GTX 1080 Ti vs RTX 2080
The complete head-to-head: full specs, real per-game FPS with 1% lows, ray tracing, a live CPU-bottleneck check, thermals, power and value, with a straight verdict.
GeForce GTX 1080 Ti
GeForce RTX 2080
avg, 1080p
The short answer
This is the closest matchup in this whole comparison series, and it's close for a reason: the GTX 1080 Ti and RTX 2080 launched at the same $699 price about 18 months apart, and NVIDIA never fully resolved that overlap until the RTX 2080 Super arrived a year later. The RTX 2080 is about 6% faster on average and adds genuine, if first-generation, ray tracing and DLSS support. The GTX 1080 Ti answers with 3 GB more VRAM, a wider 352-bit bus, and, surprisingly, more raw bandwidth despite being the older card. If ray tracing doesn't matter to you, this is close enough to buy on price alone.
Few matchups on this site are as genuinely close as this one. The GTX 1080 Ti was Pascal's massive, 11 GB flagship, and the RTX 2080 arrived as Turing's near-flagship a generation later at the exact same launch price. Reviewers argued for years about which was the better buy, and the honest answer has always been: it depends whether you value raw raster power and VRAM, or ray tracing and DLSS.
The full spec sheet is below along with 1% lows game by game, how the 2080's first-generation RT cores and DLSS actually perform since the 1080 Ti has neither, thermals, power draw, price history, and a verdict that doesn't dodge how close this one really is.
At a glance
Seven dimensions, two cards. Bigger area is not automatically better, each card owns different corners.
Strengths radar
Each axis scored 0-100 relative to the pair
Performance vs price
Up and to the left is better value
Grey dots are neighbouring cards for context.
Which should you buy?
Pick your use case, the winner changes with what you actually do.
A 6% average edge plus lower power draw make the 2080 the slightly better pick for a fresh build.
11 GB on a 352-bit bus gives the 1080 Ti real headroom at 1440p and in heavy texture packs where the 2080's 8 GB is comparatively tight.
The 1080 Ti has zero Tensor cores. Even the 2080's first-generation ones make basic local AI work possible at all.
Turing's 7th-gen NVENC gives noticeably cleaner HEVC output than Pascal's, even though neither card has AV1.
215 W against 250 W is a real if modest gap, and the 2080 runs a touch cooler for it.
484 GB/s over a 352-bit bus actually out-bandwidths the newer 2080's 448 GB/s, a rare case of the older card winning outright.
Already own a GTX 1080 Ti? Skip the stock RTX 2080.
Short answer: probably not
A 6% average gain is not worth trading away 3 GB of VRAM and a wider, faster memory bus. The 2080's real advantages are ray tracing and DLSS, both first-generation and both far better implemented on newer cards, not raw speed.
If you want ray tracing and DLSS without giving up ground elsewhere, look at a used RTX 3060 Ti or RTX 3070 instead, both a genuine step up in every category rather than a sideways trade. If you don't care about ray tracing at all, there's little reason to move off the 1080 Ti yet.
Full specification comparison
Every spec side by side. The highlighted figure wins each row.
