Compare · 2 products compared
NVIDIA GeForce RTX 5080vsNVIDIA GeForce RTX 5090: Spec Comparison
The NVIDIA GeForce RTX 5080 and NVIDIA GeForce RTX 5090 both run on the Blackwell architecture and use the PCIe 5.0 interface. Choosing between them depends entirely on whether you prioritize raw compute and massive memory or a lower price and reduced power draw.
eBay prices refresh hourly — check the listing before you buy. We may earn a commission.
Best for Value-focused gamers seeking high clock speeds and lower power consumption.
Wins6 / 13- MOST WINS
Best for Professionals and enthusiasts requiring maximum compute power and 32 GB of VRAM.
Wins7 / 13
Biggest differences
Shader cores
NVIDIA GeForce RTX 5090 leads
2.02×
higher is better
- NVIDIA GeForce RTX 508010,752 cores
- NVIDIA GeForce RTX 509021,760 cores
Launch price
NVIDIA GeForce RTX 5080 leads
2×
lower is better
- NVIDIA GeForce RTX 5080999 USD
- NVIDIA GeForce RTX 50901,999 USD
Video memory
NVIDIA GeForce RTX 5090 leads
2×
higher is better
- NVIDIA GeForce RTX 508016 GB
- NVIDIA GeForce RTX 509032 GB
Memory bus width
NVIDIA GeForce RTX 5090 leads
2×
higher is better
- NVIDIA GeForce RTX 5080256 bit
- NVIDIA GeForce RTX 5090512 bit
Where each one wins
NVIDIA GeForce RTX 5080
- Launch price
- Boost clock
- Base clock
- Memory speed
- Board power
- Recommended PSU
NVIDIA GeForce RTX 5090
- Shader cores
- FP32 compute
- Video memory
- Memory bus width
- Memory bandwidth
- Last-level cache
- Video encoders
What both have
- Release date2025-01
- ArchitectureBlackwell
- Memory typeGddr7
- Power connectors12V-2x6
- PCIe version5.0
- PCIe lanes16 lanes
- HDMI version2.1b
- DisplayPort version2.1b
- HDMI outputs1 ports
- DisplayPort outputs3 ports
- Max displays4 displays
- Hardware encodeH.264, HEVC, AV1
- Card length304 mm
- Card height137 mm
- Slot width2 slots
Every spec
Red marks the better value in each row.
| Spec | NVIDIA GeForce RTX 5080 | NVIDIA GeForce RTX 5090 |
|---|---|---|
| Overview | ||
| Release date | 2025-01 | 2025-01 |
| Launch price | 999 USD | 1,999 USD |
| Graphics chip | ||
| Architecture | Blackwell | Blackwell |
| Shader cores | 10,752 cores | 21,760 cores |
| Boost clock | 2,620 MHz | 2,410 MHz |
| Base clock | 2,300 MHz | 2,010 MHz |
| FP32 compute | 56.3 TFLOPS | 104.8 TFLOPS |
| Memory | ||
| Video memory | 16 GB | 32 GB |
| Memory type | Gddr7 | Gddr7 |
| Memory bus width | 256 bit | 512 bit |
| Memory bandwidth | 960 GB/s | 1,792 GB/s |
| Memory speed | 30 Gbps | 28 Gbps |
| Last-level cache | 64 MB | 96 MB |
| Power | ||
| Board power | 360 W | 575 W |
| Recommended PSU | 850 W1 | 1,000 W2 |
| Power connectors | 12V-2x6 | 12V-2x6 |
| Interface | ||
| PCIe version | 5.0 | 5.0 |
| PCIe lanes | 16 lanes | 16 lanes |
| Display outputs | ||
| HDMI version | 2.1b | 2.1b |
| DisplayPort version | 2.1b3 | 2.1b4 |
| HDMI outputs | 1 ports | 1 ports |
| DisplayPort outputs | 3 ports | 3 ports |
| Max displays | 4 displays | 4 displays |
| Video engine | ||
| Hardware encode | H.264, HEVC, AV1 | H.264, HEVC, AV1 |
| Video encoders | 2 encoders | 3 encoders |
| Size | ||
| Card length | 304 mm | 304 mm |
| Card height | 137 mm | 137 mm |
| Slot width | 2 slots5 | 2 slots6 |
Notes on 6 values — test conditions and caveats
- 1NVIDIA GeForce RTX 5080, Recommended PSU: NVIDIA's minimum for a PC with a Ryzen 9 9950X; it recommends a PCIe CEM 5.1 compliant PSU.
