Compare · 2 products compared
NVIDIA GeForce RTX 5070 TivsNVIDIA GeForce RTX 4080 Super
These two graphics cards both offer 16 GB of video memory and require a 750 W power supply. Choosing between them comes down to balancing launch pricing against raw shader power and newer interface technology.
eBay prices refresh hourly — check the listing before you buy. We may earn a commission.
- MOST WINS
NVIDIABest for Buyers seeking newer display outputs and lower starting costs
9of 13 specs better
NVIDIABest for Users prioritizing raw compute power and shader cores
4of 13 specs better
+ Add a product to compare
Biggest differences
Launch price
NVIDIA GeForce RTX 5070 Ti leads
−25%
lower is better
- NVIDIA GeForce RTX 5070 Ti749 USD
- NVIDIA GeForce RTX 4080 Super999 USD
Memory bandwidth
NVIDIA GeForce RTX 5070 Ti leads
+22%
higher is better
- NVIDIA GeForce RTX 5070 Ti896 GB/s
- NVIDIA GeForce RTX 4080 Super736 GB/s
Last-level cache
NVIDIA GeForce RTX 4080 Super leads
+33%
higher is better
- NVIDIA GeForce RTX 5070 Ti48 MB
- NVIDIA GeForce RTX 4080 Super64 MB
Shader cores
NVIDIA GeForce RTX 4080 Super leads
+14%
higher is better
- NVIDIA GeForce RTX 5070 Ti8,960 cores
- NVIDIA GeForce RTX 4080 Super10,240 cores
Where each one wins
NVIDIA GeForce RTX 5070 Ti
- Launch price
- Base clock
- Memory type
- Memory bandwidth
- Memory speed
- Board power
- PCIe version
- HDMI version
- DisplayPort version
NVIDIA GeForce RTX 4080 Super
- Shader cores
- Boost clock
- FP32 compute
- Last-level cache
What both have
- Video memory16 GB
- Memory bus width256 bit
- Recommended PSU750 W
- Power connectors12V-2x6
- HDMI outputs1 ports
- DisplayPort outputs3 ports
- Max displays4 displays
- Video encoders2 encoders
Every spec
Red marks the better value in each row.
| Spec | NVIDIA GeForce RTX 5070 Ti | NVIDIA GeForce RTX 4080 Super |
|---|---|---|
| Overview | ||
| Release date | 2025-02 | 2024-01 |
| Launch price | 749 USD1 | 999 USD |
| Graphics chip | ||
| Architecture | Blackwell | Ada lovelace |
| Shader cores | 8,960 cores | 10,240 cores |
| Boost clock | 2,452 MHz | 2,550 MHz |
| Base clock | 2,300 MHz | 2,290 MHz |
| FP32 compute | 43.9 TFLOPS | 52 TFLOPS |
| Memory | ||
| Video memory | 16 GB | 16 GB |
| Memory type | Gddr7 | Gddr6x |
| Memory bus width | 256 bit | 256 bit |
| Memory bandwidth | 896 GB/s | 736 GB/s |
| Memory speed | 28 Gbps | 23 Gbps |
| Last-level cache | 48 MB | 64 MB |
| Power | ||
| Board power | 300 W | 320 W |
| Recommended PSU | 750 W | 750 W2 |
| Power connectors | 12V-2x63 | 12V-2x64 |
| Interface | ||
| PCIe version | 5.0 | 4.0 |
| PCIe lanes | 16 lanes | Not published |
| Display outputs | ||
| HDMI version | 2.1b | 2.1a |
| DisplayPort version | 2.1b5 | 1.4a |
| 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 | AV1 |
| Video encoders | 2 encoders | 2 encoders |
| Size | ||
| Card length | Not published | 304 mm |
| Card height | Not published | 137 mm |
| Slot width | Not published | 3 slots |
Notes on 5 values — test conditions and caveats
- 1NVIDIA GeForce RTX 5070 Ti, Launch price: NVIDIA sells no Founders Edition of this card; $749 is the starting MSRP for partner cards.
- 2NVIDIA GeForce RTX 4080 Super, Recommended PSU: NVIDIA's 750 W minimum assumes a PC with a Ryzen 9 5900X; needs vary by configuration.
- 3NVIDIA GeForce RTX 5070 Ti, Power connectors: One 16-pin 12V-2x6 socket; an adapter for two 8-pin cables comes in the box.
- 4NVIDIA GeForce RTX 4080 Super, Power connectors: Card takes a 16-pin 12V-2x6 connector; a 3x PCIe 8-pin adapter ships in the box for PSUs without a PCIe Gen 5 cable.
- 5NVIDIA GeForce RTX 5070 Ti, DisplayPort version: DisplayPort 2.1b at the full UHBR20 rate.
Performance and Architecture
The RTX 4080 Super relies on the Ada Lovelace architecture to deliver 52 TFLOPS of FP32 compute performance. It features 10,240 shader cores and a 2,550 MHz boost clock to drive demanding games.
In contrast, the RTX 5070 Ti uses the newer Blackwell architecture with 8,960 cores and reaches 43.9 TFLOPS. While it trails in raw compute throughput, it introduces a base clock of 2,300 MHz and GDDR7 memory running at 28 Gbps for an 896 GB/s bandwidth advantage.
Price and Power
The RTX 5070 Ti carries a starting MSRP of $749 for partner cards, though most sold above that at launch. It requires a 300 W board power and connects via a single 16-pin 12V-2x6 socket.
The older RTX 4080 Super launched at $999 and draws 320 W of board power. Both cards use the same minimum 750 W power supply recommendation, making either option compatible with existing high-end power setups.
Connectivity and Video
For display outputs, the RTX 5070 Ti upgrades to PCIe 5.0, HDMI 2.1b, and DisplayPort 2.1b at the full UHBR20 rate. It also expands hardware encode support to include H.264, HEVC, and AV1 across its two encoders.
The RTX 4080 Super uses PCIe 4.0 alongside HDMI 2.1a and DisplayPort 1.4a connections. Its video engine supports AV1 hardware encoding across two encoders.
The Verdict
Buy the NVIDIA GeForce RTX 5070 Ti if you want modern PCIe 5.0 support, faster GDDR7 memory bandwidth, and a lower starting price. Buy the NVIDIA GeForce RTX 4080 Super if you prefer higher raw FP32 compute performance and shader core counts.
Where to buy
NVIDIA GeForce RTX 5070 Ti
Best for Buyers seeking newer display outputs and lower starting costs
NVIDIA GeForce RTX 4080 Super
Best for Users prioritizing raw compute power and shader cores
FAQ
Do both cards require the same power supply?
Yes, both graphics cards recommend a 750 W power supply, though NVIDIA notes this minimum assumes a system with a Ryzen 9 5900X.
What video memory do they use?
Both cards feature 16 GB of video memory, but the RTX 5070 Ti upgrades to GDDR7 memory while the RTX 4080 Super uses GDDR6X.
Which card has newer display ports?
The RTX 5070 Ti includes DisplayPort 2.1b at the full UHBR20 rate and HDMI 2.1b, whereas the RTX 4080 Super uses DisplayPort 1.4a and HDMI 2.1a.
Spotted something wrong? Tap the next to any section or spec group to let us know. Every report is read.





