On raw throughput metrics, the AMD Radeon RX 9070 XT holds a commanding lead. Its 48.7 TFLOPS of floating-point performance is roughly 58% higher than the RTX 5070's 30.87 TFLOPS, and that advantage carries through to texture and pixel fill rates as well — the RX 9070 XT pushes 760.3 GTexels/s versus 482.3 GTexels/s, and 380.2 GPixel/s versus just 201 GPixel/s. In practical terms, higher pixel fill rate translates to the GPU's ability to render more pixels per second, which matters at high resolutions and with demanding anti-aliasing; the RX 9070 XT's advantage here is substantial. Its 128 ROPs versus the RTX 5070's 80 ROPs further reinforce this edge in pixel output.
The clock speed picture is more nuanced. The Inno3D RTX 5070 runs a higher base clock of 2325 MHz, while the RX 9070 XT starts much lower at 1660 MHz but boosts aggressively to 2970 MHz — nearly 460 MHz above the RTX 5070's turbo of 2512 MHz. This wide boost range on the AMD card is what enables its superior computed throughput figures. The RTX 5070's larger shader array (6144 shading units vs 4096) is an architectural differentiator, but it does not overcome the RX 9070 XT's advantages in clock headroom, ROPs, TMUs, or memory speed — the RX 9070 XT's 2518 MHz memory speed also outpaces the RTX 5070's 1750 MHz, which benefits bandwidth-hungry workloads.
Based strictly on the provided performance specs, the RX 9070 XT has a clear and consistent advantage across nearly every compute and throughput metric in this group. The RTX 5070 counters with a higher base clock and more shading units, but those individual points do not offset the RX 9070 XT's leads in TFLOPS, fill rates, ROP count, TMU count, and memory speed. For users prioritizing raw rendering horsepower as reflected in these figures, the RX 9070 XT is the stronger performer here.