| Specification | GTX 1080 Ti | RTX 2080 |
|---|---|---|
| Architecture | ||
| GPU / architecture | GP102 · Pascal | TU104 · Turing |
| Process node | TSMC 16 nm FinFET+ | TSMC 12 nm ▲ |
| Release date | Mar 10, 2017 | Sep 20, 2018 |
| Launch MSRP | $699 | $699 |
| Compute | ||
| CUDA cores | 3,584 ▲ | 2,944 |
| RT cores | None (pre-RTX) | 46 · 1st gen ▲ |
| Tensor cores | None (pre-RTX) | 368 · 1st gen ▲ |
| Base / boost clock | 1481 / 1582 MHz | 1515 / 1800 MHz ▲ |
| L2 cache | 2.75 MB | 4 MB ▲ |
| Memory | ||
| VRAM | 11 GB GDDR5X ▲ | 8 GB GDDR6 |
| Memory bus | 352-bit ▲ | 256-bit |
| Bandwidth | 484 GB/s ▲ | 448 GB/s |
| Power & form factor | ||
| TDP | 250 W | 215 W ▲ |
| Power connector | 1× 8-pin + 1× 6-pin | 1× 8-pin + 1× 6-pin |
| Suggested PSU | 650 W | 650 W |
| PCIe interface | 3.0 ×16 | 3.0 ×16 |
| Features | ||
| Upscaling | None (pre-DLSS) | DLSS 2 / Super Res ▲ |
| AV1 decode / encode | No / No | No / No |
| Display outputs | 3×DP1.4 · HDMI 2.0b · DVI-D | 3×DP1.4 · HDMI 2.0b · USB-C (VirtualLink) |
Average FPS by game
Switch resolution, metric, ray tracing and upscaling, the bars update live. Hover a bar for the avg / 1% / 0.1% breakdown.
Average FPS · 1080p · Rasterization. Representative data, hover a bar for avg / 1% / 0.1%. Note the RTX 2080's 1% lows fall off at 1440p and up from its 8 GB VRAM.
Smoothness & frametimes
Average FPS hides stutter. This is one run plotted as frametime, a flat line is smooth, spikes are hitches. Watch what the RTX 2080's 8 GB VRAM does in a texture-heavy game.
Frametime: Alan Wake 2, 1440p Ultra
Lower and flatter is better · 16.7 ms = 60 fps
Counterintuitive but real: the slower GTX 1080 Ti delivers a smoother line here because 12 GB holds the texture pool the 8 GB RTX 2080 keeps evicting, those spikes are VRAM stutter, not raw speed.
Lab scores
Clean apples-to-apples tests, useful, but weaker at predicting real games than the FPS above.
Ray tracing & upscaling
Ray tracing tanks both cards at native resolution. What makes it playable is upscaling, and here the RTX 2080 has a real edge.
Is frame generation "real" performance?
Neither card supports frame generation, that arrived with DLSS 3 on the RTX 40 series. The RTX 2080 is limited to DLSS 2 Super Resolution upscaling, which raises frame rate by rendering at a lower internal resolution and reconstructing detail with its Tensor cores. The GTX 1080 Ti has no Tensor cores and no DLSS support of any kind.
The 1080 Ti can use AMD's FSR 1 or FSR 2 in supported games since it isn't hardware-locked, a decent fallback that narrows the gap in modern titles that ship both options. It won't match DLSS 2's image quality, but it's a real, usable upscaler where the 1080 Ti would otherwise have nothing.
Will your CPU bottleneck these cards?
A faster GPU only helps if your CPU keeps up. Pick your CPU and resolution to see expected FPS and bottleneck for each card.
CPU + GPU pairing check
The willitbottleneck engine, built right into the comparison.
Estimated from a 43-CPU model across 1440p. A weak CPU mostly bites at 1080p; at higher resolutions the GPU becomes the limiter and the bottleneck shrinks.
Thermals & noise
The efficiency gap shows up here: the newer card runs cooler and quieter, day in and day out.
Real draw, efficiency & PSU
TDP is the rated figure; what matters is real gaming draw and how many frames you get per watt.
Both use a single 8-pin connector. If your PSU already ran the GTX 1080 Ti, it will run the RTX 2080 easily. Leave about 30% headroom for spikes.
Will it fit, and what to pair it with?
Size to scale
Card length matters for small cases
What monitor to pair
GTX 1080 Ti: strong at 1440p 100-144 Hz and workable at 4K 60 fps in lighter or older titles, thanks to its wide bus and 11 GB. RTX 2080: similar 1440p strength plus DLSS 2 stretching a handful of supported titles further toward 4K. Neither is a comfortable native 4K card by 2026 AAA standards.
Running an older PCIe 3.0 board?
