Well let's see. Intel holds x86 patents and does not allow anyone to use said unless under severe duress (thank for IBM for at least trying to prevent a monopoly) I'd say you asked the most loaded question of the year.
You can call it a loaded question, you can call it a banana, call it whatever suits you.
FWIW it wasn't intended to be a loaded question, rather it was intended to be a thought-provoking leading question ala the
Socratic teaching method.
The point I was originally making is that Intel has always had plenty of x86 business carcasses to pick through over these past 30 years, but they never once felt the ROI was sufficient to warrant doing so. That internal vetting process for M&A ROI is still there today, so one should interrogate history of past decision making if one wishes to proceed with insight towards projecting how Intel's decision makers might assess an M&A with AMD.
So the question would be what makes AMD any different than every other x86 business that Intel has opted to not purchase in the decades leading up till now?
Or to state it differently, outside of AMD's NexGen purchase for the K6 micro-architecture jumpstart, why has AMD also elected to not bother purchasing any other available x86 business in the past two decades?
My guess is that the
perceived ROI many laymen assume would come from M&A activity in the x86 industry is not the same order of magnitude as the actual ROI that data-empowered decision makers know it to be...and so those data-empowered decision makers (be them at AMD or Intel) have intentionally opted to not bother pursuing such projects.
Also, FWIW IBM didn't force Intel to outsource to AMD so as to prevent an x86 monopoly. What IBM stipulated was that Intel had to have two geographically disparate fabs producing the x86 chips so that IBM would have acceptable risk management in the supply chain.
Intel ran the numbers and decided that it would be less costly and more profitable for Intel if they outsourced production to AMD rather than taking on the front-loaded capex hit of building out another fab themselves at that time. Intel's solution the supply-chain risk management concern was acceptable to IBM, and so it became a matter of history after that.
Few companies ever bothered worrying about licensing their architecture. DEC didn't worry about it, IBM didn't worry about it, Cray didn't worry about it. In those days it was very common for businesses to retain all rights to their homegrown architectures, but they would have multiple production locations for supply chain risk management. Intel had a business choice to make, and that was all that it was at the time.