I see Power architecture is going into a sunset. There has been talks of selling Power to external parties, but Power shares the basic processor architecture with Z as well as IBM i. I cannot think IBM will give up Z.
Will IBM keep spending money developing Power11?
Yes you are correct, IBM could sell off both storage and Power, BUT most storage makers are moving to a cloud enviro, and we most likely should expect IBM to follow. That doesn’t mean they wouldn’t just sell all of the manufacturing essentially exiting the HW business and do a GF deal with the purchasers. IBM makes a substantial amount of their profit from their HW sales on IP. Moving to an exclusive IP strategy would definitely make a lot of sense, BUT you have to find a willing partner. Who wants to be in a commodity business? I would say a partner who bulk buys components thus leveraging their supply chain, along with someone who can benefit from IBM’s labs and IP technology thus reducing their research.
15 - 20% of all the power chips sets land into IBM storage, so IBM has a vested interest to keep power
Or to sell both Power AND Storage.
Would IBM sell off Power brand?
Seriously??? IBM Management would sell their own mothers for the right price 😏
AIX is already re-deployed to Mexico - and even my product has been trying to deprecate it. Offhand I think there's mostly one HUGE customer that is AIX based that's indirectly keeping it alive.
Linux and ZLinux and Zos. ixNay ixAay
15 - 20% of all the power chips sets land into IBM storage, so IBM has a vested interest to keep power, or at least to keep developing the chip. It’s aligns quite nicely with the mainframe offerings. Also IBM loves the 1080 enterprise boxes, as they are real money makers with good margins, and they drive consulting engagements. (they feed the new IBM strategy) The rest of the Power line doesn’t drive consulting engagements, and IBM would most likely like to exit those HW offerings. (the margins are much lower, and the manpower required to sell the smaller boxes is actually higher due to the volumes. That’s why IBM just fulfills thru the channel. I expect IBM will find a buyer and do a global foundries deal (do the development so they can sell the IP, but exit the manufacturing business with the promise to buy smaller power boxes when they need to for approx 10 years) They will license the OS’s (Unix and OS/400) as they are unique to IBM. The problem is IBM has ignored the Power install base for so long that it is fading away and as such there are only a couple of potential buyers of the product who could benefit from purchasing. IF and it’s a big IF IBM can come to terms with someone, don’t expect a big transaction. Best guess is 1 years worth of revenue
Power as a bare system is no questionably dying(less and less people trying to develop on it since its very proprietary to hardware) but IBM can't actually get rid of power until all of it's underlying technologies are replaced first. For example, IBM actually would have 2 paths to replacing power, They would need to either 1. develop a new storage controller on x86/redhat system or port AIX/iSeries operating systems and it's software to x86 systems. This is a lot of work either way and I honestly don't think anyone at IBM is even capable of doing something like this anymore. All the young talent has left, and the old timers want to retire now. In my opinion, The correct answer is to completely port over all the operating systems and software and sell whats left of power architecture(patents/designs) to Apple or ARM.
Maybe I should be CEO? :P
IMHO, they would like to dump P-series to someone else to build/develop with contracts to use it for awhile longer in servers and storage. They have to keep Z for the customer base.
IBM has had the death rattle before, but this time, I'm not sure they will recover. You can only cut costs so much....