RTX 2080 vs RTX 5070: Is the Upgrade Worth It?

If you still game on an RTX 2080, the RTX 5070 is a worthwhile upgrade for 1440p players and anyone running into the 8 GB VRAM wall: it is roughly 80-90% faster in modern games, adds 4 GB of memory, and unlocks DLSS 4 Multi Frame Generation, while 1080p players on a tight budget can reasonably hold out a little longer. The jump is big enough that it feels like a new tier of card rather than a refresh, and it comes at a lower launch price than the RTX 2080 carried in 2018.

The RTX 2080 arrived in September 2018 as the first wave of RTX cards, introducing hardware ray tracing and the original DLSS. Seven years later, the RTX 5070 sits in the middle of the Blackwell lineup, and the gap between them is a good illustration of how much has changed: more than twice the CUDA cores, twelve times the L2 cache, a newer memory standard, and a far more capable upscaling and frame generation stack.

msi Gaming GeForce RTX 2080 8GB GDRR6 256-bit HDMI/DP/USB Ray Tracing Turing Architecture Graphics Card (RTX 2080 Ventus 8G)

Included for 'RTX 2080 vs RTX 5070: Is the Upgrade Worth It?' as a relevant option in this category; details come from the product listing.

View on Amazon →

ASUS Prime GeForce RTX 5070 12GB GDDR7 OC Edition Gaming Graphics Card | SFF-Ready, PCIe 5.0, HDMI/DP 2.1, 2.5-slot, Axial-tech Fans, Dual BIOS

Included for 'RTX 2080 vs RTX 5070: Is the Upgrade Worth It?' because the listing specifies SFF-Ready and PCIe 5.0, details this guide uses to compare options.

View on Amazon →

PNY NVIDIA GeForce RTX™ 5070 EPIC-X RGB® OC Triple-Fan Graphics Card | 12GB GDDR7, 192-bit, Boost: 2685 MHz, SFF-Ready, PCIe® 5.0, HDMI®/DP 2.1, 2.4-Slot, NVIDIA Blackwell Architecture, DLSS 5

Included for 'RTX 2080 vs RTX 5070: Is the Upgrade Worth It?' because the listing specifies 12GB GDDR7 and 192-bit, details this guide uses to compare options.

View on Amazon →

What Changes on Paper

The spec table tells most of the story. Turing on 12 nm versus Blackwell on 5 nm is a jump of three process generations, and it shows up in clock speed and core count at the same time.

Note that the memory bus actually narrows from 256-bit to 192-bit. That does not hurt the RTX 5070, because GDDR7 still delivers 50% more raw bandwidth (672 vs 448 GB/s) and the 48 MB L2 cache absorbs a large share of memory traffic before it ever reaches VRAM. The narrower bus is a cost decision, not a performance limitation, at the resolutions this card targets.

How Big Is the Performance Jump?

Published reviews rarely put a seven-year-old card on the same chart as a 2025 release, so the gap has to be stitched together from overlapping data. According to TechPowerUp’s RTX 5070 Founders Edition review, the RTX 5070 lands roughly level with the RTX 4070 Ti overall and about 56% ahead of the RTX 3070 at 1440p. The RTX 3070 itself was widely measured at around RTX 2080 Ti level, which in turn ran roughly 20-25% ahead of the RTX 2080 in launch-era reviews. Chaining those results puts the RTX 5070 at roughly 80-90% faster than the RTX 2080 in rasterized games at 1440p. Aggregate benchmark sites built on synthetic tests show a smaller lead of around 50%, but synthetic scores tend to understate the gap in newer, heavier games.

In practice that means a game that runs at around 50-60 FPS on an RTX 2080 at 1440p high settings can be expected to land somewhere around 90-110 FPS on an RTX 5070 with the same settings. These are approximate ranges that vary with game, settings, CPU and patch version, but they illustrate the scale of change.

Ray tracing and path tracing

Ray tracing widens the gap further. Turing’s first-generation RT cores were capable of light ray-traced effects but struggled with heavier workloads, and the 8 GB frame buffer often runs short once ray-traced reflections and high-resolution textures are combined. The RTX 5070 carries fourth-generation RT cores and enough cache and VRAM to run demanding ray tracing at 1440p with upscaling. Path-traced modes in titles like Cyberpunk 2077 or Alan Wake 2 are effectively out of reach on the RTX 2080 at playable frame rates, while the RTX 5070 can run them at 1440p when combined with DLSS upscaling and frame generation.

DLSS: what you keep and what you gain

One of the overlooked facts is that the RTX 2080 still benefits from part of DLSS 4. The transformer-based super resolution model rolled out to all RTX cards, so Turing owners get sharper upscaling than they had before. What the RTX 2080 does not get is any form of DLSS Frame Generation. The RTX 5070 supports DLSS 4 Multi Frame Generation, which can generate up to three frames per rendered frame on supported titles. It is best used when the base frame rate is already reasonable (roughly 50-60 FPS or more) to keep input latency acceptable, which the RTX 5070 can usually deliver at 1440p.

Where 8 GB Starts to Hurt

The RTX 2080’s biggest weakness in 2026 is not raw speed but memory. Several recent releases want more than 8 GB at 1440p with high textures, and a few push past it even at 1080p. When VRAM fills up, the result is stutter, texture pop-in or textures that silently drop to lower quality rather than a simple lower average frame rate.

