Apple's self-made modem is a massive challenge, but with big rewards at stake
Apple's move towards creating its own modems could revolutionize how it produces mobile devices like the rumored Apple Glass, but only if it can match or exceed the performance of its existing modem supplier, Qualcomm.

Following an unexpected settlement with Qualcomm in 2019 to end patent infringement legal action, Apple has been a major client of the modem maker's products. However, with Apple working on its own modems, the days of using Qualcomm's communications hardware may be numbered.
In a Saturday profile of Apple's modem efforts, the Wall Street Journal outlines the challenge Apple faces in creating modems that are good enough for Apple to use instead of Qualcomm's version.
The rewards are numerous, including 5G in items like the MacBook Pro and extremely fast speeds for iPhones. For future hardware augmented reality headsets and smart glasses could benefit from the fast speeds, with the former needing high bandwidth but minimal latency to be worthwhile.
Apple's work in the field was bolstered by the acquisition of most of Intel's smartphone modem business and the onboarding of some 2,200 engineers, but Apple continues to expand its talent pool in the field.
Approximately 140 job postings based in an Apple office in San Diego, the hometown of Qualcomm, center around the creation of cellular chips. Meanwhile a satellite engineering office in Irvine, California has around 20 similar open positions, potentially to try and tempt employees of Broadcom over to the company.
Current expectations has Apple moving to its own modems from 2023, with TSMC expected to be the producer of the chips for the iPhone maker.
Making its own modem offers Apple advantages in a number of areas, including cost-savings and a reduction in reliance from suppliers like Qualcomm, which Apple has a tension-filled relationship with, CCS Insight senior director of research Wayne Lam told the report.
The ability to fine-tune the modem for Apple's intended purposes is also a big benefit, as it could adjust the modem to work in specific ways for one product, but in others to be more effective for another item.
As for what this could look like, the report points to Apple Silicon's impact, which involved high power efficiency chips that out-performed Intel's alternates. Lam offers the same sort of design improvements could improve the connectivity of smaller devices like the Apple Watch.
However, Tantra Analyst founder Prakash Sangam offers that "in some ways a modem is more complex" than a processor like the M1, in part due to the complexity of dealing with many circumstances that can affect a signal. This could make it harder for Apple to produce, increasing the relative development time.
"If you throw enough time and resources and money at it, it can be done," says Sangam. "But whether they can do it by 2023, I don't think anyone other than Apple can say."
Read on AppleInsider

Following an unexpected settlement with Qualcomm in 2019 to end patent infringement legal action, Apple has been a major client of the modem maker's products. However, with Apple working on its own modems, the days of using Qualcomm's communications hardware may be numbered.
In a Saturday profile of Apple's modem efforts, the Wall Street Journal outlines the challenge Apple faces in creating modems that are good enough for Apple to use instead of Qualcomm's version.
The rewards are numerous, including 5G in items like the MacBook Pro and extremely fast speeds for iPhones. For future hardware augmented reality headsets and smart glasses could benefit from the fast speeds, with the former needing high bandwidth but minimal latency to be worthwhile.
Apple's work in the field was bolstered by the acquisition of most of Intel's smartphone modem business and the onboarding of some 2,200 engineers, but Apple continues to expand its talent pool in the field.
Approximately 140 job postings based in an Apple office in San Diego, the hometown of Qualcomm, center around the creation of cellular chips. Meanwhile a satellite engineering office in Irvine, California has around 20 similar open positions, potentially to try and tempt employees of Broadcom over to the company.
Current expectations has Apple moving to its own modems from 2023, with TSMC expected to be the producer of the chips for the iPhone maker.
Making its own modem offers Apple advantages in a number of areas, including cost-savings and a reduction in reliance from suppliers like Qualcomm, which Apple has a tension-filled relationship with, CCS Insight senior director of research Wayne Lam told the report.
The ability to fine-tune the modem for Apple's intended purposes is also a big benefit, as it could adjust the modem to work in specific ways for one product, but in others to be more effective for another item.
As for what this could look like, the report points to Apple Silicon's impact, which involved high power efficiency chips that out-performed Intel's alternates. Lam offers the same sort of design improvements could improve the connectivity of smaller devices like the Apple Watch.
However, Tantra Analyst founder Prakash Sangam offers that "in some ways a modem is more complex" than a processor like the M1, in part due to the complexity of dealing with many circumstances that can affect a signal. This could make it harder for Apple to produce, increasing the relative development time.
"If you throw enough time and resources and money at it, it can be done," says Sangam. "But whether they can do it by 2023, I don't think anyone other than Apple can say."
Read on AppleInsider
Comments
Seems an odd statement to just hang out there.
Then the finished product has to actually play well with the deployed carrier infrastructure out there where Qualcomm and Huawei etc will have a major advantage, as both of them are actively involved in making that hardware as well as moving it forward (5.5G, 6G...).
Of course, financially, there is no getting away from paying patent fees to both of them in the process.
The reality is what it is. There is no getting away from that. If you want to live in denial, that is fine.
but, making a modern modem is very difficult. It’s not “just” a radio. It’s a very complex transmitter in both directions. And yes, it does have to meet numerous compliance standards around the world. There are also numerous frequencies it has to operate over. Then there are the antennas, complicated by the new hi bands. Power draw is a problem. A modern modem can draw as much as a high end SoC.
Apple never integrated the modems into their SoC as many Android manufacturers did, because Qualcomm never allowed it. Only Qualcomm;s SoC and modem could be mated together that way. Now, possibly Apple might be interested in doing that, though the major advantage of less power draw has never been a problem for Apple before.
but one thing is true. The very first modem Apple comes out with had better be about as good as the Qualcomm model it will replace, or Apple will never hear the end of it, even if in the real world, the performance is about the same. If there are just a couple of edge cases where it falls slightly behind, it will get ripped.
