Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
I don’t understand why building a modem is so challenging. I wonder if it is more of a patent legal obstacle course rather than a technical hurdle.
Modems need to take very faint, very noisy analog input that often bounces and overlaps with itself and other signals, digitize that and interpret the results. It has to do that quickly, reliably, without burning through batteries and without costing $100/chip. It also has to not step on patents owned by Qualcomm.

Easy Peezy! 🤣
 
We knew about Apple Silicon in 2018. M1 launched two years later.

Apple bought Intel's 5G modem business in 2019. It's four years and counting. Five if you consider Apple has put cellular job postings since 2018.


The M1 was genuinely a culmination of ~10 years experience of in-house full CPU mircroarchitecture design. Apple had in fact considered switching the Mac from PowerPC directly to Arm (thus no Intel era at all). It was Jobs that preferred Intel, instead of Arm, at that time.

The two companies were allegedly sharing software engineering work to such a degree that P.A. Semi's CEO, Dan Dobberpuhl, thought that all the Intel whispers were just Apple's way of trying to get a little more leverage. P.A. Semi presumed that the design win was practically a sure thing, and the company was stunned and pretty darn upset at Apple's surprise switch to Intel.


//

2018 is a bit late, IMO. Even in 2015, AnandTech wrote how ridiculously fast Apple's "mobile" SoCs were compared to Intel's Skylake.

Everything is positive and in the double-digits even after factoring out the clockspeed increase, and with it nothing is less than 44% faster.

At this point we also have to start looking at not only who is chasing Apple, but who Apple is chasing. With yet another round of architectural improvements and a clockspeed approaching 2GHz, comparing Apple’s CPU designs to Intel’s is less rhetorical than ever before. By the time we get to iPad Pro and can start comparing tablets to tablets, we may need to have a discussion about how Twister and Skylake compare.

Apple should have been more humble in its modem launches because even M1 took 10+ years of of iteration, talent, and experience:

Apple acquires P.A. Semi (Jim Keller) + Intrinsity in 2008
Apple launches M1 on macOS in 2020 (+12 years)

Apple acquires Intel's modem business in 2019
Apple launches a custom modem in 2031 (+12 years)
 
  • Like
Reactions: bumblebritches5
Might as well just scrap it and start work on the 6G modem right now to get a head start ;)
 
The M1 was genuinely a culmination of ~10 years experience of in-house full CPU mircroarchitecture design. Apple had in fact considered switching the Mac from PowerPC directly to Arm (thus no Intel era at all). It was Jobs that preferred Intel, instead of Arm, at that time.




//

2018 is a bit late, IMO. Even in 2015, AnandTech wrote how ridiculously fast Apple's "mobile" SoCs were compared to Intel's Skylake.





Apple should have been more humble in its modem launches because even M1 took 10+ years of of iteration, talent, and experience:

Apple acquires P.A. Semi (Jim Keller) + Intrinsity in 2008
Apple launches M1 on macOS in 2020 (+12 years)

Apple acquires Intel's modem business in 2019
Apple launches a custom modem in 2031 (+12 years)
Based on the word though We will likely see something beyond 10 years like you indicate with the 2031 mark.

They indicate not only is it not ready but it will be many years behind whatever is out at that time.

And the benchmarks showing the current modem compared to previous years and the year before has been a really significant YoY increase as well from qualcomm.

Qualcomm is DAMN GOOD at what they do.

Nearly all of Qualcomms patents are FRAND and can be had cheaply (compared to other patents not in FRAND). So apple could use whatever they need from QC if needed.

But The comments from internal goes to show the beast of Modems and how apple was not only not prepared but kept up the facade that they did well which does not lead well to SWE and HWE projects which we are seeing.



I suspect 26-28 before we see ANYTHING and i expect qc x60/65 level of performance

which means QUALCOMM will be on xx90 to x100 generations
 
The M1 was genuinely a culmination of ~10 years experience of in-house full CPU mircroarchitecture design. Apple had in fact considered switching the Mac from PowerPC directly to Arm (thus no Intel era at all). It was Jobs that preferred Intel, instead of Arm, at that time.




//

2018 is a bit late, IMO. Even in 2015, AnandTech wrote how ridiculously fast Apple's "mobile" SoCs were compared to Intel's Skylake.





Apple should have been more humble in its modem launches because even M1 took 10+ years of of iteration, talent, and experience:

Apple acquires P.A. Semi (Jim Keller) + Intrinsity in 2008
Apple launches M1 on macOS in 2020 (+12 years)

Apple acquires Intel's modem business in 2019
Apple launches a custom modem in 2031 (+12 years)
It’s amusing to read that article in light of the P.A. Semi acquisition and the Apple Silicon chips. That work (and that relationship building with Apple) ended up not being wasted after all (and in a huge way). Now P.A. Semi DNA is in every device Apple ships.
 
