ARM, Nvidia, RISC-V, Intel

Over the last few years we’ve been seeing an interesting case of The Innovator’s Dilemma playing out. The case is ARM, and how it through smartphones have grown from energy efficient but comparably slow CPUs, into something which is turning to an overall better choice for general use CPUs.

Looking at how Apple is moving to ARM in its computers, not just the iPhone, and how Amazon are using ARM in their cloud servers, it is clear that we’re very close to having ARM replace Intel across the board. Intel, on its side, have not managed to keep up, which, when viewed from the top of the S-curve, fits perfectly with what we’d expect to happen.

So with that, ARM is not just an interesting technology, but also an interesting business. It’s not surprising to see current interest in buying ARM. If and when that will happen remains to be seen, but to expect a new owner of ARM in the next few months is pretty clear in my view.

Whoever ends up owning ARM, Intel is in trouble no matter. They do not have an architecture which is capable of much further growth. It’s stagnant. I’m sure AMD can easily enough turn around and start selling ARM based CPUs, as they see that market growing. But for Intel to jump on the ARM path, that’s hard to imagine. There’s too much history tied to the x86 architecture for them to make that jump. So I’m not hopeful for Intel, and I wouldn’t be surprised if they almost end up dead in a few years.

But only then will they also be able to move forward, and that’s where I see RISC-V becoming interesting.

If we have Nvidia owning ARM, I concur with many others that, they probably will eventually introduce too much friction. Companies like Apple, Amazon, Google, and maybe even Microsoft through Azure, not to forget non-U.S. companies, will become unhappy with their working relationship with Nvidia.

And because we at that point just recently switched from x86 to ARM, a new switch to RISC-V doesn’t seem that farfetched. Also, because ARM and RISC-V are both a RISC designs, you can argue its an easier switch than from the CISC design of the x86.

It’s at this point, with a new switch to RISC-V in 10 years or so, that Intel can jump back on the path to growth. If they can see this happening, and plan ahead, they could revive themselves through RISC-V.

RISC-V will remain an open source project, which would be a requirement from the industry, given their, at that point, lessons learned from ARM as owned by Nvidia. But open source projects need business backing to grow, and Intel could be one of those, as a significant member of a greater RISC-V community.

Time will tell..