Apple can do anything it wants within the boundaries of the system that they own. When it comes to telecommunications Apple does not own the system, they are just one player in a much larger system.
I get your point about asserting that Apple undoubtedly has the technical chops and "smarts" to take on very complex technical challenges. But being smart is not enough. They also need problem domain and subject matter expertise, experienced and knowledgable staff ready to go, and design and manufacturing resources available "yesterday" that have been working towards solving the kind of highly specialized problems they are facing to build their own modem.
Apple can certainly grow or buy everything they need to get to where they need to be. However, it takes time and money and lessons learned along the way, exactly what this Apple Insider article is laying out in good detail, to get there. Simply having a bunch of really smart people on staff, all of whom are already heavily engaged in solving other big problems that need solutions, is not enough.
This is a high bar for Apple to get over. Simply being good enough or comparable to what they are getting from Qualcomm isn't going to cut it for Apple. Just like the M1 and Intel, they have to be significantly better to really make it worth the huge time, effort, and money needed to solve this with engineering versus solving this with business negotiation. Apple could negotiate with to Qualcomm cut its prices or give Apple more favorable terms for timing, deliverables, and volumes. Once Apple decides to take on the huge engineering effort on their own this becomes a case in burning their (Qualcomm) boats. It had better be worth it, especially without Intel to fall back on this time around.
You are the one in denial like Intel, Apple has the money time, and talent, to git it done and they will, it took 13 years to kick Intel’s ass, it won’t take that long for Qualcomm.
The first modems that Apple fabs will go into Mac product lines, not iPhones, and given that there are no modems at all for Mac's, the complaints, if any, will fall on deaf ears. The reason for this is quite obvious; Apple can support a few millions of modems a month for Mac's, and they won't have to be as energy efficient for the first pass, as an iPhone would require.
As for competing with Qualcomm on modems, Apple has a number of years left in its agreement with Qualcomm, so given Apple's historic advances in just about all technologies, I'm not seeing the risk that you are.
avon b7 said:
Here you go again, bringing Huawei into the conversation.
Perhaps you should mention that Huawei doesn't actually have any current Kirin processors being fabbed. I'd speculate as well that Ukraine will be the example to all of the West of why you don't use an adversary for critical infrastructure, and Huawei's close ties to the PRC are abundantly clear. On top of that, Finland and Sweden are likely to join NATO this month, and both of those countries are leaders in telecom.
Oddly, you have never, ever, agreed with me that Telecom is a National Security issue; here's your chance...
What are not national security issues are the vendors of telecom infrastructure.
All telecom infrastructure is standards based and cannot reach the market without certification. Security is part of that certification. In the case of Huawei, it has at least three 'transparency' centres around the world dedicated to security. On top of that its code is available for inspection. It even offered the US its entire 5G stack under licence.
Huawei has installed and manages some of the world's most important undersea communications cabling. Three such lines (from three continents) hit mainland Spain and that is one of the reasons why Huawei is creating more data centre hubs in Spain.
Anyone wishing to inflict significant harm on the world's internet traffic would only have to take out undersea cables. Something that most major states can do without too much effort.
In terms of actual network operations though, it is the carriers, not the vendors, which manage everything. Any attacks on networks are not dependent on infrastructure vendors and are equally 'vulnerable' to any party with an interest in breaching them.
Infrastructure vendors are not the problem and never ever have been. That has been demonstrated by decades of interoperability and, in the case of Huawei specifically, not a single major breach to its name nor evidence of any problems.
Of course, none of that has anything to do this thread. Neither does Kirin so you can let all that go.
I mentioned Huawei because it has key patents. It produces CE and industrial 5G products within a wide domain of activity. It is relevant in terms of competition and the bars that are raised.
Obviously Qualcomm too as it is discussed in the article even though it doesn't ship end user products.
In patent terms, it was rumored a few years ago that Apple licences almost 800 patents from Huawei and that Huawei licences around 80 from Apple.
With the purchase of Intel's modem division Apple suddenly has a chest of around 17,000 wireless patents to use.
It is already squaring up to Ericsson in court in a patent dispute (which it could lose) and no doubt will use that influence to try and bring down its patent licencing fees.
That won't change the difficulty surrounding Apple's efforts though as I detailed above.
FWIW, I've always said it was the right way to go even if they are very late to the party and have had to 'buy' almost everything in (and basically with no alternative because it landed itself in a strategic planning nightmare with the Qualcomm fiasco).
Yeah, of course, let's have a PRC company, Huawei, manage undersea communications; what could possibly go wrong?
Facebook/Microsoft have a joint 5,000KM cable down there too and there are plenty more so please don't worry.
Now, would it be asking too much to get back on topic?
Well, one of the topics you brought up is undersea communications, and I reiterate that having a PRC company, Huawei, as a gatekeeper is fraught with risk to the West. Yeah, it's about national security, the same as allowing Huawei into the telecom chain of the EU; fraught with risk, but more to the point, China's mercantilism in the early stages of 5G buildout war also reason enough to restrict Huawei telecom.
Thankfully, the high water market for Huawei in the West is in the distant past.
https://www.cnbc.com/2022/04/28/huaweis-first-quarter-revenue-tumbles-as-smartphone-sales-plunge.html
https://www.lightreading.com/opticalip/fttx/the-slow-death-of-huaweis-european-fixed-business-has-begun/d/d-id/773242