite similar to the Founders Edition, a bit higher in gaming, due to the factory OC, Non-gaming power consumption was very high on the RTX 5090 (30 W), the RTX 5080 is doing much better here. While the default power limit of the Expert matches the reference setting of 360 W, the manual adjustment range has been increased by +10 W, bringing it to 400 W.
Overclocking
Overclocking the RTX 5080 Expert worked very well, we gained +11% in real-life performance on top of the factory OC, 12% with a maxed out power limit. This is much more than what we usually see on modern graphics cards. Unfortunately NVIDIA is limiting the maximum overclocking for the GDDR7 memory chips to +375 MHz—usually NVIDIA doesn't have any OC limits. At first, I was wondering why NVIDIA left so much performance on the table, especially when the card's gen-over-gen gains are so small, but then I realized that they might want to build a RTX 5080 Super next year. The problem is that RTX 5080 already maxes out the GB203 GPU, so additional units can't be enabled, and they'll have to rely on increases to clock speeds only. Looking at our numbers, higher clocks and memory speeds, some firmware optimizations, maybe even faster, higher capacity GDDR7 chips can definitely yield +10% in mass production—RTX 5080 Super spotted.
Pricing & Alternatives
Officially, the GeForce RTX 5080 is priced at an MSRP of $1000, which is a fantasy of course, at least in the United States. While prices and supply have improved, the cheapest RTX 5080 that you can find goes for $1350. Things are much better in Europe, which has plenty of RTX 5080 stock and many cards can be found for the equivalent of $1000. According to MSI, the US price of the RTX 5080 Expert OC is $1300, which is very reasonable given the current price situation (in the States). This would make it one of the most affordable models available. In Europe, the price is expected to be €1135, including VAT, which converts to $1100. Considering that the Expert offers a greatly improved experience over the Founders Edition, with a clean, high-quality, metal design and no distracting "gamer" fluff, I think it's reasonable to price this card at the lower end of the price spectrum. Especially in markets that don't have a Founders Edition, or where it's impossible to find one in stock, the card could also be a good alternative to the NVIDIA FE, at not that much higher cost. For those focused on cooling efficiency or the quietest operation, premium models like the MSI Vanguard, Suprim, ASUS TUF, and Astral are potential choices, but at a higher price tag. If the MSI Expert would end up at a similar price point as these, I would find it too expensive.
ASUS recently announced their creator-focused RTX 5080 ProArt, which comes with an additional USB-C output. I don't think that's an essential capability on such a card, but of course it could be a unique selling point for the (small) crowd of people who want to connect and daisy-chain monitors using USB-C, or use the standard for VR headsets.
The alternatives in this price segment are quite limited though. RTX 4080 ($850) and RTX 4080 Super ($1000) are still too expensive, no way you should buy those if you can find a RTX 5080 for just a bit more. RTX 5070 Ti is priced around $750 right now—definitely much better price/performance, but of course lower overall FPS. RX 9070 XT for $615 is definitely a good option if value is your focus—much better price/performance, no compromises with RT performance, but you'll lose out on all the DLSS 4 features, but you get support for FSR 4 and its high-quality AI-based upscaling model. Unless you know that you will probably never use RT or upscaling I don't think I'd pick a last-generation RX 7000-series GPU, unless it came with a huge discount.
If you already have a high-end GeForce RTX 40-series card, then there is no reason to upgrade. You're just missing out on multi-frame-generation, the DLSS Transformer model is supported on all older RTX cards, too. On the other hand, if you're coming from GeForce 30-series, then suddenly you'll get to experience frame generation, which will make a huge difference for your gaming experience.
Looking at future options, there isn't much. Maybe Intel's Arc B770, but I doubt it can reach RTX 5080 performance levels, and it's completely unknown when that launches, if ever. We've been hearing that NVIDIA has been working on their RTX 50 Super refresh, which brings bigger VRAM sizes thanks to 3 GB-sized memory chips. A RTX 5080 Super with 24 GB VRAM would definitely be interesting, even though the extra VRAM will ONLY make a difference in games that are VRAM starved with 16 GB, and I'm sure it will come with a price increase, too.