Ever since NVIDIA unveiled the new Blackwell-based 50-series GPUs, people have been speculating about what this could mean for gaming and content creation and whether it would truly be a leap compared to the 40-series. Well, we now have the RTX 5080 Founders Edition in the office and have been testing it over the past week. While it does not compete against an RTX 5090 or even an RTX 4090, it is a solid offering that will be the mainstream gaming card to beat going into 2025.
With the latest features like DLSS 4, new GDDR7 memory and a slimmer design, this feels like the card made for people who crave the latest and greatest but can’t justify the price tag of a 90-range card. Looking at the specs, it brings a fair amount to the table, especially when it comes to what DLSS 4 and the new transformer models can do for frame generation and delivering smoother 4K video. But I am getting a bit ahead of myself here; it is worth looking at what you get when you finally unbox the RTX 5080 and prepare to slot it into your gaming or content creation rig.
If you have used the NVIDIA RTX 4080 or RTX 4080 Super, you know how big those cards are. NVIDIA has changed things up with this release, and much like the RTX 5090 Founders Edition, the RTX 5080 Founders Edition is a much smaller card compared to the 40-series offerings, now taking up only two slots in your PC. It is still a beastly GPU, with a length of 304 mm (12 inches), a height of 48 mm (1.9 inches) and a width of 137 mm (5.4 inches). It maintains the design language and clean look we saw with the RTX 5090, boasting a dual-fan design and minimal lighting that looks fantastic in and out of the PC.
| RTX 5090 | RTX 5080 | RTX 5070 Ti | RTX 5070 | |
|---|---|---|---|---|
| CUDA cores | 21760 | 10752 | 8960 | 6144 |
| Boost clock (GHz) | 2.41 | 2.62 | 2.45 | 2.51 |
| Base clock (GHz) | 2.01 | 2.3 | 2.30 | 2.16 |
| Tensor core TOPS | 3352 | 1801 | 1406 | 988 |
| Ray tracing core TFLOPS | 318 | 171 | 133 | 94 |
| Memory | 32 GB GDDR7 | 16 GB GDDR7 | 16 GB GDDR7 | 12 GB GDDR7 |
| Memory bus width | 512-bit | 256-bit | 256-bit | 192-bit |
| Total Graphics Power (watts) | 575 | 360 | 300 | 250 |
| Required system power (PSU wattage) | 1000 | 850 | 750 | 650 |
| Power connector | 1x 600 W PCIe Gen 5 OR 4x PCIe 8-pin adapter | 1x 450 W PCIe Gen 5 OR 3x PCIe 8-pin adapter | 1x 300 W PCIe Gen 5 OR 2x PCIe 8-pin adapter | 1x 300 W PCIe Gen 5 OR 2x PCIe 8-pin adapter |
| Price | $1,999 | $999 | $749 | $549 |
| Release date | Jan 30, 2025 | Jan 30, 2025 | Feb 2025 | Feb 2025 |
Looking at the specs, the RTX 5080 is a step down from the monster of a GPU the RTX 5090 delivers. It is built on the same 4nm TSMC process and features a base clock of 2295 MHz with a solid boost clock of 2617 MHz, with 10752 CUDA cores, a Tensor core TOPS of 1801, a Ray Tracing core TFLOPS of 171, and 16GB GDDR7 memory on a 256-bit bus. This is all powered on a TGP of 360W, with a suggested PSU of 850W to power everything this GPU can do.
There are a lot of numbers to consider, but a few key takeaways help explain how this GPU fits into the 50-series lineup. The CUDA core count is about half of what you’d find on an RTX 5090. Even though it’s clocked faster, it has fewer cores to work with, leading to lower performance than NVIDIA’s new flagship. Additionally, it runs on a 256-bit memory bus compared to the massive 512-bit bus on the RTX 5090. While this doesn’t directly translate to half the performance of the RTX 5090, it clearly limits the RTX 5080’s capacity to render frames or handle computing workflows efficiently.