  • Like
Reactions: ikjadoon
The M1 was genuinely a culmination of ~10 years experience of in-house full CPU mircroarchitecture design. Apple had in fact considered switching the Mac from PowerPC directly to Arm (thus no Intel era at all). It was Jobs that preferred Intel, instead of Arm, at that time.

Intel had a working 5G modem available. Apple didn't need to start from scratch, nor are they doing that.

Intel absorbed the Infineon 3G team and created 4G modems. Apple absorbed Intel's 5G team, including all the Israeli engineers which already had a working product. Four years later, still no product.
 
I for one do not want Apple to give up because if they can do it, a modem specifically optimized for iPhones and iPads (and, please, Apple - Macs!) would likely offer benefits in features, performance and battery life over using even the best "off the shelf" Qualcomm model that is designed to be "good enough" for everyone.

Obviously, until (or if) Apple gets there, continue to source from Qualcomm.

Absolutely correct.

It's not like the Apple's division for this is worth nothing. Platforms change over time too. As we speak, patents are already being approved for 6G standards, and Qualcomm doesn't hold these.
 
Last edited:
  • Like
Reactions: Tagbert
Intel had a working 5G modem available. Apple didn't need to start from scratch, nor are they doing that.

Intel absorbed the Infineon 3G team and created 4G modems. Apple absorbed Intel's 5G team, including all the Israeli engineers which already had a working product. Four years later, still no product.
I thought that Intel had a 4G modem, not a 5G modem.
 
Not sure if building their own components is a sustainable strategy compared to just buying the best components the market has to offer. What happens for example if Intel or Qualcomm develop CPUs that are much fast and much more efficient than Apple Silicon? Would Apple stick to their own chips anyway?

What's next? Their own displays? Their own camera sensors? Their own lenses? Their own batteries?

They might have a modem ready in 2026, but until then Qualcomm will also make their modems much better unless they are stopped by patents.
 
Will take a long time before it is available on the main iPhone lineup and iPads
 
I thought that Intel had a 4G modem, not a 5G modem.
Yes. And to call it working is a stretch. The worst iphone I ever had by far was the only one that has an Intel modem. Mine would suddenly just stop receiving calls for no reason. Diagnostics all looked good and even did a clean reinstall. Apple made me treat everything from requiring I get a new SIM card from Verizon (it had a brand new one in it already) to attempting to use it with an eSIM. I finally had to threaten a BBB complaint before they replaced it. New one did not have the same issue but had terrible cellular performance overall. In the same location as a previous year Qualcomm model it was always much slower. I was glad to be rid of that phone.
 
I thought that Intel had a 4G modem, not a 5G modem.

All reports say Intel had a working 5G modem. Apple got a copy of Intel's homework in 2019. Four years later, Apple has nothing to show for it.

In the same four years, Huawei was sanctioned, built their own semiconductor supply chain bottom up, and integrated their 5G modem into the main SoC. Then, added satellite voice and SMS capability this year and sold 2.5 million units so far.

I have no idea what Apple is doing.
 
  • Like
Reactions: Tagbert
I don’t understand why building a modem is so challenging. I wonder if it is more of a patent legal obstacle course rather than a technical hurdle.
It’s ok if you don’t understand. Everyone only lives so long.
As for your guesses, it’s both, plus regulatory bodies of every single major Apple market on the planet and then some. Many other posters have better answer than mine.
Cellular modems are incredibly complex devices that have to operate with many constraints. It's absolutely challenging, and a different world than building CPUs and GPUs.
It’s like its own league in terms of complexity and then some. What’s more, if all core technologies are firmly controlled by Qualcomm and patented, Apple will never be able to build their own modem without infringing Qualcomm patents and maybe also upsetting worldwide regulatory bodies due to non-compliance chips.
 
  • Like
Reactions: Tagbert
A company is never “forced” to settle. They can decide however that they want to avoid the expense and uncertainty of trial.

You mean it's like no one is ever "forced" to give up their wallet at gun point - they can just choose to be shot instead?
 
That sad part is, only us fossils will even know what you're referring to. Anyone under the age of... 30 probably? ... is all like, "wat. robotic modem?"


That secondary handshake sound that starts at the :22 second mark always felt like you were entering hyperspace.
That sound is pure nostalgia for some of us. You just made me realise how old I am 🤣
 
It's easy to see the rationale of Apple creating their own modems (not tied to a single, rather oppressive supplier + higher profit margins), but it certainly seems that this is unlikely to benefit Apple customers much in the near term.

Hopefully over time Apple can up their modem game and start producing industry leading modems, but we seem to be a LOOOONG way from that and Qualcomm's patents might prevent this from happening at all.
 
It’s not a mystery, Apple just wanted to have more control over the components inside their phone, but they overestimated the challenge and had to back down.

I’m glad they did because their Intel modems were bad.

Nope.

Apple wants to lower production costs while keeping prices the same. Thus more $$$$$.

Apple don't care if their 5G modem sucks, as long as they increase their profits doing it in house.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.