I mean yeah if you listened to the MACOM/AMCC merger call and look at that slide it just solidified what I thought was the issue the whole time. Connectivity was being used as a cash cow for computing. They couldn't invest properly into R&D for both so it was basically a parasitic relationship. Also, as much as they wanted to ditch dealing with legacy PowerPC stuff in the end it too was providing them with cash. Finally, even with the sale of their old Sunnyvale building it still wasn't enough to keep them afloat long enough for this supposed revenue stream. Sad but a combination of poor execution, bad market timing, and lackluster management.
When x-gene3 comes out it will be competing with INTC's future products (they will not stand still), so it might have happened, if all went well, for x-gene4 or x-gene5, but that entails very deep pockets and more than a little luck. I still prefer ARM's business model to INTC's vertical integration. I don't think we have seen the last of ARM in the datacenter.
Processor-wise MACOM seems content to stay within the carrier network with their processors: