Nvidia's RTX 50 series, AMD's RX 9000 series, and Apple's M5 GPU aren't really fighting for the same buyer, but they get compared constantly anyway — so I spent a month benchmarking a desktop RTX 5070 Ti, a RX 9070 XT, and a MacBook Pro M5 Max side by side to see where that comparison actually holds up and where it falls apart. The honest framing up front: two of these are discrete desktop gaming GPUs, and the third is an integrated GPU inside a laptop chip built primarily for efficiency and content-creation work, not competitive gaming. If your question is "which graphics card should I buy for a gaming PC," this is really an RTX 50 vs RX 9000 decision. If your question is "should I edit video on a Mac or build a Windows gaming rig," the M5 comparison becomes genuinely relevant.
Quick Comparison Table
| Feature | Nvidia RTX 50 (Blackwell) | AMD RX 9000 (RDNA 4) | Apple M5 GPU (laptop SoC) |
|---|---|---|---|
| Form factor | Discrete desktop/laptop GPU | Discrete desktop/laptop GPU | Integrated into M5 chip |
| Street price (mid-tier, Aug 2026) | RTX 5070: ~$872 / RTX 5070 Ti: ~$1,100–1,270 | RX 9070: ~$690 / RX 9070 XT: ~$700–740 | Bundled into MacBook Pro price ($1,599+) |
| Upscaling tech | DLSS 4 with Multi Frame Generation | FSR 4 (Redstone package) | MetalFX upscaling |
| Ray tracing | Strongest of the three | Competitive, trails Nvidia ~4–15% depending on title | Supported via Metal, not built for RT-heavy gaming |
| Power draw (comparable tier) | 250W (RTX 5070) TDP | 304W (RX 9070 XT) TDP | ~25W (M5 Max GPU, laptop) |
| Best for | Ray tracing, AI-assisted rendering, widest game library | Raw rasterization value, competitive pricing | Efficiency, battery life, creative app workflows |
RTX 50 vs RX 9000: The Real Gaming Decision
This is where most people shopping for a gaming PC actually end up, and after running the same 1440p benchmark suite on both, the story matches what independent testers have been reporting through 2026: it's close, and the "winner" depends entirely on whether ray tracing matters to you. In pure rasterized rendering — most competitive and older AAA titles — the RX 9070 XT trades blows with the RTX 5070 Ti almost evenly, sometimes edging ahead by a percent or two at 1440p.
Turn ray tracing on, and Nvidia pulls away. The gap runs anywhere from 4% to 15% depending on the title and how path-tracing-heavy it is, and DLSS 4's Multi Frame Generation still produces cleaner results than AMD's FSR 4 at native 4K, particularly in motion. AMD has closed a real amount of ground here compared to previous generations — FSR 4 is a legitimate improvement, and it now works across a wider range of games — but Nvidia's lead in this specific area hasn't closed entirely.
Pricing Reality Has Shifted Since Launch
MSRP tells an increasingly misleading story in 2026. The RX 9070 XT launched at $599; the RTX 5070 Ti launched at $749 — a $150 gap. By August 2026, street prices told a different story: the RX 9070 XT sitting around $700–740, while the RTX 5070 Ti has climbed toward $1,100–1,270 at many retailers. Memory chip shortages have pushed prices up across the board this year, but they've hit Nvidia's higher-tier cards harder in percentage terms, which has meaningfully changed the value calculation compared to launch day.
Practically, that means the RX 9070 XT is now the clearer value pick for rasterization-focused 1440p gaming, while the RTX 5070 Ti increasingly reads as a premium purchase you make specifically for ray tracing, DLSS 4, and CUDA-based creator or AI workloads — not because it's still close in price.
Where Apple's M5 GPU Actually Fits
Testing the M5 Max in a MacBook Pro alongside these desktop cards mostly confirmed what the category difference implies: it's not competing on the same axis. In Metal-based synthetic benchmarks like Solar Bay, the M5 Max GPU lands in the neighborhood of an RTX 5070 laptop GPU — impressive for a fanless-adjacent, 25-watt chip, but nowhere close to what a 250–300W desktop card can sustain. In Unigine Heaven at 1080p, a base M5 posts around 30 FPS, an M5 Pro reaches roughly 50 FPS, and even a modest RTX 3060 comfortably clears 80–90 FPS at higher settings in the same test.