As with the RTX 5090, the RTX 5080 includes all the DLSS 4 features that made this generation so exciting. This means it benefits from transformer model improvements to frame generation and all the enhancements in speed and performance these advancements provide. I discussed DLSS 4 extensively in the RTX 5090 review, but I’m still impressed with how NVIDIA strategically enhanced this generation’s technology. The transformer models operating in the background deliver a gameplay experience so smooth it almost defies belief. This also means that while the RTX 5080 is less powerful than the RTX 4090 on paper, incorporating DLSS 4 allows it to achieve framerates close to the RTX 4090’s, even with ray tracing. I’ll delve further into this in the benchmarking section.
On the topic of benchmarking, as noted in the RTX 5090 Founders Edition review, CGMagazine has updated its benchmarking process. We now use two nearly identical systems: both equipped with an AMD Ryzen 7 9800X3D in an ASUS ROG Crosshair Hero motherboard, paired with 32 GB of DDR5 G.Skill RAM running at EXPO 6000. For storage, the systems feature PCIe Gen 4 SK Hynix 2 TB M.2 SSDs and are powered by 1,200-watt PSUs—one from NZXT and the other from be quiet!
We’ve aimed to make these systems as similar as possible, but some components differ based on availability during assembly. One system uses an NZXT Kraken AIO, while the other employs a Corsair Titan AIO. Both SSDs are clones to avoid software discrepancies, and the Windows 11 installations include all updates. The latest NVIDIA press driver was used for testing NVIDIA cards, while the most recent AMD Radeon driver was applied when benchmarking the RX 7900 XTX for comparison.
Starting with the synthetic benchmarks, the NVIDIA RTX 5080 Founders Edition GPU delivers strong results across the board, even when compared to the previous-generation RTX 4090. As expected, it outperformed both the RTX 4080 and the AMD Radeon RX 7900 XTX. It boasts an 11 per cent increase in performance on 3DMark Time Spy and was only around five per cent behind the RTX 4090. In the 3DMark Port Royal benchmark, the RTX 5080 showed a 20 per cent improvement over the RTX 4080 but fell short of the RTX 4090 by just over 14 per cent in this test.
“Starting with the synthetic benchmarks, the NVIDIA RTX 5080 Founders Edition GPU delivers strong results across the board, even when compared to the previous-generation RTX 4090.”
Things become more interesting when examining the 3DMark DLSS benchmark, which tests DLSS 4 technology. Here, the RTX 5080 demonstrated a significant advantage, particularly over the RTX 4090. With DLSS 4’s 4X frame generation enabled, the RTX 5080 delivered an impressive 316 FPS compared to the RTX 4090’s 192 FPS. However, when looking at raw performance without DLSS 4, the RTX 4090 still leads, scoring 56.89 compared to the RTX 5080’s 48.17. These results highlight the substantial impact of DLSS 4 frame generation in boosting potential framerates, especially when ray tracing is factored in.
| Synthetic Benchmarks | RTX 5090 | RTX 5080 | RTX 4090 | RTX 4080 | RX 7900 XTX |
|---|---|---|---|---|---|
| 3D Mark Time Spy | 47148 | 31770 | 33 405 | 28 418 | 30 434 |
| 3D Mark DLSS | 81.36 | 48.17 | 56.89 | 39.26 | – |
| 3D Mark DLSS + FG | 434.79 | 316.58 | 192.86 | 149.24 | – |
| 3D Mark Port Royal | 35780 | 21889 | 25295 | 17885 | 16278 |
| V-Ray (RTX) | 15076 | 9236 | 10759 | 7519 | – |
| Procyon Video GPU Editing | 65694 | 61896 | 55 896 | 52868 | 52460 |
| Procyon PHI 3.5 | 6050 | 4643 | 5155 | 4285 | 2298 |
| Procyon MISTRAL 7B | 6674 | 4798 | 5242 | 3957 | 2426 |
| Procyon LLAMA 3.1 | 6492 | 4588 | 4933 | 3693 | 2264 |
| Procyon LLAMA 2 | 7002 | 4698 | 5147 | 3912 | 2567 |
| Blender – Junkshop | 3927 | 2293 | 2700 | 2062 | 1144 |
Jumping to creative software and AI benchmarks, the RTX 5080 outperforms all previous-generation cards tested in the Procyon Video GPU Editing benchmark, even surpassing the RTX 4090 in this test. When evaluating different LLM workflows, the RTX 5080 delivers solid results, though it falls behind the RTX 4090 due to the latter’s raw power, increased memory, and greater number of cores.
