> I'd expect them to have the resources to start working on ARM now, while planning for RISC-V further out.
I doubt that strategy would work. The problem in the PC space is that there's no Apple-like leader who can boldy drive a change in CPU architecture. Whoever is first out of the gate bears all the risk of the change not catching on. If you lose the bet, you've wasted massive R&D spend to make a lemon. Take a look at the reviews of Microsoft's ARM surface laptops to see what I mean here.
I think there's an opportunity over the next 5 years or so for PC makers to ride the M1's coattails and shift from x86 to ARM. Especially given intel's failure to move off 14nm. But I doubt lightning will strike twice. If windows moves to ARM now, we'll be stuck there for at least another few decades. And the technical argument for RISC-V will be much weaker if ARM rules the roost.
Weirdly the strongest counterargument I can think of is due to electron. Chrome will maintain first class support for any and every popular CPU architecture. So the more that desktop software is written on the web and in electron apps, the easier any architecture transition will be down the road. That and the server space. Linux already supports RISC-V extremely well given how few linux-capable RV chips there are in the wild.
I doubt that strategy would work. The problem in the PC space is that there's no Apple-like leader who can boldy drive a change in CPU architecture. Whoever is first out of the gate bears all the risk of the change not catching on. If you lose the bet, you've wasted massive R&D spend to make a lemon. Take a look at the reviews of Microsoft's ARM surface laptops to see what I mean here.
I think there's an opportunity over the next 5 years or so for PC makers to ride the M1's coattails and shift from x86 to ARM. Especially given intel's failure to move off 14nm. But I doubt lightning will strike twice. If windows moves to ARM now, we'll be stuck there for at least another few decades. And the technical argument for RISC-V will be much weaker if ARM rules the roost.
Weirdly the strongest counterargument I can think of is due to electron. Chrome will maintain first class support for any and every popular CPU architecture. So the more that desktop software is written on the web and in electron apps, the easier any architecture transition will be down the road. That and the server space. Linux already supports RISC-V extremely well given how few linux-capable RV chips there are in the wild.