Always connected PC, the Qualcomm initiative pushing towards a PC that can last up to 20jsh or hours that is always connected and can sync emails and social waste of time services, meant that in late 2018 Intel presented us a Sony Cove and a 10nm contender that will get you always connected 20ish hours of battery.
This is what competition brings, and Intel only started to seriously think about investing in modems when an Apple offer came to the table. Apple wants multiple sources, this is nothing extraordinary to the computer industry. If one has two or more suppliers, you can make them compete for the deal and essentially get a better price for the device.
Companies with volumes of Apple want to squeeze every cent. The irony is that a 1200 Euro phone has a Bill of Material of $400ish - or roughly one third of the price.
Let’s take, for example, the FTC trial revelation that Apple wanted to pay $1.5 per iPhone instead of the previous $7.5 license. Assuming that Apple sells 45 million phones per quarter. This simple match $1.5 dollar per phone could earn Qualcomm $67.5 million - a far cry from $337,5 million. A modem fee comes on top of this number and it is rumored that it can be as high as $30 per device.
Modem generation cost $2 billion
Based on our investigations, it costs at least $2 billion dollars in R&D to make a modem. For Qualcomm it might be a bit easier to digest the modem fee as it uses the technology in its Snapdragon SoCs that it sells to hundreds of customers.
For Intel, the $2 billion modem commitment is based mainly on keeping Apple happy. It is unlikely that Intel is making money on that deal, and is probably not even close to breaking even. Intel tried and failed to make a phone SoC as the management at that time was to0 stubborn to drop X86 and take the ARM license - obviously necessary to build a decent phone SoC.
Intel didn't want to invest in 4G
Another decision confirmed by now ex Intel Aicha Evans, not to invest in 4G at time was a good business decision. Intel had the money, Intel had more than 100,000 employees and could definitely develop a decent modem if it wanted to. Back in the earlier days of 4G, Intel decided that this would not be a great business opportunity. Bear in mind that the deceased Paul Otellini decided to pass on the opportunity to make a chip for what would be called the iPhone. The PC and data-centric company didn't see much sense in going mobile, but, of course now it does. Intel is heavily investing in 5G as it sees that it needs to be a player in edge computing and the "always connected" world.
It was ironic to see that Aicha Evans called Qualcomm a great technology company and she said that "They've [Qualcomm] earned what they have". Part of the testimony has Aisha claiming that she "always been super respecful of Qualcomm". and she even called Qualcomm the "number one and undisputed leader. She also mentioned that back in 2012 when it acquired Infineon, Intel knew that it would lose Apple iPhone deal due to the lack of LTE. Qualcomm invested and led the development of LTE and made the choice for Apple easier. Intel invested billions of dollars and engineers to do the 4G / LTE catch up.
PC vendors like Dell, HP, Lenovo, Acer and many others will start adding modems to their laptops at least in the high end and later mainstream, but it might take years to get these units in decent numbers. The PC market is not where Intel will make money on 4G and soon 5G. A PC user in 2019 - especially the millennials - expect that a device comes with always connected access, otherwise it is quite useless as it won't tweet, Snapchat or Uber to name just a few. Parents of millennials are using Facebook and that needs connectivity too. Apple in 2017 wanted Intel modems to get to 600 Mbps instead of 450 Mbps but Aishe testified that this could not have been done in the desired time frame.
CDMA from VIA was must for Apple deal
The FTC trial also revealed that VIA telecom acquisition in 2015 was made with Apple in mind. Aische itself made a presentation stating: "Add 6-Mode CDMA solution to win Apple" calling it the last piece of the puzzle. To put things in perspective, without CMDA Intel could never replaced Qualcomm as Apple is the sole supplier as Verizon in US and China Telekom - the two extraordinary large and important carriers use CDMA.
When asked why Intel acquired VIA instead of developing CDMA in house, Aische replied “I mean, you need experience, you need talent, expertise, and at this point basically building it from scratch was never -- we were not going to make it". Infineon and Intel lost the Apple deal for not having LTE, a piece of crucial technology and Apple went solo to Qualcomm models until 2017.
Great 4G is necessary for successful 5G
As a pure wild crystal ball speculation, making a modem for Apple phones is good practice for Intel automotive 5G solutions and future 5G enabled PCs. This part is clear and logic sounds rock solid and Intel has a decent chance to have better sucess in 5G than in 4G. The only downside is that based on its 4G introduction it looks like Intel is at least 1.5 to 2 generations behind Qualcomm and - based on Intel's announcements - Chipzilla will be late to 5G too. The irony is that Intel will be betrayed by its big customers as Cupertino boys are expected to make its own modem and join the modem club that includes Huawei and Samsung, its biggest threat and competitors.
Another speculative point is that Apple is not using its SoCs for its laptops to keep Intel happy, and only time will tell. So blaming Qualcomm for success in the modem market is obviously a business decision and a path that Intel refused to take. It would be as ironic as Qualcomm blaming Intel for dominance in the X86 and overall PC market, a market where Qualcomm now slowly enters and sees the challenges - but, compared to Intel, Qualcomm invested dramatically less.