It is worth noting that while the RTX 5080 performs well in LLM tests, its 16GB of memory may struggle with larger models, potentially resulting in slower performance. The RTX 5090 is better equipped to handle AI workflows, but it is encouraging to see the RTX 5080 perform strongly overall in this area. Smaller models will work fine, but once you try to push things into the 70B range, you will have a bad time, or at the very least, things will run much slower than you may want.
| 4K Benchmark | RTX 5090 | RTX 5080 | RTX 4090 | RTX 4080 | RX 7900 XTX |
|---|---|---|---|---|---|
| Shadow of the Tomb Raider | 150.00 | 97.00 | 117.00 | 85.00 | 80.00 |
| Shadow of the Tomb Raider + Frame Generation | 269.00 | 165.00 | 211.00 | 168.00 | – |
| Returnal | 138.00 | 98.00 | 110.00 | 85.00 | 95.00 |
| Assassin’s Creed Valhalla | 142.00 | 120.00 | 138.00 | 108.00 | 108.00 |
| Watch Dogs Legion | 94.00 | 91.00 | 103.00 | 81.00 | 43.00 |
| Cyberpunk 2077 (Ultra) | 102.30 | 71.60 | 70.19 | 56.00 | 63.56 |
| Cyberpunk 2077 RT | 56.29 | 35.47 | 42.36 | 29.90 | 20.08 |
| Cyberpunk 2077 RT + Frame Generation | 351.77 | 241.25 | 169.00 | 79.25 | 59.12 |
| F1 24 (Ultra Preset) | 217.00 | 134.00 | 162.00 | 120.00 | 67.00 |
| F1 24 + Frame Generation | 259.00 | 166.00 | 197.00 | 147.00 | 112.00 |
| Metro Exodus (Extreme) | 96.85 | 57.46 | 74.37 | 52.18 | 61.62 |
| Black Myth: Wukong | 59.00 | 37.00 | 43.00 | 31.00 | 8.00 |
| Black Myth: Wukong + Frame Generation | 99.00 | 61.00 | 70.00 | 52.00 | 16.00 |
This brings us to what most people are looking at this card to deliver on, and that is 4K gaming. This is when things get very interesting and where the new DLSS 4 features truly shine. Looking at the raw performance without any extra features thrown into the mix, the RTX 5080 is a notable, if not massive,leap when comparing it to the RTX 4080.
In games like Assassin’s Creed Valhalla, the RTX 5080 Founders Edition managed just above a 10 per cent uptick in performance compared to the RTX 4080, with it delivering 120 FPS over 108 FPS. This average level of performance is consistent across the board, giving a general taste of what the new GPU can do and how much faster it is compared to the last generation.
Things get exciting when you throw in DLSS 4 into the mix, and once you throw on 4X frame generation, and as we saw with the 3D Mark test, things get very interesting. Looking at Cyberpunk 2077, the game moves from a slow 35.47 FPS with ray tracing on, beating out the RTX 4080’s 29.90 by only around 5 FPS, to a staggering 241.25 FPS compared to the DLSS 3 performance of 79.25, a 2X jump in performance, delivering a very smooth experience that makes 4K ray tracing very possible, even on a card half the cost of the flagship 5090. It even beat out the DLSS performance of the RTX 4090, a card that costs significantly more even today than the $999 RTX 5080.
