I really think intel needs to be taken to with a big meat cleaver, and chopped into two distinct parts - one for chip design, and one for fabs. The poor sod chip designers were held back for years, being tied to intel's inferior process nodes (when compared to the likes of TSMC), whereas AMD were free to get on the cutting edge. Also, if I was Lisa Su from AMD, would I take my advanced CPU designs to TSMC for production, or to a direct competitor like Intel - even if both TSMC and Intel could offer the exact same thing for the same price? I'd go with TSMC, and not give away IP and profit to a direct competitor. Even if I wasn't in direct competition with Intel, you still wouldn't want to use Intel fabs as you'd imagine that Intel would give top billing tog Intel's own chips, and any fab customer would play second-fiddle. If there were fab delays, then they are more likely to prioritise their own chips and bump others, to keep their retail arm happy.
So both sides of the business hold each other back. It's a terrible model. The so-called "synergies" don't outweigh the costs. The rest of the industry has basically proven that.
This is a bad move by Softbank. They are just delaying the inevitable. What needs to happen is for intel to fully crash and burn into the ground like a Boeing jumbo jet. Then from the ashes, after liquidation they can split into a dedicated small and nimble design team, where real innovation is rewarded, and also a separate fab team, who are outwardly focussed on winning contracts across the whole Industry. The design team wouldn't be tied to x86 either - they can also embrace Arm and RISC-V.