If Apple really want to be the good guys they claim to be they need to start making their technologies more open.
So what if the make the worlds best processor technology and lock it to macs and iPhones?
So what if they have the most private messaging system, locked to iPhone
So what if they have the best/most secure App Store only available on iPhones
and same goes for a modem
It would be nice if they acted how they preach and improve the worlds technology, but then they should still be able to make the iPhone the best place to enjoy it
This is probably a bigger deal than most people realize. The full communication stack for any given protocol includes much more than the modem. The modem is just one (or a couple) of the layers in the comm stack. Using someone else’s modem, one that’s designed for broad market applications by multiple vendors means having to conform to interfaces, timings, and state machines the modem vendor provides - the good, the bad, and the ugly.
Apple will still have to conform to the “wire” protocol for all of the communication standards their modem must support. That is non-negotiable. But by building more of the comm stack layers themselves Apple will be able to provide optimizations and tailor their allocation of functionality between hardware, firmware, and software to improve the efficiency and performance of their modems to eke out the very best performance possible for their host systems.
This is yet another implementation of a strategy that Apple exploited with the M1. Optimizing for Apple’s specific needs and strategies is always going to be better for Apple than conforming to a generic interface intended to serve a wide range of customers.
This is exactly what Huawei has been doing for almost a year with WiFi 6 (which for Huawei devices is called WiFi 6+ because it takes 5G technology and rolls it into its WiFi 6 products.
It does exactly the same with Bluetooth in areas like audio to offer better quality to its own devices, not with 5G technology though.
It also converges WiFi and Bluetooth for certain system features.
All the while keeping its products certified for regular 5G, WiFi, Bluetooth etc.
Once it has the capability, I agree it would make sense for Apple to enhance it to fit its own products.
In a more limited way, it is not dissimilar to Wi-Fi and AirPort years ago although it was using technology from other companies.
In its day AirPort was a revolution.
Thanks for this post. This is getting closer to what I’d like to understand better — how will apple making their own modems lead to a better product?
I can understand in general terms that wireless communications is super important, there are lots of protocols, integration can be good, etc. But I don’t have as clear a sense of the tech in this area, what apple can do better than, Qualcomm, and how this will lead to better products. I’m eager to learn more about this!
It was coming. When Apple signed that 8 years deal with Qualcomm 5G tech, whole point was using Qualcomm 5G chip in iPhone 12 and later embed 5G tech onto Apple's own SOC.
What we don't know is what is going on behind the scenes between Apple and Qualcomm.
It sounds strange that Apple would have swallowed their pride and kissed and made up with Qualcomm so they could use Qualcomm modems for only a single year.
So, that leaves two possibilities:
1) The Apple-Qualcomm reunion didn't work and Apple is going its own way on their own as quickly as possible -- like next September
2) This is, as has always been thought, a long term project much like Apple Silicon replacing the Intel CPU's.
The iPhone 12 reportedly does not have the latest and greatest of Qualcomm modems -- that may have influenced this and its timing. That is: Did Qualcomm stick it to Apple and pawn off an inferior modem on them?
Apple has Qualcomm’s latest and greatest modem, the x55. The x60 won’t be in new devices until mid 2021, a few months before apple comes out with their new phones. It’s the timing of product cycles, nothing more.
I don't think that is a correct statement to make. SD 888 powered Android phones with x60 modem (Samsung Galaxy S21 series, MI 11 pro etc) will be out in Jan-2021 itself and available for customers. So, majority of the Android flagship phones in 2021 will have a better 5G modem than iPhone 12 series phones for at least 6 months (even the ones which launch in Mar-2021).
Most of those phones won’t actually be on sales for a time after they’re announced. But Samsung sometimes announces phones three months before they become available. They’re no the only ones. At any rate, it doesn’t make a difference. It will still be months.
Feels like they should have started 4 years ago, but good to hear they are committed. They could have switched Macs to Apple Silicon in 2018. And their own cellular modems could have been started in 2016 or 2017. They basically wasted billions and 4 years to groom Intel to be a secondary or primary modem provider, but Intel has poorly performed in virtually everything they have touched the past 3 to 4 years.
