The spec sheet: same memory, different muscle
Nvidia announced the RTX 5080 at a $999 launch MSRP and the RTX 5070 Ti at $749. That is a $250 gap, or about 33% of the 5070 Ti's price. These were "starting at" prices. Partner cards often sold above them.
| Spec | RTX 5070 Ti | RTX 5080 |
|---|---|---|
| Launch MSRP | $749 | $999 |
| CUDA cores | 8,960 | 10,752 |
| Boost clock | 2.45 GHz | 2.62 GHz |
| Memory | 16GB GDDR7 | 16GB GDDR7 |
| Memory bandwidth | 896 GB/s | 960 GB/s |
| Total graphics power | 300W | 360W |
| Nvidia minimum PSU (reference system) | 750W | 850W |
The 5080 has about 20% more CUDA cores. That does not translate into 20% more frames. Performance depends on the game, the settings and the test rig. The 5080 has no VRAM-capacity advantage over the 5070 Ti, so any gap comes from compute and bandwidth. The wattage and PSU figures are Nvidia's rated numbers. They are not measurements of power draw during gameplay, and they do not show which card is more efficient.
Gaming: a real lead, but a modest one
TechRepublic cites TechPowerUp testing of one demanding scenario. Without upscaling, the 5080 reached 35 FPS and the 5070 Ti reached 31 FPS, a gap of roughly 13%. With DLSS Transformer Quality enabled, it was 62 FPS against 53 FPS, a gap of roughly 17%. I could not independently confirm these exact figures in the underlying test. Treat them as one tested scenario, not a universal average.
TechRepublic also notes that the 5070 Ti in that test was a factory-overclocked model. That narrows the real-world gap a little and is a reminder that retail cards vary.
Multi Frame Generation needs the most care:
- With 4x Multi Frame Generation, the reported displayed rates were 120 FPS on the 5080 and 108 FPS on the 5070 Ti.
- Reported latency rose from 55 ms to 74 ms on the 5080 and from 62 ms to 86 ms on the 5070 Ti.
- Smoother-looking motion is not the same as more responsive input. For competitive play, conventional frame rate and latency matter more than the displayed number.
The 5080's extra resources help most when the GPU is the bottleneck, such as 4K with heavy ray tracing. At 1440p, or at 4K with sensible settings, the 5070 Ti stays capable. Check benchmarks for the games you actually play, because the ray-tracing gap varies by title.
Creator workloads: it depends on the app
Puget Systems' 5070 Ti content-creation review found the card 8.5% ahead of the RTX 4070 Ti Super in Premiere Pro, nearly matching the 5080. Puget's own conclusion was that the 5070 Ti is generally more compelling for Premiere Pro than the 5080. The caveat is that the extra decoder on the 5080 could matter for LongGOP footage.
That hardware difference is real. Puget's later RTX 50-series video-editing piece says the 5080 and 5090 carry two 6th-gen NVDEC decoders, while the 5070 Ti and 5070 carry one. Puget suggests the single-decoder cards suit simpler timelines. Multi-camera sequences, or sequences with several video layers, benefit from the second decoder.
The 5080 also pulls ahead where GPU effects dominate. In DaVinci Resolve GPU Effects, Puget measured the 5080 14% ahead of the 5070 Ti. Its summary said that heavy GPU-effects users may find the 80-class card a better value.
One caution on codecs. TechRepublic repeats Puget's early finding that 10-bit 4:2:2 hardware acceleration was not enabled in Premiere Pro at the time. That reflected the software of that test, and Adobe has since moved toward support for RTX 50-series cards. Verify support in your current application and driver versions before assuming either way.
For local AI work, both cards top out at 16GB, which caps the size of models that fit entirely in VRAM. The 5080 has more compute, but a higher price only pays off if your own workload shows the gain.
Pricing: the part that changed
Retail prices have drifted far from MSRP, and the numbers depend heavily on the tracker, retailer and date:
- PC Tech Vault, a US new-retail tracker, showed a 90-day median of $1,150 for the 5070 Ti. The range was $901 to $1,350, and the newest capture was dated October 19, 2026. Its cheapest in-stock card was $1,190 at Best Buy.
- Pangoly's US average for the 5080 was $1,847, which it said was 5% above its 90-day average of $1,760.
- Videocardprices.com showed the 5080 at $1,848.43, 85% above MSRP, with a tracked low of $1,249.99 back in June.
- Digital Citizen's October roundup put the lowest US listings at $1,869 for the RTX 5080 and $1,189 for the RTX 5070 Ti.
TechRepublic cites a PC Tech Vault 90-day median of about $1,600 for the 5080. I could not verify that figure, and other trackers in this sample read higher. Retail listings on a single day can also be lower or higher than any median.
The trackers do not agree, and each measures something different: lowest listing, average or median. The same applies to the 5070 Ti. Compare two specific, in-stock cards from the same retailer on the same day.
Doing the premium math
The arithmetic is simple. Divide the price difference by the 5070 Ti's price:
- $1,200 versus $1,600 is a 33% premium, matching the original launch gap.
- $1,200 versus $1,400 is about 17%.
- Using the lowest-listing figures above, $1,189 versus $1,869 is a premium of about 57%. That is a hypothetical pairing of different sources, not a verified same-day quote.
Put that against the measured gaming lead of roughly 13–17% in TechRepublic's cited scenario. A 57% price jump for that margin is hard to justify unless 4K is your daily workload. If the gap narrows toward 15–20%, the 5080 starts to make sense.
Practical checks before buying
- Match the cards. Compare exact partner models at the same retailer and date.
- Check the PSU. Nvidia's reference minimums are 750W for the 5070 Ti and 850W for the 5080. Your CPU and the card's factory settings change the real requirement.
- Check clearance and connectors. Measure card length and thickness against your case, and confirm the power connector fits your supply.
- Verify software support. For editing, confirm that your application version and driver support the codecs you shoot.
- Read benchmarks for your own workload. Frame-generation figures and averages from other titles will mislead you.
Verdict
Based on the evidence here, the RTX 5070 Ti is the sensible default. It matches the 5080's VRAM, DLSS 4 support and encoder generation. It draws less rated power and, by the trackers above, costs far less. The 5080 earns its place for buyers who regularly play demanding games at 4K. It also suits editors who lean on multi-layer timelines or GPU-heavy effects, and anyone who finds one priced close to a 5070 Ti.
TechRepublic is a general publisher that may earn affiliate revenue, as its own disclosure notes. Its article does stay appropriately cautious about its price snapshots. Even so, retail prices move quickly. Treat any premium calculation as valid only for the day you check.