Intel Corp (INTC)

[This Causes an Error]

It's just too much hassle for no payoff. What advantage is there to nerfing Core and trying to stick it into a phone when you could build an SoC around CPU cores that are designed for the task?

By using ARM cores, Intel is guaranteed to have a competitive architecture (even Qualcomm's Kryo is kind of a dud compared to ARM's A72) and all it needs to worry about is the SoC integration and focus on the key intellectual properties that it can actually differentiate on (aka modem, memory controller, etc.)

Atom will never be competitive with ARM's high performance mobile cores, and it will always have nagging compatibility issues. The fact that Intel basically abandoned ship on smartphones and Android tablets for x86 means that the software people are going to give precisely zero hoots about making sure their code works on x86, let alone is optimized for it.

Atom will do its duty as a low-cost PC processor core, an IoT core, and even as the control processor in the basebands (XMM 7560 for example uses an Atom processor). It has a future, and it will make Intel money, but it is not leading edge high performance smartphone AP material.

Anyway, if Intel wants to make some real $$$ in mobile, it should be trying very hard to win Apple's A-series chip biz from a foundry perspective. Intel already won the Apple modems, so fabbing the SoC for them, too, would be really nice.

However, it would appear that since Intel's TMG has stumbled so badly with 14nm and based on everything Intel hasn't said about 10nm (and the fact that they will be releasing 14nm Core processors even in 2018), that TSMC's consistent yearly execution could keep Intel from winning that Apple biz.