I don’t think it’s that simple. Apple knew that Intel was starting years behind in R&D and patents. I think that Intel did a pretty good job considering that starting position. In fact, if 5G had come in one year later, it’s possible that Intel would have made Apple’s 5G modem.
it’s only because Qualcomm now seems to have been forced to license its essential patents as FRAND, as they should have been made to do years before, that Apple can even do this.
Yes, it's always not clear cut. Not producing a 5G modem for Apple on time was likely the reason the division was put up for sale. I think Intel's dabbling with cell modems ended up being longer than their effort with ARM processors, for maybe an interesting bit of trivia (don't recall when Intel's first StrongARM shipped though).
Apple likely felt it less risky to have Intel develop modems than themselves way back when. Intel's starting position wasn't that bad, and they had manufacturing advantages which by 2018, ended up never being used for their phone related chips. Intel bought Infineon's cellular division in 2011 if you recall, so they weren't starting from zero. Infineon modems were shipping in iPhones all the way up to the pre-Verizon iPhone 4. So, the skills were there, but Intel never ended up shipping any of their Atom and modem chips on their state of the art fabs, thus negating the real advantage they had.
Given this, Apple buying Infineon's cell division would perhaps be the proper course of action if we could rewind the clock. It wouldn't have stopped a Qualcomm vs Apple licensing war, but it would have meant Apple being responsible for the other very important piece of silicon in iPhones. In the grand scheme of things, the product ecosystem is not ready to have modems in every product just yet, but it's coming. Like having an LTE modem in an AirPods Max or iPod could support an Apple services subscription (music, assistant). Obviously, Mac laptops could have them. So, having a wholly own, specifically develop cellular chips for Apple's products is going to be really important. Having the development machine for modems working well in the past few years would be better than rushing it over the next 3 to 4 years.
I agree with all of that. Additionally, Apple like other manufacturers, can’t afford to fall behind too much in any area that actually matters. Screen refresh rates, for example, are a totally trivial matter for Apple devices, while they’ve been almost critical for Android devices, so Apple can wait until the low power screens are ready, screens that Apple actually invented, but which Samsun can’t make enough of yet for Apple’s use.
‘’but modems are more important, at least in that one year means nothing, but after that it does, people may remember coming into 4G when Apple also decided to wait. I remember posts on Arstechnica that were not complementary, and where some predicted the end of Apple’s phone market. One genius said that the problem was that Apple users were too “lazy to manage their phones”. I asked why one should have to do that, and the answer was a total jumble, similar to what we get here from a couple of members.
Then next year Apple went to 4G, and it was forgotten. The same thing here too. Apple hopefully will have their own modem, but what’s the rush? Too many people are making a big deal out of this. Ironically, if there was no mention from third party wags or Apple, that they would be making their own modems, this entire silly discussion would never have happened.
Feels like they should have started 4 years ago, but good to hear they are committed. They could have switched Macs to Apple Silicon in 2018. And their own cellular modems could have been started pin 2016 or 2017. They basically wasted billions and 4 years to groom Intel to be a secondary or primary modem provider, but Intel has poorly performed in virtually everything they have touched the past 3 to 4 years.
Lol, right. I mean really they should have started 8 years ago so they could switch Macs to Apple silicon in 2014. And their own cellular modems could have been started in 2012 or 2013. Making a technological roadmap after the fact is a breeze. I wonder if there are any paying jobs in that field.
2018 is when TSMC caught up to Intel in fab. That would have been the perfect time to make the switch as Apple could have rode a six year tick tock cycle of TSMC going from 10nm, 7nm, 5nm. And if they are on cadence with 3nm, 8 years. It would have been relentless march. Before that, Intel had the best fabs, and the best decision was to stay with Intel.
