Intel Corp (INTC)

[Unkwn]

Thanks for sharing that info (though you certainly cherry picked the bad stuff)!

Emerging markets (the previous PC growth driver) has completely left the PC market in favor of mobile devices. Intel used to break out geographic sales information in their earnings report. They no longer do this.

I question this as being a long term trend. Emerging markets are under pressure, especially the ones that largely depend on commodities (Brasil, Russia) and China in general. They won't do their business stuff on tablets, for sure, so there is a lot of potential for growth in that area. In addition, once consumers in those areas will become technically more capable, they will demand for devices which allow them to perform more complex tasks (including higher compute power, which, as of today, only Intel delivers). Not to mention gaming - huge potential for Intel in the EMs.

Caused by lack of innovation and "good enough" computing

That's the one I probably underestimated, to be honest. The problem is that Intel, without fierce competition from AMD, isn't advancing fast enough anymore, which makes people ask why they should replace their old computer. 10% more compute power doesn't justify a new 1000$ PC. There is one good thing, though: Device quality starts to improve, which may cause more replacements. New laptops from Dell, HP, Asus and the likes are of much higher build quality than they were in the past. I still think, though, that many are not happy to move over to Windows 10, since Microsoft just isn't able to provide real user friendliness, in my opinion. I see big potential for Chrome OS, though that will strengthen ARM's position (and Intel finally has a new opponent to vitalize the market).

One additional thought: Mobile devices, due to their strong market presence, lead software developers to not waste as much performance as they were used to in the past. If a website or application runs on a mobile phone just good enough, it runs great on a PC, leading to lower performance requirements for PCs. The same can be seen with websites, which optimize better for performance and the improvements on web browsers, lead by Google, are outstanding. This all leads to the impression of "good enough computing". I still don't think there will be a stop, it's more like a pause in the ever going cat and mouse game between hardware (faster) and software (more demanding).

The carriers want the make the most out of their spectrum which is their golden goose. They will not allow inferior modems to waste spectrum. Good luck catching QCOM. 5G is extremely complex and requires decades of know-how.

Well, Intel is pretty close to Qualcomm in terms of modems, from a technical standpoint. I think the gap is pretty much closed. An Intel modem in the iPhone will improve the acceptance further. I see no issue here.

The only problem is the integration of the modem into Intel's own APs. That's where this company failed horribly. I would really like Krzanich to share some of the reasons for this. The problem now is that there is no process advantage left anymore, which makes Intel compete on the same technical and cost level, in an extremely aggressive market. Not meeting milestones and releasing one generation every 2 years just isn't enough in this market. Either Intel improves execution, or they're out (and in danger).

Intel is double counting traditional x86 sales as IoT sales (eg. Gas stations using x86 PCs for unit displays) to paint a rosy picture

Intel is also counting Wind River as IoT even though Wind River supports ARM, MIPS and POWER as a majority fraction of their revenue

Yup, that's also what I think. There is some x86 market in the embedded world, e.g. for industrial PCs, Linux or Windows based terminals and so on. This is a good market to be in, but IoT, for me, is the cheap microcontroller with built in communication I can't see Intel (or baiscally anyone else, except ARM) making a dime on it. Qualcomm, Mediatek and others won't benefit from it either. Just the foundries and ARM.

Regarding the process advances: The announcements from ARM, Mediatek, TSMC and the likes are just virtual papwer as of now. They are for sure having the same problems as does Intel. There is no free lunch for them either. I bet we'll see devices in the market with 10nm in volume at about the same time from Intel as from TSMC. I am not talking about some lighthouse products where TSMC and others are basically paying for prestige (something Intel isn't doing for good reason). Maybe Intel will have a small lead regarding process characteristics towards TSMC, not so sure about Samsung though.

There remains still the question about how important this is these days. The last process steps didn't bring the big improvements in power consumption and/or clock frequency as they did in the past. They are also getting more expensive to produce, compensating at least for some of the area reductions. I think with 7 nm, things will slow down abruptly. New solutions are needed, like 3D transistors, something where cards are shuffled again. Only Samsung and Intel/Micron are showing progress in this with their memory business, not TSMC! I honestly don't know if it can be done with digital circuits in the same or similar way, but it might be an important step for Intel and certainly Samsung (already in market with this technology for a while).

To get to the point: Intel isn't out of the race. I still see TSMC as the weakest player of the big three. Intel needs to better compete on the design level, though, since processes won't cut it for them in the near future anymore. They should concentrate much better on foundry and mobile (modem integration, faster iterations, cost competitiveness) and, last but not least, should invest more in memory. I would like to see an investment in Micron with closer tie up between the two. Together, they can easily compete with Samsung (well, not probably not that easy, but still ...). The NAND market is about to grow huge, with PCs and data centers starting to convert from magnetic drives. I really should have shorted Seagate, damn it ...