iPhone 12 and the Reality of Qualcomm 5G Modems

The lawsuits are over. The drama has ended. It’s 2020 and Apple and Qualcomm have finally buried their legal hatchets — at least enough to work together on bringing 5G modems to the iPhone 12 this fall.

Almost certainly. We think. Look… listen.. it’s complicated. And here’s why.


Now, to understand the iPhone, you need to understand one very important thing about Apple.

Ok, a couple of very important things, but only one of them really matters right now:

Apple is super conservative when it comes to battery life.

That might sound obvious, but its effects are both subtle and frustrating. Because, it means anything that affects battery life in any way. Anything that radiates. Light, like the display. Heat like the chipset. And, yeah, radio waves like the… radios. Cellular, Wi-Fi, Bluetooth, all of them.

Are just beyond super tightly controlled.

It’s why, once Apple got to Retina, they became far less aggressive about display density and started focusing on the more energy neutral aspects of display quality.

It’s why they started making their own chipsets, so they could balance performance with efficiency cores, and make the graphics wider instead of just faster.

It’s why they adopt new Wi-Fi and Bluetooth standards so quickly, so they can maintain low-energy states as much as possible, and better race to sleep when it just isn’t possible. Which means transfer as fast as they can and then shut down as fast as they can to save as much power as they can.

But it’s also why they’re slower to adopt new cellular standards. Because the early versions of those modem chips are typically bigger and run hotter, on networks that have far less deployment, that if hit by hundreds of thousands of iPhones suddenly coming online all at once, would just fall over like Megatron hit by Optimus Prime in full on Mack truck mode.

Now, I know what you’re already typing, why doesn’t Apple just make up for the extra power draw by making the iPhone twice as thick and squeezing double the battery cells into it like… a jelly donut?

Well, that’s the second really important thing to understand about Apple, and if you want to see a video on that, let me know in the comments below.


Never mind 5G in 20-20, the original iPhone didn’t even support 3G in 20-oh-7.

It used modems by a German company named Infineon Technologies, which spun out of Siemens AG. And, all it had to do at launch was carry GSM and EDGE data for AT&T in the U.S.

GSM, or Global System for Mobile Communications, is what most of the world was using. Most of the world, but not Verizon or Sprint in the U.S. Also Bell and Telus in Canada but they would eventually switch to GSM while Verizon and Sprint… would decidedly not.

Verizon and Sprint used and stuck to CDMA, or Code-Division Multiple Access. It was a technology that allowed for far fewer towers to cover far more distance and serve far more people, so even though it had a host of other problems, it’s what they went with to build their networks out bigger and cheaper.

What that meant was, once Apple had sold iPhones to everyone willing and able to use AT&T, in order to keep growing, they had to add support for Verizon.

That meant adding CDMA… and that meant dealing with Qualcomm.


In order for your technology to become part of a standard, you have to agree to license it in a fair, reasonable, and non-discriminatory way — what’s called FRAND.

Which means, in theory, anyone capable of spinning a chip should be able to license and implement it to spin exactly that chip. That’s how we get and preserve competition.

In practice, though, Qualcomm had CDMA wrapped up so tightly, there was just no way not to deal with them.

That’s why the iPhone 4 variant designed for Verizon, CDMA, and EVDO data, used a Qualcomm modem.

As Apple continued to expand iPhone carriers and eventually went to 4G LTE with the iPhone 5s, those Qualcomm modems were a huge benefit not just in terms of performance but also compatibility.

We got the first iPhone that was truly a world phone in that, if you bought it or could get it unlocked, you could use it almost anywhere in the world.

But it also came at a huge cost — specifically Qualcomm’s huge fees.

See, Qualcomm considered their modems so important to the success of any phone that they demanded not just a regular fee for the modem, but a fee based on the whole phone.

Now, you can just imagine if every important component in a phone wanted to do this, like — gotta have a screen, that’s a percentage! Wi-Fi, yeah, also a percentage. You want to plug a cable into it, right? Percentage! Processor? Oh, wait, have you seen how much Qualcomm is charging for those now as well? Galaxy S20 prices say hiiiiii.

I joke, but if every component vendor acted that way, every manufacturer would run out of percentages before they made a single phone.

But Qualcomm acts exactly that way because they’re not mostly or even mainly a modem or chipset company. They’re mostly and mainly a patent licensing company.

That meant Apple was paying those hefty, hefty modem fees — which meant we were paying those hefty, hefty modem fees — on every iPhone. Yes, even on iPhones that weren’t and would never be used on CDMA in the U.S.