Apple’s chips weren’t ready in 2018. You think you know more about it than Apple? And the question about fabs isn’t so direct either. TSMC, Samsung and others have been using very relaxed, as they call it, manufacturing rules for this. Intel’s problem wasn’t so much that their fans were behind, as it was that they refused to do the same, struggling with something others gave up on. I’m not saying that Intel was doing the right thing.
But a couple of years ago Intel made a proposal that instead of using process node, which these others aren’t totally honest about, they should use transistor density. That’s a much more useful way of measurement, as it actually accounts for all the manufacturing points. But the others said no. Why, because it takes the marketing advantage of using smaller numbers, which they can do, and using bigger numbers instead, which they can’t.
‘’admittedly, Intel has screwed itself because of their stubbornness. But it’s not really so much a matter of technical superiority, as it was then trying to make 10nm transistors in a 10nm space, which was something the others had given up on before.
Apple was ready with the A12X. An A12X in the same set of Macs that have the M1 today would have made very very good Macs in 2018. It wouldn't have been the 1.5x, 2x, 3x advantages that M1 has over the Intel chips that it replaced, but it would have been 1.2x to 2x. The 2018 and 2019 MBA shipped with a dual core i5-8210 chip. The A12X crushes it in CPU and GPU, and it would have been fanless.
What wasn't ready was probably the software. But, if they wanted to switch to Apple Silicon in 2018, they would have started dropping 32 bit software sooner, started trying to get everyone to used Xcode toolchains more, ie, getting the ecosystem prepared earlier. Apple's biggest weakness on the PC side is 3rd party software. This is where they really have to push.
For what to call fab nodes, what does that matter? Intel has to ship. Even they don't care about transistor density anymore, and they are seemingly obfuscating their 10nm transistor density numbers. Pretty darn hard to find transistor density values for Ice Lake and Tiger Lake chips. If you know them, let me know. And yes, we should be mindful of the density numbers of each foundry's node. You already know TSMC has lapped Intel and Samsung will basically catchup if not surpass Intel a little in a few months.
My point was 2018 was the best time for the switch because Apple would have added a couple more years to the cadence. When you can progressively improve the hardware across 6 to 8 years, it creates a branding advantage all on its own. It takes time to convince people of things. A continuing progression of year over year improvements does that.
When we look at the performance of the 12z in the developer machines, we see pretty good, but not great performance. I think Apple did the right thing. By waiting, they have a chip that well surpasses machines in the price and size range. Instead of people thinking that’s nice, but it’s not really a big deal, the response has been that it IS a big deal.
but of course, you’re right too. The software likely wasn’t ready either. Apple wanted to, or had to normalize things between the iPad and the Mac with Catalyst too, which was pretty poor in its first iteration. Other problems existed as well. But the increased chip performance enables those x86 apps to run faster too, and that encourages people to buy in now rather than to wait, if they all ran slowly.
If Apple really want to be the good guys they claim to be they need to start making their technologies more open.
So what if the make the worlds best processor technology and lock it to macs and iPhones?
So what if they have the most private messaging system, locked to iPhone
So what if they have the best/most secure App Store only available on iPhones
and same goes for a modem
It would be nice if they acted how they preach and improve the worlds technology, but then they should still be able to make the iPhone the best place to enjoy it
No. They don’t have to make their technologies open to everyone. Nobody expects them to either.
This comes to no surprise. I'm sure Qualcomm is readying its army of lawyers ready to accuse Apple of stealing its IP.
Not how essential patents and licensing works. Qualcomm HAS TO license its IP. That's part of the deal when your IP is included in a standard.
Well, wasn’t that the problem before? They predicated licensing on other contracts and fees?
The way it works is kind of odd. You can have “essential” patents that you don’t declare as being essential. So you don’t have to license them out unless lawsuits end up forcing you to. But if you declare a patent as essential, you’re effectively stating that you intend to license them. Then, you’re supposed to license them as FRAND.
“Apple sued the company in 2017 over patent licensing fees, an action that sparked a worldwide legal scrum that was settled in April.”
“Settled” is one way of putting it. Apple was resoundingly trounced by Qualcomm, ended up making a significant payment to clear outstanding royalties, and got their licensing fee jacked up to an even higher amount than they were paying before.
