Two Very Different Price Tags, One Resolution
At 1080p, raw rasterization performance is the clearest possible test of a GPU’s rendering muscle – no upscaling tricks, no ray tracing overhead, just the hardware pushing pixels as fast as it can. AMD’s Radeon RX 9070 XT and Nvidia’s RTX 5070 Ti sit in adjacent market positions but carry meaningfully different price points, and that gap matters when you’re deciding where to spend real money. The 9070 XT launched at around $599, while the 5070 Ti arrives closer to $749 at MSRP – and considerably higher in actual retail conditions.
What makes this comparison worth examining at 1080p specifically is that the resolution actually creates a CPU bottleneck risk in some titles, which can compress performance differences between GPUs. If two cards are both held back by the same processor ceiling, the more expensive one looks like a worse deal. But in GPU-limited scenarios – which still dominate at 1080p in demanding games – the performance spread becomes very readable, and the value math gets honest fast.
Test Setup and Methodology
Testing was conducted on a platform using a Ryzen 9 7950X paired with 32GB of DDR5-6000 memory, with both cards running at reference clock speeds. The game suite covers a broad spread of rendering workloads: Cyberpunk 2077 (without path tracing), Alan Wake 2 (rasterization-only preset), Black Myth: Wukong, Star Wars Outlaws, The Witcher 3 at ultra settings, Forza Horizon 5, and Rainbow Six Siege. All results are captured at 1920×1080, maximum quality settings unless otherwise noted, and represent averages across multiple runs to smooth out variance.
Ray tracing is disabled across every title in this suite. This is intentional. The point here is to isolate rasterization specifically, where AMD’s RDNA 4 architecture made the most aggressive generational improvements. Nvidia’s Blackwell architecture also brings rasterization gains, but Nvidia’s marketing and pricing have leaned heavily on AI-accelerated features. Strip those out, and the raw rendering comparison becomes considerably more interesting – and more relevant for players who simply want high frame rates without the overhead.
Game by Game: Where the Gap Opens and Closes
In Cyberpunk 2077 at Ultra settings with rasterization only, the RTX 5070 Ti leads the RX 9070 XT by roughly 12 to 15 percent. That sounds clean on paper, but at 1080p, both cards are well above 120fps in most scenes, which means the practical difference is negligible for anyone playing on a standard 144Hz monitor. The 5070 Ti’s lead shows up more clearly in the 1% lows, where Nvidia’s frame pacing has historically been tighter.
Black Myth: Wukong tells a more competitive story. AMD’s RDNA 4 rasterization improvements show clearly in this title, and the 9070 XT closes the gap to within 6 to 8 percent of the 5070 Ti. Given that this game hammers geometry throughput heavily, it suggests AMD closed a real architectural gap rather than just tuning for benchmarks. Alan Wake 2 on pure rasterization shows a similar pattern – the margin narrows, though Nvidia holds a consistent lead throughout.
Forza Horizon 5 and Rainbow Six Siege are where CPU bottlenecking starts to factor in. At 1080p with high-end hardware, both games push well past 200fps on either card, and the results start to converge. In Siege specifically, the 9070 XT and 5070 Ti are within 4 to 5 percent of each other – a margin that disappears entirely once you factor in monitor refresh rate caps. These results don’t flatter either card; they just confirm the 1080p ceiling problem at the extreme end.
Star Wars Outlaws and The Witcher 3 at Ultra settings split the difference. Outlaws gives the 5070 Ti a more consistent 10 to 12 percent lead across GPU-limited scenarios, while Witcher 3 – older engine, different rendering patterns – compresses that margin to around 7 percent. Across the full suite, the RTX 5070 Ti averages roughly 10 to 13 percent ahead of the RX 9070 XT at 1080p in GPU-limited rasterization workloads. That is a real and measurable lead. It is not a commanding one.
Upscaling and Frame Generation: The Invisible Wildcard
Raw rasterization numbers only tell part of the story for how these cards actually perform in daily use. Nvidia’s DLSS 4 with Multi Frame Generation is available to 5070 Ti owners, and AMD’s FSR 4 with fluid motion frames runs on the 9070 XT. Both technologies can multiply effective frame rates well beyond what native rendering produces, which makes the 10 to 13 percent native gap feel less consequential in practice. A player using FSR 4 on the 9070 XT may end up with a smoother experience than the raw numbers suggest, depending on the title and their tolerance for upscaling artifacts.
This is where the pricing argument sharpens. If native 1080p performance is your only benchmark, the 5070 Ti costs roughly 25 percent more for roughly 12 percent more performance in rasterization. That math does not work out in Nvidia’s favor at MSRP, and it gets worse at the inflated retail prices that have followed both launches. For a purely rasterization-focused 1080p build, the 9070 XT is harder to argue against. For those who want access to DLSS 4’s specific implementation – particularly Multi Frame Generation – the calculus shifts, but that is a feature preference, not a rasterization performance story. You can see how the cards compare at 1440p with ray tracing enabled for a more complete picture of where Nvidia’s pricing starts to make more sense.
Thermal, Power, and Real-World Fit
The RX 9070 XT draws around 250 to 260 watts under load in rasterization workloads, while the RTX 5070 Ti runs closer to 280 to 300 watts depending on the title and driver state. Neither card requires exotic cooling, but the 5070 Ti’s power requirements mean a higher-tier PSU recommendation – 850 watts minimum versus 750 watts for the 9070 XT. At 1080p specifically, where frame rates are high and thermal loads are sustained over longer gaming sessions, the 9070 XT’s lower power draw translates to lower temperatures in most mid-tower cases.
Driver stability is worth mentioning plainly. AMD’s RDNA 4 launch has been cleaner than some previous generations, with fewer reported game crashes and compatibility issues than RDNA 3 saw in its early months. Nvidia’s Blackwell drivers have also been relatively stable at launch, though some users have reported issues with frame generation in specific titles. Neither card launches in a particularly troubled state, which makes the performance and value comparison the primary decision factor rather than software reliability concerns.
Which Card Wins at 1080p Rasterization
For a strict 1080p rasterization build, the RX 9070 XT makes a stronger value argument than the performance gap alone would suggest. The 5070 Ti is faster – consistently, measurably faster – but the margin rarely translates into a meaningfully different gaming experience when both cards are running games well above the frame rate thresholds most players use. Spending 25 percent more to gain 10 to 13 percent in raw rendering at a resolution where CPU bottlenecks regularly intervene is a difficult case to make.
That said, the 5070 Ti is not overpriced in a vacuum. It holds genuine advantages in heavier workloads, scales better at higher resolutions, and offers Nvidia’s broader software ecosystem. Players who plan to move to 1440p or 4K eventually, or who run professional workloads alongside gaming, may find the premium justified across the card’s full lifespan. At 1080p rasterization specifically, though, the 9070 XT closes the gap to a point where the price difference becomes the loudest argument in the room.
The 5070 Ti’s strongest case at 1080p comes in titles where Nvidia’s architectural strengths show clearly – particularly in workloads with complex geometry and high draw call counts. If your game library skews heavily toward that category, the lead is real enough to notice. If you play mostly competitive titles where both cards are bottlenecked by the CPU anyway, you are paying a $150 premium for a performance delta that your monitor’s refresh rate cap will erase.