- 2NVIDIA GeForce RTX 5090, Recommended PSU: NVIDIA's minimum for a PC with a Ryzen 9 9950X; it recommends a PCIe CEM 5.1 compliant PSU.
- 3NVIDIA GeForce RTX 5080, DisplayPort version: DisplayPort 2.1b at UHBR20 on all three DP outputs.
- 4NVIDIA GeForce RTX 5090, DisplayPort version: DisplayPort 2.1b at UHBR20 on all three DP outputs.
- 5NVIDIA GeForce RTX 5080, Slot width: The card is 2-slot (40 mm thick), but NVIDIA's install guide asks for 3-slot (61 mm) case clearance.
- 6NVIDIA GeForce RTX 5090, Slot width: The card is 2-slot, but NVIDIA's install guide asks for 3-slot (61 mm) case clearance.
Performance and Compute
The NVIDIA GeForce RTX 5090 delivers significantly heavier processing power with 21,760 shader cores and 104.8 TFLOPS of FP32 compute. This massive core advantage allows it to handle extreme rendering loads and complex calculations much faster than the competition.
By contrast, the RTX 5080 features 10,752 shader cores and 56.3 TFLOPS, but it pushes ahead in clock speeds with a 2,620 MHz boost clock. This higher frequency helps it maintain swift frame rates in standard gaming workloads without demanding as much electrical overhead.
Memory and Video Engine
For high-resolution texture streaming and professional workflows, the RTX 5090 equips 32 GB of GDDR7 video memory across a 512-bit bus width. This configuration drives a memory bandwidth of 1,792 GB/s alongside a larger 96 MB last-level cache.
The RTX 5080 steps down to 16 GB of video memory and a 256-bit bus width, yielding a memory bandwidth of 960 GB/s and a 64 MB cache. While both cards use GDDR7 and achieve fast memory speeds, the RTX 5090 adds an extra encoder for a total of 3 hardware video encoders to aid heavy video pipelines.
Power and System Requirements
Power consumption creates a wide gap between these two graphics cards. The RTX 5090 demands a hefty 575 W board power and requires a 1,000 W power supply unit as NVIDIA's minimum for a Ryzen 9 9950X system.
The RTX 5080 draws a much tamer 360 W of board power, needing only an 850 W power supply under the same processor minimum. Both options utilize a 12V-2x6 power connector and comply with PCIe CEM 5.1 guidelines.
Price and Dimensions
Cost is a major differentiator, as the RTX 5080 launches at $999 while the RTX 5090 doubles that investment at $1,999. Buyers must weigh whether the substantial performance upgrade justifies paying twice the entry price.
Despite their different power profiles, both graphics cards share identical physical dimensions of 304 mm in length and 137 mm in height. Each card is built as a 2-slot unit measuring 40 mm thick, though NVIDIA's installation guide asks for a 3-slot case clearance of 61 mm for both models.
The Verdict
Buy the RTX 5080 if you want top-tier gaming performance at half the price while keeping power requirements manageable. Buy the RTX 5090 if your workflow demands maximum compute power, a wider memory bus, and 32 GB of VRAM regardless of cost.
Where to buy
NVIDIA GeForce RTX 5080
Best for Value-focused gamers seeking high clock speeds and lower power consumption.
NVIDIA GeForce RTX 5090
Best for Professionals and enthusiasts requiring maximum compute power and 32 GB of VRAM.
FAQ
What video outputs do these graphics cards include?
Both models feature three DisplayPort 2.1b outputs running at UHBR20 and one HDMI 2.1b port, supporting up to 4 displays total.
Do the RTX 5080 and RTX 5090 support the same hardware video encoding formats?
Yes, both graphics cards support H.264, HEVC, and AV1 hardware encoding, though the RTX 5090 includes three encoders compared to two on the RTX 5080.
What PCIe version and lane count do they use?
Both graphics cards operate on PCIe 5.0 and utilize 16 lanes.
Spotted something wrong? Tap the next to any section or spec group to let us know. Every report is read.
