I would take a guess that Apple is shipping (far) more TFlops of GPU power than Nvidia or anyone else in the mobile GPU market. Few people are buying laptops with 80-150w TDP GPUs, as those start to stretch the definition of both 'laptop' and 'battery powered'. Big gaming laptops with an hour of battery life are more akin to the luggables of yore.
That's fair. Nvidia has issues scaling their full systems down to laptop spec, and Apple almost has the opposite problem. They're both impressive in their own right, but right now Nvidia has both the performance and performance-per-watt crown in this space. The disparity in 3D applications (like gaming and Blender[0]) so ugly it's not even close.
And in all fairness - Apple's products might not need more GPU power. Cyberpunk and Elden Ring appear to be CPU-bottlenecked, if people are comfortable upscaling they could get a pretty comfortable Retina experience. The 2D optimization and media accelerators are a good focus for mobile hardware. For more demanding applications though, it looks like Apple's current approach is not scaling well.
Yeah I'm really curious what Apple's next-gen GPU (with raytracing and a bunch of other stuff) brings to fix some of these shortcomings. It was supposed to show up on last year's iPhone 14 followed presumably by inclusion in the M-series, and the 3nm process was supposed to be shipping this year, but everything got set back a year. In Mac-land the M2 wound up just being an overclocked M1, so we're left waiting for M3 to bring us a more competitive GPU.
The other half of the story is a lot of software (inc Blender, looking at these crazy results) just isn't well optimized and Apple is still struggling to win over developers in certain sectors of the market. Nvidia's decade+ investment in the software side has paid off so incredibly well for them, it's basically made the company.
Fact? Which Apple chips are outperforming the laptop 3070, much less the current-gen mobile 4090?