Well, looks like Softbank is very much intent on shooting themselves in the dick (because that's a lot more embarrassing than simply shooting one's self in the foot) and killing off their gold egg laying goose. First they tried it by trying to sell ARM to Nvidia and regulators were like "LOL...no." Now they're trying it again by attempting to extort their current licensees and this IPO stupidity.

Softbank's shenanigans with trying to sell ARM to Nvidia already "persuaded" Intel to partner with SiFive to develop the Horse Creek platform. That's a real thing, not vaporware. I can envision a future where Intel focuses on e-cores for "high performance" and RISC-V cores for "efficiency" in every segment that does not involve spittle flecking rubes screaming "I NEEDS ME 60 HOJILLION GHZ FOR TEH BENCHMARKS!" Which, let's be real here, is a very, very tiny segment of the global market. Enterprise customers don't give a shit about HEDT or even "gaming" CPUs and that is where a vast amount of Intel's revenue stream comes from. They care about "good enough" at cheap enough and Xeons for the datacenter because that's the creme de la creme of revenue streams. That's literally a money printing machine. But they're also feeling the pinch there as the arguably world's largest cloud provider doesn't use x86 all that much. They use their own ARM based CPU design, the Graviton line (among others). So if Softbank decides to go balls deep in to the stupidity, you'll have Amazon and Intel partnering up to develop the next generation of AWS machines. Amazon saves money by not paying extortion fees to ARM and by not needing to develop their own in-house CPUs and Intel makes more money because they'll have a product that does what ARM is currently best at; performance per watt per dollar. Intel can guarantee availability to Amazon because RISC-V is an open source ISA. They'll just serve as the development and manufacturing house to Amazon. Much in the same way that TSMC does for AMD, Nvidia, and other right now.

AMD is pretty tight lipped on their future plans but they're already a RISC company. Implementing any RISC ISA is easier for them than it is for Intel. Then there's Samsung, Qualcomm. MediaTek, and Rockchip. Put all of them in to an industry group ala Vulkan and now ARM is no longer a value proposition.

We need to remember that ARM is profitable to the tune of ~$2 billion/year. It's their parent company, Softbank, that is absolutely hemorrhaging money. $5.4 billion last yer. So I am not shaking my finger at ARM but ARM is not going to survive the decade if they let Softbank dictate how they run a CPU business, business daddy or not.

Me, personally, I have abandoned ARM and am fully backing RISC-V. Yes, RISC-V is behind the Cortex-X line in terms of performance but that's a gap that is largely due to a lack of funding and a fragmented development landscape. You get a large company, or several large companies, behind it that agree on a block layout and starting taping out and RISC-V will be on-par or have surpassed ARM in 2-3 years. All you really need is someone like Jim Keller to devote their time to it.

Oh, wait, he's already doing that right now.