| 1440P Benchmark | RTX 5090 | RTX 5080 | RTX 4090 | RTX 4080 | RX 7900 XTX |
|---|---|---|---|---|---|
| Shadow of the Tomb Raider | 259.00 | 182.00 | 206.00 | 159.00 | 144.00 |
| Shadow of the Tomb Raider + Frame Generation | 338.00 | 282.00 | 281.00 | 238.00 | – |
| Returnal | 206.00 | 157.00 | 168.00 | 141.00 | 155.00 |
| Assassin’s Creed Valhalla | 191.00 | 179.00 | 205.00 | 178.00 | 170.00 |
| Watch Dogs Legion | 135.00 | 95.00 | 145.00 | 128.00 | 79.00 |
| Cyberpunk 2077 (Ultra) | 191.85 | 100.43 | 149.62 | 123.88 | 133.44 |
| Cyberpunk 2077 RT | 102.88 | 91.10 | 100.15 | 61.23 | 55.00 |
| Cyberpunk 2077 RT + Frame Generation | 467.57 | 386.26 | 251.34 | 135.72 | 83.45 |
| F1 24 (Ultra Preset) | 228.00 | 204.00 | 242.00 | 183.00 | 124.00 |
| F1 24 + Frame Generation | 374.00 | 262.00 | 308.00 | 241.00 | 219.00 |
| Metro Exodus (Extreme) | 155.88 | 154.20 | 128.72 | 141.84 | 104.41 |
| Black Myth: Wukong | 96.00 | 64.00 | 72.00 | 57.00 | 25.00 |
| Black Myth: Wukong + Frame Generation | 151.00 | 109.00 | 114.00 | 89.00 | 50.00 |
Jumping over to our 1440p and 1080p benchmarks, the story stays consistent, with the RTX 5080 delivering around a 10 per cent uptick in performance when compared to its last-generation counterpart. Games such as Returnal, Shadow of the Tomb Raider, Watch Dogs: Legion, Assassin’s Creed Valhalla and even Black Myth: Wukong all deliver well over 60 FPS on both 1080p and 1440p—at least until you throw some significant ray tracing into the mix. This is where the limitations of the RTX 5080 come to the surface. While it can deliver buttery-smooth frame rates with most modern AAA games, ray tracing is taxing on even the most powerful GPUs, and it can drastically reduce frame rates and playability.
This is where DLSS comes into the mix and can take a game that was only running at 50-60 FPS and push it into a much smoother, more playable experience. Once again, looking at Cyberpunk 2077, the game was very playable at 1080p even with ray tracing, but with DLSS 4, the overall frame rate was significantly smoother, potentially ideal for a 1080p or even 1440p experience with this GPU. It is when you jump over to 4K that it starts to really make sense, and you can take an experience that struggles to run and make it very playable.
| 1080P Benchmark | RTX 5090 | RTX 5080 | RTX 4090 | RTX 4080 | RX 7900 XTX |
|---|---|---|---|---|---|
| Shadow of the Tomb Raider | 319.00 | 259.00 | 264.00 | 218.00 | 200.00 |
| Shadow of the Tomb Raider + Frame Generation | 342.00 | 319.00 | 304.00 | 266.00 | – |
| Returnal | 242.00 | 193.00 | 208.00 | 181.00 | 199.00 |
| Assassin’s Creed Valhalla | 223.00 | 211.00 | 244.00 | 224.00 | 209.00 |
| Watch Dogs Legion | 165.00 | 127.00 | 152.00 | 147.00 | 112.00 |
| Cyberpunk 2077 (Ultra) | 249.25 | 219.00 | 235.53 | 200.88 | 194.58 |
| Cyberpunk 2077 RT | 150.05 | 134.00 | 141.50 | 99.06 | 65.53 |
| Cyberpunk 2077 RT + Frame Generation | 562.61 | 494.70 | 316.64 | 172.50 | 100.78 |
| F1 24 (Ultra Preset) | 326.00 | 243.00 | 286.00 | 220.00 | 172.00 |
| F1 24 + Frame Generation | 445.00 | 331.00 | 376.00 | 308.00 | 318.00 |
| Metro Exodus (Extreme) | 188.57 | 131.13 | 163.56 | 90.77 | 136.04 |
| Black Myth: Wukong | 119.00 | 84.00 | 96.00 | 79.00 | 25.00 |
| Black Myth: Wukong + Frame Generation | 183.00 | 131.00 | 143.00 | 117.00 | 49.00 |
It is worth noting that while only a selection of games were included in our benchmark charts, Alan Wake 2 and Hogwarts Legacy were tested to assess how DLSS 4 performs in those titles, and the results were consistent. The visuals were smooth, clear, and free of major issues with fidelity or artifacts. However, as with any technology, it can be hit or miss, and DLSS 4 frame generation is not a universal solution that will improve every game.