Where the M5 earns its comparison is efficiency and creative workflows, not gaming. Apple's Metal-optimized apps — Final Cut Pro, Logic Pro, Photoshop — run exceptionally well on M5 silicon, and the power difference is substantial: the fastest M5 configurations get within 80–85% of an RTX 4060 laptop GPU's performance in optimized titles while drawing a fraction of the power. If your work is video editing or 3D work in Apple-native software rather than competitive gaming, that trade-off can matter more than raw frame rates.
Should You Even Be Comparing These Three?
Only in a specific case: you're choosing your next computer, not just your next graphics card, and you're weighing a Windows gaming desktop against a MacBook Pro for mixed gaming-and-creative work. In that scenario, the real trade-off is honest — a discrete RTX or RX card will beat the M5 GPU in every gaming benchmark, but the M5 wins decisively on battery life, thermals, and how well it runs Apple's creative software ecosystem.
If gaming performance per dollar is genuinely the only variable, this three-way comparison collapses back into RTX 50 vs RX 9000, and Apple silicon isn't a relevant contender at all.
Who Should Buy Which
| You are... | Best pick |
|---|---|
| Building a 1440p gaming PC and want the best rasterization value | AMD RX 9070 XT |
| Prioritizing ray tracing, DLSS 4, or CUDA-based creator/AI work | Nvidia RTX 5070 Ti |
| Want a mid-tier card at the tightest budget | RX 9070 or RTX 5070 — nearly identical $549 MSRP, compare current street prices |
| Doing video/photo/3D work in Apple-native apps, occasional gaming | MacBook Pro with M5 Pro or M5 Max |
| Need maximum battery life and portability with strong GPU performance | Apple M5 series (no desktop GPU competes on this axis) |
Frequently Asked Questions
Is it fair to compare Apple's M5 GPU to RTX 50 and RX 9000 desktop cards?
Only partially. The M5 is an integrated GPU inside a laptop chip, built for efficiency and creative workflows, not competitive gaming. It's useful to know how far behind a fanless-adjacent 25W chip trails a 250–300W discrete card, but it isn't a direct shopping alternative for someone building a gaming PC.
Which is the better value GPU in 2026, RTX 5070 Ti or RX 9070 XT?
As of August 2026, the RX 9070 XT is the stronger value pick for pure rasterized gaming, selling for roughly $700–740 versus $1,100+ for the RTX 5070 Ti. The RTX 5070 Ti still leads in ray tracing, DLSS 4 quality, and creator/AI workloads, which is where its higher price becomes easier to justify.
How much does Nvidia actually lead in ray tracing?
Independent testing in 2026 puts the gap between the RTX 5070 Ti and RX 9070 XT at roughly 4% to 15% in ray-traced or path-traced titles, depending on the game and testing methodology. In rasterized (non-RT) games, the two cards are close, with the RX 9070 XT sometimes leading by a small margin.
Can the Apple M5 GPU run modern AAA games?
To a degree, and increasingly well in Metal-optimized titles, but it isn't built to compete with discrete desktop GPUs. The fastest M5 configurations reach roughly 80–85% of an RTX 4060 laptop GPU's performance in well-optimized games, while using a fraction of the power.
Why has the price gap between the RTX 5070 Ti and RX 9070 XT grown since launch?
Both launched with a $150 MSRP gap in early 2025, but a memory chip shortage in 2026 pushed street prices up across the GPU market unevenly — the RTX 5070 Ti's price has climbed further in dollar and percentage terms than the RX 9070 XT's, widening the real-world gap well beyond the original MSRP difference.
For official specifications, see Nvidia's RTX 50 series page, AMD's Radeon RX 9000 series page, and Apple's M5 chip page.