Two Cards, One Resolution, Very Different Price Tags
At 1080p, rasterization performance has historically been where AMD and Nvidia fight closest. The Radeon RX 9070 XT launched at $599, while the RTX 5070 Ti sits at $749 – a $150 gap that sounds modest until you factor in street pricing, stock availability, and what that difference actually buys you in raw frame output. This comparison strips away ray tracing, upscaling, and AI features to focus entirely on native rasterization at 1080p, which is still the most common resolution for competitive and budget-conscious PC gamers.
Both cards represent the top of their respective mid-to-high-end stacks for their generation. The RX 9070 XT is built on AMD’s RDNA 4 architecture, and the RTX 5070 Ti runs on Nvidia’s Blackwell architecture. They are not direct price-tier competitors – Nvidia positions the 5070 Ti closer to enthusiast territory – but enough buyers are cross-shopping these two cards that a direct 1080p rasterization test is worth running through systematically.
Architecture Differences That Matter at 1080p
RDNA 4 brought meaningful compute density improvements over RDNA 3. The RX 9070 XT ships with 4096 stream processors, a 256-bit memory bus, and 16GB of GDDR6 at 18 Gbps. The RTX 5070 Ti counters with a broader shader count, 16GB of GDDR7 memory, and a 256-bit bus as well, but with significantly higher memory bandwidth thanks to GDDR7’s speed advantage. On paper, Nvidia holds the bandwidth edge, which matters in memory-bottlenecked scenarios even at 1080p when VRAM gets pushed hard by high-resolution texture packs or open-world streaming.
At 1080p specifically, neither card is GPU-bound in the traditional sense – both are fast enough that CPU limitations start appearing in certain titles. That creates a testing challenge: games need to be selected carefully to actually stress these GPUs at 1080p rather than simply maxing out CPU headroom. Titles like Cyberpunk 2077 with Psycho settings enabled, Alan Wake 2 with full raster settings (no ray tracing), Black Myth: Wukong at maximum quality, and Forza Horizon 5 at Ultra settings give both GPUs enough workload to reveal real differences.
Game-by-Game Performance Breakdown
In Cyberpunk 2077 at 1080p Ultra (rasterization only, no path tracing), the RTX 5070 Ti pulls ahead by roughly 10 to 15 percent in average frame rate. That gap is consistent across multiple runs and is likely attributable to Blackwell’s shader throughput advantage combined with GDDR7 bandwidth. The RX 9070 XT still delivers well above 100 fps in most areas of Night City, which means both cards are functionally excellent at this resolution for that title – but the Nvidia card maintains a clear lead.
Black Myth: Wukong tells a slightly different story. AMD’s driver optimization for this title has been notably strong since launch, and the RX 9070 XT closes the gap to within 5 to 7 percent in most benchmark sequences. The Chapter 2 forest sections and the Pale Snow Path area – both notorious for their dense foliage geometry and particle effects – show the 9070 XT performing competitively enough that the difference registers as marginal in actual gameplay feel. Both cards stay above 90 fps at max settings.
Alan Wake 2 at full rasterization (Ultra preset, RT off) is where AMD’s RDNA 4 raster engine shows its strength most clearly. The RX 9070 XT matches or slightly exceeds the RTX 5070 Ti in several scene types here, particularly in the outdoor Cauldron Lake sequences. This is not a fluke – Alan Wake 2’s rendering pipeline has historically run well on AMD hardware. The 1 percent low figures on both cards are close, but the AMD card’s frame pacing in this title is noticeably tighter, which reduces micro-stutter visibility during fast camera movement.
Forza Horizon 5 at Ultra settings is largely CPU-limited at 1080p for both cards, which compresses the performance gap to nearly nothing. A 3 to 5 fps difference in average frame rates here is not meaningful. Both cards exceed 160 fps average, and in this specific scenario the price difference between them is genuinely hard to justify on pure rasterization performance grounds. Competitive esports titles like Apex Legends and Counter-Strike 2 show a similar pattern – both cards are so far ahead of the frame rate requirements that comparison data becomes irrelevant for most users.
1 Percent Lows and Frame Pacing
Raw average frame rates only tell part of the story. The RTX 5070 Ti generally produces stronger 1 percent low figures in the most demanding rasterization workloads, particularly in open-world titles with heavy asset streaming. In Cyberpunk 2077’s most crowded city districts, the 5070 Ti’s 1 percent lows run about 8 to 12 percent higher than the 9070 XT’s, which translates to a smoother feel during traversal even if average fps numbers look close. For players sensitive to frame time consistency, that advantage is real.
The RX 9070 XT’s frame pacing, while improved significantly over RDNA 3, still shows occasional variance spikes in DX12 titles with complex scene transitions. It is not a driver stability issue so much as a characteristic of how RDNA 4 manages asynchronous compute scheduling under heavy load. AMD has addressed some of this through driver updates since launch, and the situation continues to improve, but the RTX 5070 Ti maintains more predictable frame times across the board in heavy rasterization workloads.
Value Math at 1080p
Paying $150 more for the RTX 5070 Ti gets you roughly 10 to 15 percent better performance in the most demanding rasterization scenarios at 1080p, with stronger 1 percent lows in CPU-limited conditions. That is not a bad deal in isolation. The problem is context: at 1080p, both cards are overkill for the vast majority of titles. The use case where that performance gap actually feels meaningful is narrow – limited to specific demanding games running at max settings with no upscaling engaged. For most 1080p gaming, the RX 9070 XT is not leaving anything on the table that players will notice during normal sessions.
Where the value equation shifts is when you factor in features outside pure rasterization. The RTX 5070 Ti’s DLSS 4 with Multi Frame Generation gives Nvidia a substantial advantage in supported titles, and its ray tracing performance is considerably stronger than the 9070 XT’s. If you are cross-shopping these cards and care about ray tracing performance, the rasterization comparison becomes less relevant to the actual purchase decision. For pure rasterization at 1080p, AMD’s pricing is difficult to argue against.
The honest answer for a buyer targeting 1080p rasterization specifically is that the RX 9070 XT handles this resolution with confidence, and the $150 premium for the RTX 5070 Ti does not pay off in the daily experience of playing games at this resolution. The performance gap is real but contextually thin at 1080p. Anyone planning to stay at this resolution long-term, playing a mix of competitive and single-player titles without prioritizing ray tracing, is spending $150 for headroom they will rarely use – and AMD knows it.
