- A YouTuber has been outlining some compelling theories on RX 9070 efficiency versus the RTX 5070
- This theorizing relies on purported inside AMD benchmarks for the RX 9070, and serviette maths for RTX 5070 body charges
- If it pans out, the RX 9070 XT might simply beat the RTX 5070 Ti, and even come inside 10% of the RTX 5080
AMD’s RX 9070 fashions might handily outgun Nvidia’s mid-range RTX 5070 Ti and 5070 graphics playing cards, if a thought of prediction from a YouTuber pans out.
That is common rumor peddler Moore’s Legislation is Useless (MLID), who in his newest video (embedded under) engages in some serviette maths to work out the place the efficiency of the RTX 5070 and 5070 Ti GPUs are more likely to weigh in (extra on the intricacies therein shortly). The YouTuber then compares that to inside benchmarks purportedly carried out by AMD a month in the past with its RX 9070 fashions.
The upshot is that this: going by these inside exams from AMD – add skepticism with all this, which means the benchmarks, and in addition MLID’s personal theories – Staff Purple was focusing on a slight win for the RX 9070 XT over the RTX 4080 Founders Version (to the tune of three% or so).
- AMD exec drops hints on RX 9070 pricing and a few PC players are panicking – however this appears like excellent news to me
- AMD RDNA 4 GPU rumors flood forth, together with attainable identify change to RX 9070 – as a result of greater is healthier, in comparison with Nvidia’s RTX 5070?
MLID then took that degree of estimated efficiency and overlaid it on a graph of benchmarks (a 17-game common) from {Hardware} Unboxed that features the RTX 5080. From this, we see that in principle, the RX 9070 XT is inside 10% of the RTX 5080 for rasterized (non-ray tracing) efficiency at 4K decision.
On prime of that, the YouTuber added within the talked about serviette maths approximations of RTX 5070 efficiency, which is that the vanilla RTX 5070 is more likely to are available in at about 20% sooner than Nvidia’s RTX 4070 (so within the ballpark of the RTX 4070 Ti). And that the RTX 5070 Ti is more likely to be a relatively minor generational uplift, and perhaps solely barely sooner than the RTX 4070 Ti Tremendous.
Granted, that provides in a great deal of uncertainty, and ifs-and-buts, although it’s based mostly on sound sufficient reasoning. (Particularly the uplifts we’ve seen for the RTX 5090 and 5080, on common – a robust taste of the architectural features for Blackwell, in different phrases – after which the relative specs of the brand new RTX 5070 fashions versus their predecessors).
RTX 5070 efficiency could not pan out like this, but when it roughly does, MLID theorizes that the RX 9070 XT (based mostly on these inside AMD benchmarks) might probably be 15% sooner than the RTX 5070 Ti. And that the RX 9070 versus the RTX 5070 might see a win for AMD, too, extra to the tune of 10%, however nonetheless, a marked victory.
Evaluation: Efficiency means little with out value
Proper, so is AMD set to personal the GPU mid-range this yr? Nicely, as I’ve already mentioned a few occasions – however it bears one other point out simply to underline – a whole lot of that is up within the air theorizing, albeit workings-out that make sense to me. MLID lays some heavy caveats on all this himself, though the YouTuber does assert that he’s assured sufficient in these predictions on the entire.
We must always keep in mind that the graphs used (from {Hardware} Unboxed) are simply straight rasterized efficiency. Though MLID additionally notes he’s assured AMD has nearly caught as much as Nvidia with ray tracing on this era, however one other main piece of the puzzle is DLSS 4 and Staff Inexperienced’s new body era. The latter MFG function, and different enhancements in DLSS 4, are literally enormous – and that shouldn’t be underestimated. We don’t but know the way FSR 4, AMD’s rival next-gen tech, will shake out, and so that continues to be a reasonably weighty query mark right here.
One other vital level right here is that it’s all very properly analyzing (potential) relative efficiency ranges in a theoretical train like this, however even when this proves right as to the comparative body charges we’ll get from the RX 9070 versus RTX 5070 fashions, there’s AMD’s pricing to think about. We all know the tough worth proposition of the RTX 5070 flavors as now we have the MSRPs, however we don’t with the RDNA 4 graphics playing cards.
My fear, then, is that AMD can be calculating the place to pitch RX 9070 asking costs based mostly on the Nvidia RTX 5070 critiques after they arrive in February (properly, if Staff Inexperienced sticks to its promised launch timeframe for these mid-range graphics playing cards). It’s actually been rumored that AMD continues to be very a lot weighing up pricing, and the query then turns into: how a lot does Staff Purple wish to take Nvidia down within the mid-range area?
If that’s a robust motive right here, AMD would possibly are available in with actually aggressive MSRPs for the RX 9070 fashions. However, if maximizing income and return is larger up the precedence listing for RDNA 4, then we might get weightier than rumored asking costs.
Who is aware of, is absolutely the purpose, and the RX 9070 will solely be an RTX 5070 killer – assuming MLID’s serviette scribbling and GPU hypothesizing is in the fitting ballpark – if AMD costs it to be an RTX 5070 killer. Hopefully that’s the intent, and MLID suggests $499 and $649 (US) as attainable value tags for a suitably aggressive transfer with the RX 9070 and its XT sibling respectively.
Beforehand, there have been hopes of a sub-$500 value for the RX 9070, but when efficiency does form up something like as steered right here, there’s no cause AMD would want to dip decrease than the talked about $499. And once more, this comes again to my fear that AMD would possibly be at liberty to simply push pricing tougher than initially meant, maybe, if the RX 9070 fashions are outmuscling the RTX 5070s on this vein.
- MLID additionally tells us Nvidia RTX 5090 GPU inventory is ‘mainly non-existent’
- Nonetheless frightened RTX 5000 GPUs don’t have sufficient VRAM? Nvidia’s secret weapon is highly effective AI texture compression, and it’s calmed a few of my fears
- AMD vs Nvidia: who’s the graphics card champion?
No Comment! Be the first one.