And Apple, like Sauron does not share power. I mean profits. Profits.

So, Apple started trying to disentangle themselves from Qualcomm.

Infineon II

At around the same time Apple was switching to Qualcomm, Infineon was bought out by Intel — yeah, that Intel — and became Intel Mobile Communications. IMC.

The initial plan was to replace Qualcomm on all the non-CDMA phones. In other words, even in an increasingly LTE world, all the phones that didn’t need to still run on Verizon or Sprint’s legacy networks in the U.S.

Now, Qualcomm, of course, had a ton of LTE patents, and they treated FRAND there pretty much how they treated FRAND everywhere — in an unfair, unreasonable, and quite discriminatory way.

So Intel and Apple had to work around Qualcomm’s patents as best they could.

And that resulted in an iPhone 7 rollout where the GSM phones all had Intel modems that.. just didn’t quite work as well as the CDMA models that still had Qualcomm modems. Especially in areas with more obstacles or weaker signals.

But, for Apple, that didn’t really matter. They said, in order to get on Apple’s Olympic modem team, all you had to do was sprint 100 meters in under 10 seconds. Didn’t matter if Intel could do it in 9.7 and Qualcomm in 7.9, both were just had to be under 10 seconds.

And, it turned out, it also didn’t matter that Qualcomm could do it on pavement, grass, or mud, and Intel… not so much.

If it meant not having to pay the exorbitant licensing fee Qualcomm demanded, that’s what Apple was going to do.

Up to and including the iPhone 11, which was all-in on Intel modems, with not a dollop of Qualcomm inside or in sight.

Much better Intel modems as well… if still not quite as good as Qualcomm.

Qualcomm II

Now, while Apple and Intel were busy making modems for the iPhone, Apple and Qualcomm were busy trying to sue the pants off each other over Qualcomm’s licensing terms and Apple helping Intel make those modems.

Also, 5G was on its way, and Qualcomm had gotten every bit the stranglehold over that technology that they’d gotten over the ones before. More even.

And just as importantly, Qualcomm had a huge lead in terms of actually making modems that worked on the various implementations of 5G.

As Apple and Qualcomm battled it out in the courts, Apple and Intel battled it out in the labs but, in the end, it just became 100% crystal clear that there was no freaking way Apple and Intel would be able to deliver a 5G modem anywhere nearly as good as Qualcomm’s or anywhere nearly on time… if at all, like ever.

That meant, Apple needed Qualcomm and Qualcomm still wanted Apple’s money, and neither really wanted to risk a decision going the other way.

So, both decided to bury their very large legal hatchets and sign a new deal that would ensure 5G support for the iPhone.

Just in time for the iPhone 12.


Now, the story doesn’t end here. When Intel lost Apple’s modem business, they kinda also lost… their modem business. Which is why it came as absolutely no surprise to absolutely anyone when Apple bought that business from Intel shortly after settling with Qualcomm.

So, now it’s Apple’s modem business. Of course, the modem is only part of the solution. There’s also the RF front end, the antennas, and more.

Apple already makes their own antennas. Perhaps you remember them from the iPhone 4 and… free bumpers? But we’ll have to see what happens with the RF front end.

But Apple didn’t just make a deal with Qualcomm for their modems, they made a deal to license the underlying technologies.

So, in a couple or few years, Apple could have their own, custom modems ready for the iPhone 14 or iPhone 15, alongside or as part of their custom silicon, the A14 or A15.

iPhone 12

But that’s then, this is now, and Apple and Qualcomm are both full steam ahead just trying to get 5G modems into the iPhone 12 in time for it’s launch later this year.

That could mean Qualcomm’s X55 modem, which is what we’ve seen in recent 5G Android phones. But, there have also been some rumors saying, if time and yield allows, it could also be the next-generation X60, which will apparently be fabricated on TSMC’s new 5 nanometer process, just like Apple’s next generation A13 system-on-a-chip.

That would certainly play into Apple’s more conservative modem strategy, where they prefer to wait and let other manufacturers and phones suffer with the quirkier, less power-efficient, more battery devastating early-generation chipsets. And carriers’ load strategies, where they prefer to keep the huge volumes of iPhone users off next-generation networks until they’re deployed enough to handle all the added traffic. Because no one likes a network when it’s continuously being taken down.

Now, I still haven’t even touched on the mess that is 5G itself. Sub-6 vs. Sub 9 or FR1 vs. mmWave and F2, and whether that’s even ever going to be a real commercial things, so let me know in the comments if you want a video on that as well.