That said, it is incredibly powerful and results in a much smoother experience, particularly with the new transformer model. Still, because it generates so much content, it has the potential to introduce minor issues with visual fidelity. NVIDIA, however, has done an excellent job training the models to deliver reliable results, and all the tests we conducted supported that conclusion.
Jumping over to content creation, the RTX 5080 carries over much of what we saw in the RTX 5090, such as fast rendering times and access to NVIDIA’s range of tools designed for GPUs that enhance content creation. Software like NVIDIA Broadcast leverages the card’s capabilities to improve audio and visuals and even add effects that can reduce the workload after recording. With a broad range of software supporting CUDA, NVIDIA maintains an edge—at least on Windows—over AMD in rendering and encoding tasks.
While the RTX 5090 features three Gen 9 encoders that speed up rendering times, the RTX 5080 is a close second with two Gen 9 encoders, enabling efficient and fast exports in software like DaVinci Resolve Studio.
As noted with the RTX 5090, one of the more significant advancements is the introduction of 4:2:2 chroma subsampling support—a feature that revolutionizes video production workflows. This capability allows creators to work with higher-quality footage without the typical processing bottlenecks, maintaining professional-grade colour accuracy throughout their projects. The GPU handles H.264, H.265, and Multi-view HEVC encoding and decoding tasks with impressive efficiency. Thanks to its built-in hardware decoders and encoders, the card can manage complex workflows with H.264 4:2:2 without bringing the system to a crawl, ensuring exports are completed in seconds rather than minutes—something any video editor will appreciate.
“The NVIDIA RTX 5080 Founders Edition GPU is a solid mainstream and enthusiast GPU that delivers on its promises.”
In testing with video provided by NVIDIA, along with footage we captured ourselves, the RTX 5080 performed exceptionally well in the timeline, with no significant slowdowns, even when working with complex video featuring chroma subsampling. While exporting wasn’t as fast as with the RTX 5090, the process remained smooth and delivered clean, polished footage ready for YouTube, distribution, or team collaboration. The difference in export times between the RTX 5090 and RTX 5080 was minor, shifting from approximately 13 seconds on the flagship RTX 5090 to around 15 seconds on the RTX 5080.
In terms of power consumption, the RTX 5080 Founders Edition remained consistent with expectations for the 40-series lineup. The card reached a peak power draw of 293.28W during testing, with its highest usage occurring during our Cyberpunk 2077 ray tracing benchmarks. For most other tests conducted during the review period, it rarely exceeded 275W. This power usage aligns with what we observed during the RTX 4080’s launch testing and even drew less power than the AMD Radeon RX 7900 XTX—a GPU that fell short of any RTX 5080 benchmark results, even before factoring in DLSS.
The NVIDIA RTX 5080 Founders Edition GPU is a solid mainstream and enthusiast GPU that delivers on its promises. It is incredibly powerful, boasts features that benefit both gamers and content creators, and comes within striking distance of the last-generation flagship, the RTX 4090. I would have liked to see a bigger generational leap in raw performance compared to the RTX 4080, but with the help of DLSS 4, it bridges that gap enough that there is rarely a workload it will struggle with—even considering it has roughly half the cores of its bigger brother.
At $999, NVIDIA has positioned the RTX 5080 Founder’s Edition GPU at a performance level that makes it a worthwhile investment, especially for those upgrading from a 30-series card. The new encoders and decoders, improved codec support, and overall strong performance across a range of games—even at maxed-out visual settings—make it an easy GPU to recommend. While it may be overkill for those focused on 1080p or 1440p gaming, for anyone wanting the best in 4K AAA gaming or looking to enter the world of content creation, this is the ideal choice. It delivers where it matters most without draining your life savings to install it in your rig.