A practical rule of thumb for VRAM by resolution in current AAA games:

• 1080p, high textures: 8 GB is workable, though the newest titles may require texture settings one step down.
• 1440p, high textures: 10-12 GB is the comfortable floor; 8 GB often forces compromises.
• 1440p with ray tracing and frame generation: 12 GB is adequate, as both features add memory overhead.
• 4K, high textures: 16 GB is the comfortable target; 12 GB works with settings tuning.

The RTX 5070’s 12 GB places it squarely in the 1440p comfort zone. It can play at 4K with DLSS, but 4K buyers who keep cards for many years may prefer a 16 GB model.

Value Math: Cost per Unit of Performance

Using launch MSRP and the approximate relative performance above, set the RTX 2080 as an index of 100 and the RTX 5070 at roughly 185 (the midpoint of the 80-90% range).

• RTX 2080: $699 / 100 = about $7.00 per performance point at launch.
• RTX 5070: $549 / 185 = about $2.97 per performance point at launch.

That works out to roughly 2.3 times more performance per launch dollar. Street prices move around, but the takeaway is that the RTX 5070 offers far better value than the RTX 2080 did when it was new.

Power, PSU and Connector Changes

The RTX 5070 is rated at 250 W, only 35 W more than the RTX 2080’s 215 W, so most systems that ran an RTX 2080 comfortably will handle the newer card. A simple sizing method is GPU TDP plus CPU power plus about 150 W of headroom for the motherboard, drives, fans and transient spikes:

• With a 125 W gaming CPU: 250 + 125 + 150 = 525 W, so a quality 600-650 W unit is a sensible target.
• With a high-draw CPU around 250 W: 250 + 250 + 150 = 650 W, so 750 W gives more comfortable margin.

The connector is the bigger practical change. The RTX 2080 uses one 6-pin and one 8-pin PCIe connector, while the RTX 5070 uses a single 16-pin (12V-2×6) connector. Newer ATX 3.x power supplies include a native cable. Older units can use the adapter typically bundled with the card, which combines standard 8-pin PCIe cables. Make sure the 16-pin plug is fully seated, and avoid sharp bends right at the connector.

Case fit and older platforms

At 245 mm for the reference design, the RTX 5070 is actually shorter than the 267 mm RTX 2080, and both are dual-slot, so case clearance is rarely an issue. Partner cards with large triple-fan coolers can run longer and thicker, so check measurements for the exact model.

The RTX 5070 is a PCIe 5.0 x16 card, but it runs fine in older PCIe 3.0 or 4.0 slots. Because it uses all 16 lanes, the bandwidth penalty on a PCIe 3.0 platform is small. The more important question for older RTX 2080 systems is the CPU: a four- or six-core processor from 2017-2018 can hold back the RTX 5070, especially at 1080p. If the CPU is that old, 1440p is the resolution where the upgrade shines most, since the GPU does more of the work there.

Is the Upgrade Worth It for Your Resolution?

What an RTX 2080 Is Worth Now

Selling the old card softens the upgrade cost. On the used market, an RTX 2080 usually goes for roughly $180-$250 depending on condition, cooler design and region, and blower or heavily used mining-era cards sit at the bottom of that range. For anyone buying used instead, an RTX 2080 remains a capable 1080p card, but its 8 GB buffer and lack of frame generation make it a short-term purchase rather than a long-term one.

The Bottom Line

The RTX 5070 nearly doubles the RTX 2080’s rasterized performance, adds 50% more VRAM and a modern feature set, uses only slightly more power and fits in a smaller space. For a 1440p player, it is one of the clearest upgrade paths in the current lineup. The only good reasons to wait are playing mostly lighter games at 1080p, or planning a full platform upgrade where a higher-tier 16 GB card makes more sense.

Frequently Asked Questions

Is the RTX 5070 twice as fast as the RTX 2080?

Not quite in every game, but it comes close. Based on published review data, the RTX 5070 is roughly 80-90% faster at 1440p in rasterized games, with the gap growing in ray-traced titles. With DLSS 4 Multi Frame Generation enabled, displayed frame rates can be several times higher, though that is not the same as native rendering performance.

Can my RTX 2080 power supply run an RTX 5070?

Usually yes, if it is a quality unit of 600-650 W or more. The RTX 5070 is rated at 250 W versus 215 W for the RTX 2080. You need either a native 16-pin cable or the bundled adapter that uses your existing 8-pin PCIe cables.

Will an RTX 5070 work in a PCIe 3.0 motherboard?

Yes. PCIe is backward compatible, and because the RTX 5070 uses a full x16 link, the performance loss on PCIe 3.0 is small. An older CPU is more likely to limit performance than the slot itself.

Does the RTX 2080 support DLSS 4?

Partly. RTX 20 series cards can use the DLSS 4 transformer super resolution model for improved upscaling quality. They cannot use DLSS Frame Generation or Multi Frame Generation, which require newer RTX generations.

Is 12 GB of VRAM enough for an upgrade from 8 GB?

For 1440p gaming, 12 GB is a solid amount in current titles, including with ray tracing and frame generation. It removes most of the texture and stutter issues 8 GB cards face. For 4K with maximum textures over many years, 16 GB offers more headroom.