That lawsuit was one of the most lucrative in Qualcomm’s history.
What we don't know is what is going on behind the scenes between Apple and Qualcomm.
It sounds strange that Apple would have swallowed their pride and kissed and made up with Qualcomm so they could use Qualcomm modems for only a single year.
So, that leaves two possibilities:
1) The Apple-Qualcomm reunion didn't work and Apple is going its own way on their own as quickly as possible -- like next September
2) This is, as has always been thought, a long term project much like Apple Silicon replacing the Intel CPU's.
The iPhone 12 reportedly does not have the latest and greatest of Qualcomm modems -- that may have influenced this and its timing. That is: Did Qualcomm stick it to Apple and pawn off an inferior modem on them?
Apple has Qualcomm’s latest and greatest modem, the x55. The x60 won’t be in new devices until mid 2021, a few months before apple comes out with their new phones. It’s the timing of product cycles, nothing more.
I don't think that is a correct statement to make. SD 888 powered Android phones with x60 modem (Samsung Galaxy S21 series, MI 11 pro etc) will be out in Jan-2021 itself and available for customers. So, majority of the Android flagship phones in 2021 will have a better 5G modem than iPhone 12 series phones for at least 6 months (even the ones which launch in Mar-2021).
Most of those phones won’t actually be on sales for a time after they’re announced. But Samsung sometimes announces phones three months before they become available. They’re no the only ones. At any rate, it doesn’t make a difference. It will still be months.
I think you are mistaken on this topic. Samsung galaxy S21 with SD 888, Qualcomm x60 modem is set to launch on Jan-14-2021 and will be available for sale to customers in the same month. Considering this, iphone 12 series phones will have the inferior x55 modem for bulk of its sale period compared to its key competition (Samsung galaxy S21 series).
What we don't know is what is going on behind the scenes between Apple and Qualcomm.
It sounds strange that Apple would have swallowed their pride and kissed and made up with Qualcomm so they could use Qualcomm modems for only a single year.
So, that leaves two possibilities:
1) The Apple-Qualcomm reunion didn't work and Apple is going its own way on their own as quickly as possible -- like next September
2) This is, as has always been thought, a long term project much like Apple Silicon replacing the Intel CPU's.
The iPhone 12 reportedly does not have the latest and greatest of Qualcomm modems -- that may have influenced this and its timing. That is: Did Qualcomm stick it to Apple and pawn off an inferior modem on them?
Apple has Qualcomm’s latest and greatest modem, the x55. The x60 won’t be in new devices until mid 2021, a few months before apple comes out with their new phones. It’s the timing of product cycles, nothing more.
I don't think that is a correct statement to make. SD 888 powered Android phones with x60 modem (Samsung Galaxy S21 series, MI 11 pro etc) will be out in Jan-2021 itself and available for customers. So, majority of the Android flagship phones in 2021 will have a better 5G modem than iPhone 12 series phones for at least 6 months (even the ones which launch in Mar-2021).
Most of those phones won’t actually be on sales for a time after they’re announced. But Samsung sometimes announces phones three months before they become available. They’re no the only ones. At any rate, it doesn’t make a difference. It will still be months.
I think you are mistaken on this topic. Samsung galaxy S21 with SD 888, Qualcomm x60 modem is set to launch on Jan-14-2021 and will be available for sale to customers in the same month. Considering this, iphone 12 series phones will have the inferior x55 modem for bulk of its sale period compared to its key competition (Samsung galaxy S21 series).
The timescale is pretty correct. Yes, it could be a month either way, but it still wont actually be widely available for a time after the announcement.
and realistically, the differences between the modems in real world performance won’t be noticeable to most people.
I'm guessing Apple is at the point now with some development milestone that they are confident enough to announce their intentions or plans. I can assure you, they didn't just decide to kick this off this year. They didn't buy Intel's modem division on a whim either. They had staff dedicated to these even before the Intel purchase.
“Apple sued the company in 2017 over patent licensing fees, an action that sparked a worldwide legal scrum that was settled in April.”
