What makes you think that Intel has to be on par with NVidia for their first try? Intel will get into the high end league in much less time than it took NVidia and AMD to get where they are now and Intel's oneAPI is already CUDA compliant in case you didn't know. DLSS is an NVidia proprietary tech that needs to be implemented game by game. Intel will sponsor the upcoming Hitman 3 and many other games, and we'll see how long devs can keep justifying supporting a proprietary feature from a chip designer who doesn't have a full CPU + GPU ecosystem.
Talking about Anti-trust? Should I remind you that Intel retained their PCIe 4.0 Optane (P5800X) SSD until Ice Lake-SP and Rocket Lake-S were close to their release window. Intel is an ecosystem and they can do whatever they want with it.
I would be surprised if the people in charge will see Intel actively forbidding a nvidia/intel dGPU/CPU combo on the same level. Back when Larabee was a thing the FTC went after them:
The FTC Sues Intel Over CPU & GPU Competition (anandtech.com).
Look at it this way: Optane wasn't working with an AMD cpu, but they never went and told to Dell: "You cannot sell an Intel pc if they don't exclusively have an optane ssd." Optane made Intel cpu a bit more appealing, but you are not forced to use Optane if you got an Intel CPU. That nuance is important. (It's like saying: "look, Microsoft already got a small ecosystem going on with the surface. Then one day they decide to make their own cpu/gpu, and go the Apple way while forbidding, anyone else to make a windows compatible pc. That ain't gonna fly, and Microsoft got sued for less).
And as you're aware, lately the E.U and U.S are pretty trigger happy when it comes to antitrust.
You are the one who said that Tiger lake H won't be sold with RTX 3000, meaning that if Intel cannot compete right now, they are effectively going to be out of the gaming laptop market until they manage to get a parity. Going from a mx350 to an RTX 3060 is a big, big jump and we don't have any data suggesting they got something able to do that.
Intel OneAPI still require the dev to do some work, and history showed me that devs are either slow or won't do it at all if there no benefits. When Apple
got into a fight with Nvidia, they lost a bunch of their customers base who had to use windows even though they were so deep into the Apple ecosystem. It's only now that some dev finally announced a metal version of those apps. Even Pixar who got a tight relationship with Apple is merely considering metal, but their production tools are still deep into the Nvidia ecosystem.
So, I won't jump into any hype train until I see a strong support
for years and years people from Otoy and redshift kept saying "AMD support will eventually come" and don't even get me started on Autodesk. Maya is a software that's been available on every OS even in the Power Pc days, but there's still no words about a native ARM version in the work even though there's a fair amount of Mac based vfx/animation studio who will make the transition. So I have very little faith about seeing Arnold working in Intel gpu's in the close future.
It's not that I don't want to see Intel being able to provide a great ecosystem, it's just that I know that it's not as easy as you make it sound like. Metal was Apple reaction to Nvidia effectively neutering open CL. They tried an open API, and it failed. Everybody wants to make their ecosystem, but the E.U and U.S government are constantly looking over the shoulder of tech company getting too zealous.
.
and as long as I know GPUs have their own DRIVERS to compute graphics workloads.