Both cards run a full PCIe 3.0 ×16 link, RTX 20-series never moved to PCIe 4.0. Neither loses anything on an older PCIe 3.0 motherboard, and a modern PCIe 4.0 or 5.0 board offers no measurable gain to either since neither card can use the extra bandwidth.
Price history & dollars-per-frame
Raw price is only half the story. Cost per frame is the number that decides it, and it keeps moving as prices fall.
New price over time
Street price of new cards (USD)
The 11 GB vs 8 GB question
GTX 1080 Ti · 11 GB
- 11 GB was massive for 2017 and remains plenty for 1080p and 1440p, even some lighter 4K, in 2026.
- The 352-bit bus and 484 GB/s of bandwidth are the widest memory subsystem in this whole comparison, out-bandwidthing the newer RTX 2080.
- That bandwidth advantage carries directly into emulation and other memory-bound workloads.
- No RT or Tensor cores at all, so ray tracing and Tensor-accelerated AI work aren't just slower, they're unavailable.
RTX 2080 · 8 GB
- First-generation RT and Tensor cores make ray tracing and basic local AI work possible at all, something the 1080 Ti cannot do regardless of settings.
- DLSS 2 support lets it stretch its 8 GB and frame rate further at 1440p in supported titles.
- 8 GB is the tighter number of the two by 3 GB, and can show up first in texture-heavy 2026 AAA titles.
- A narrower 256-bit bus and lower 448 GB/s bandwidth than the older 1080 Ti, an unusual case of the newer card having the smaller memory subsystem.
Emulation & memory bandwidth
The narrower bus is a real downgrade for emulation
This is one of the rare cases where the older card wins the memory subsystem outright. The GTX 1080 Ti's 484 GB/s over a 352-bit bus beats the RTX 2080's 448 GB/s over 256-bit, and the 1080 Ti's 3 GB VRAM advantage adds to it. For high-resolution Switch and RPCS3 emulation, the 1080 Ti is genuinely the stronger pick.
Creator, AI & streaming
The newer card is quicker per task, but the older card's extra VRAM unlocks larger jobs it can't fit at all.
Streaming: the AV1 encoder is the differentiator
Neither card has AV1 support in either direction, that arrived with Ampere. What the RTX 2080 does bring is Turing's 7th-generation NVENC, with proper HEVC B-frame support that gives noticeably cleaner output than Pascal's 6th-gen encoder at the same bitrate. For anyone streaming or recording regularly, that generational encoder gap is the real reason to lean toward the 2080.
Overclocking & undervolting
Both are strong overclockers for their era. The GTX 1080 Ti in particular has a well-earned reputation for headroom, its GDDR5X memory often takes a real bump too. The RTX 2080 has less room left on the core but gains more from undervolting given its higher 215 W TDP relative to how RT-heavy workloads load it.
Pros & cons
GTX 1080 Ti
- 11 GB of VRAM and the widest memory bus in this whole comparison series, out-bandwidths the newer 2080
- Genuinely close raw performance despite being a generation older, wins raster-heavy, non-RT titles outright
- Cheaper new-old-stock and used, with the best raw $/frame here
- No RT cores, no Tensor cores, no DLSS at all, and no AV1 support
- 250 W draw on a reference blower design that runs hot and loud under load
RTX 2080
- About 6% faster on average, plus genuine, if first-generation, ray tracing and DLSS 2 support
- 215 W draw and a newer 7th-gen NVENC encoder with real HEVC quality gains over Pascal
- Tensor cores make basic local AI work possible at all
- Only 8 GB of VRAM on a narrower, lower-bandwidth bus than the older 1080 Ti
- No AV1 decode or encode either, that arrived with Ampere, so this isn't a clean streaming win
Buying pitfalls
Reference blower coolers run hot and loud on both
Founders Edition and many partner cards for both the 1080 Ti and 2080 used blower-style coolers that get genuinely loud under sustained load. If you're buying used, an aftermarket dual or triple-fan model from either generation runs noticeably cooler and quieter than a reference blower.