“Settled” is one way of putting it. Apple was resoundingly trounced by Qualcomm, ended up making a significant payment to clear outstanding royalties, and got their licensing fee jacked up to an even higher amount than they were paying before.
That lawsuit was one of the most lucrative in Qualcomm’s history.
Yup, it was indeed a short term windfall for Qualcomm. However, in the long run, it's going to be a major loss for Qualcomm and this is entirely due to Qualcomm's greed and not doing the right thing.
Comments
So what if the make the worlds best processor technology and lock it to macs and iPhones?
So what if they have the most private messaging system, locked to iPhone
So what if they have the best/most secure App Store only available on iPhones
and same goes for a modem
I can understand in general terms that wireless communications is super important, there are lots of protocols, integration can be good, etc. But I don’t have as clear a sense of the tech in this area, what apple can do better than, Qualcomm, and how this will lead to better products. I’m eager to learn more about this!
I agree with all of that. Additionally, Apple like other manufacturers, can’t afford to fall behind too much in any area that actually matters. Screen refresh rates, for example, are a totally trivial matter for Apple devices, while they’ve been almost critical for Android devices, so Apple can wait until the low power screens are ready, screens that Apple actually invented, but which Samsun can’t make enough of yet for Apple’s use.
‘’but modems are more important, at least in that one year means nothing, but after that it does, people may remember coming into 4G when Apple also decided to wait. I remember posts on Arstechnica that were not complementary, and where some predicted the end of Apple’s phone market. One genius said that the problem was that Apple users were too “lazy to manage their phones”. I asked why one should have to do that, and the answer was a total jumble, similar to what we get here from a couple of members.
Then next year Apple went to 4G, and it was forgotten. The same thing here too. Apple hopefully will have their own modem, but what’s the rush? Too many people are making a big deal out of this. Ironically, if there was no mention from third party wags or Apple, that they would be making their own modems, this entire silly discussion would never have happened.
When we look at the performance of the 12z in the developer machines, we see pretty good, but not great performance. I think Apple did the right thing. By waiting, they have a chip that well surpasses machines in the price and size range. Instead of people thinking that’s nice, but it’s not really a big deal, the response has been that it IS a big deal.
but of course, you’re right too. The software likely wasn’t ready either. Apple wanted to, or had to normalize things between the iPad and the Mac with Catalyst too, which was pretty poor in its first iteration. Other problems existed as well. But the increased chip performance enables those x86 apps to run faster too, and that encourages people to buy in now rather than to wait, if they all ran slowly.
The way it works is kind of odd. You can have “essential” patents that you don’t declare as being essential. So you don’t have to license them out unless lawsuits end up forcing you to. But if you declare a patent as essential, you’re effectively stating that you intend to license them. Then, you’re supposed to license them as FRAND.
“Settled” is one way of putting it. Apple was resoundingly trounced by Qualcomm, ended up making a significant payment to clear outstanding royalties, and got their licensing fee jacked up to an even higher amount than they were paying before.
and realistically, the differences between the modems in real world performance won’t be noticeable to most people.
https://www.theinformation.com/articles/apple-developing-modem-as-chip-efforts-expand
https://www.macrumors.com/2018/11/15/apple-hunting-qualcomm-headquarters/
https://arstechnica.com/tech-policy/2019/07/apple-poised-to-acquire-intels-smartphone-modem-business-report/
I'm guessing Apple is at the point now with some development milestone that they are confident enough to announce their intentions or plans. I can assure you, they didn't just decide to kick this off this year. They didn't buy Intel's modem division on a whim either. They had staff dedicated to these even before the Intel purchase.
The market seems to agree as Qualcomm's stock took a beating on the news.
https://www.barrons.com/articles/the-stars-align-for-emerging-market-growth-stocks-in-2021-51608283838
"Qualcomm Stock Tumbles on Report Apple Will Build Its Own Modem Chips"
This actually seems odd to me, because it was hardly an industry secret that Apple was working on their own modems.