Don't confuse the RTX 2080 with the 2080 Super or 2080 Ti
NVIDIA released a faster RTX 2080 Super about a year later, and the RTX 2080 Ti is a different, much more expensive flagship. This page covers only the original, non-Super RTX 2080, confirm the exact model before buying if the seller's listing is vague.
GTX 1080 Ti vs RTX 2080: your questions answered
Only slightly, and only in some ways. It's about 6% faster on average and adds real ray tracing and DLSS support. The GTX 1080 Ti answers with 3 GB more VRAM and, surprisingly, more memory bandwidth despite being the older card. This is one of the closest matchups on this site.
If ray tracing and DLSS matter to you, or you value lower power draw, buy the RTX 2080. If you want the most VRAM and bandwidth for the least money and don't care about ray tracing, the GTX 1080 Ti is arguably the better value today.
Both remain capable 1440p cards, and either handles 1080p comfortably. Their real limitation by 2026 is DLSS 2 at best (or none at all), no DLSS 3 Frame Generation, and RT performance that's a fraction of a current mid-range card.
Not really. A 6% average gain isn't worth losing 3 GB of VRAM and bandwidth. Upgrade to the 2080 only if ray tracing or DLSS specifically matter to you, otherwise a used RTX 3060 Ti or RTX 3070 is a far more meaningful jump for similar money.
The RTX 2080. The GTX 1080 Ti has zero Tensor cores, so all AI work runs on CUDA alone. The 2080's first-generation Tensor cores, while modest by current standards, make basic Stable Diffusion and rendering acceleration possible where the 1080 Ti simply can't do it at all.
The RTX 2080, for its newer 7th-gen NVENC encoder with proper HEVC quality. Neither card supports AV1 in any form, so if AV1 specifically matters, you need a newer card than either of these.
No. It launched before RTX and has no dedicated RT hardware whatsoever. Ray tracing is not a realistic option on the 1080 Ti in any modern game.
No. The RTX 2080 supports DLSS 2 Super Resolution upscaling only. DLSS 3 Frame Generation is exclusive to the RTX 40 series and newer. The GTX 1080 Ti has no DLSS support at all, though it can use AMD's FSR since that isn't hardware-locked.
At 1080p and 1440p a weak CPU can hold back either card, more so the slightly faster 2080. Use the pairing checker on this page, a Ryzen 5 3600 or Core i5-10400F and up keeps both cards reasonably fed.
The RTX 2080 draws less, 215 W against the 1080 Ti's 250 W, and runs a few degrees cooler on comparable coolers. Both were commonly sold with loud reference blower designs, aftermarket cards from either generation run noticeably quieter.
A quality 650 W unit is the standard recommendation for either card, each using an 8-pin plus 6-pin connector. A 650-750 W supply gives comfortable headroom for either.
Founders Edition versions of both cards run around 267 mm and 2-slot, check your case's maximum GPU length. Both use a full PCIe 3.0 ×16 link, RTX 20-series never adopted PCIe 4.0, so a modern motherboard offers no extra bandwidth to either card.
There is a mobile RTX 2080 that runs at lower power and clocks than this desktop card. There was no direct mobile GTX 1080 Ti, the closest mobile Pascal part was the GTX 1080 notebook GPU, itself a different, lower-power chip. Never compare laptop and desktop numbers directly.
Both handle 4K reasonably in lighter or older titles, the 1080 Ti's 11 GB and wide bus help there, and the 2080's DLSS 2 stretches a handful of supported titles further. Neither is a comfortable 4K card for demanding modern AAA settings.
Both are long out of production, so this is a used-market comparison either way. A used GTX 1080 Ti around $105 is the stronger raw value; a used RTX 2080 around $125 is worth the premium if you want any ray tracing or DLSS capability at all. Buy from a seller with a return window and ask about mining history on either card.
Both are on the older side. NVIDIA still ships driver updates covering Pascal and Turing, but new features and optimizations increasingly target Ampere and newer. Neither is at risk of losing basic driver support soon, but don't expect meaningful new feature support on either going forward.