Apple Considered Purchasing Intel's Smartphone Modem Chip Business

[name99]

But what’s the point in investing in all that just to still pay hefty licensing fees in the end .... go all they and develop a superior chip

The point is that at the END of the process
(a) You can have a more performant device. You can use lower power by moving the logic onto the SoC (right now it's a separate chip on the PCB). And if Apple chooses to prioritize performance over area (like they have done with their CPUs and GPUs) you can devote more area to things like Viterbi decoding, which translate into higher performance.

(b) You can save money. Yes, whatever you do you owe QC royalties. But right now, you're ALSO paying QC to buy those modems. If you make your own modems, you're not paying QC that money, and it's not costing you as much if it's just another 10% or so of the SoC die.

 

[CWallace]

But what’s the point in investing in all that just to still pay hefty licensing fees in the end .... go all they and develop a superior chip

Once Apple makes their own modems, they can optimize and tailor their performance for their products specifically as opposed to having to "least common denominator" it to work with every OEMs products as is the case now. Just as Apple's A-Series chips outperform the Snapdragon series used by most other Tier 1 smartphone manufacturers, I would not be surprised that in time iOS devices offer superior cellular performance than other devices due to that refinement. And that will become another competitive advantage for iOS devices in the marketplace.

Current fab technology is working just fine for 5G modems. Qualcomm is using 10nm while Huawei is on 7nm. It's Intel that's far behind.

Rumors are that Intel would have outsourced their 5G modem production to TSMC to take advantage of their smaller process.

Actually, no. That was THE significant aspect of the recent settlement, that Apple gets rights to QC IP for at least the next six years. That's why I score the whole thing as a win for Apple. They pay QC a little more than they were paying (not MUCH more, from about $7.5/phone to $9/phone) but in return they get something priceless, the freedom to build their own modem without QC hassling them in court for the next twenty years.

Once Apple did their own modem, Qualcomm would have been required to license them their patents under FRAND and likely at a much lower fee than they were paying since it would arguably be levied against the value of the modem chip itself and not the device (as was the case prior to the trial/settlement).

 

[JetTester]

I hope Apple will develop their own modem chips. If they end up making their own Mac chips, that will put most of the control of hardware in their hands. I think that can be a big plus if it is done right. Or it can be the last brick in the building of the world's largest monopoly.

 

[macfacts]

That's why I score the whole thing as a win for Apple. They pay QC a little more than they were paying (not MUCH more, from about $7.5/phone to $9/phone) but in return they get something priceless, the freedom to build their own modem without QC hassling them in court for the next twenty years.

Qualcomm was "hassling" apple because apple forced Foxconn and others to stop paying license fees to qualcomm.

 

[realtuner]

Consider that a 50 ms round trip latency means car traveling at 70 mph will move 5 ft even before any decisions can be made.

26 ms is good for humans but about 10x too slow for AI.

Yeah, no. I work in automotive engineering. There's no way AI or anything related to autonomous driving is going to be offloaded somewhere and have to rely on a flaky wireless connection in order to work. It will all be done on-vehicle.
[doublepost=1556311627][/doublepost]

Qualcomm was "hassling" apple because apple forced Foxconn and others to stop paying license fees to qualcomm.

Qualcomm was hassling Apple because Qualcomm had an illegal "gag order" clause in their contract to prevent companies like Apple from co-operating with antitrust and other government regulators, and Apple didn't abide by this.

 

[Ingot]

Thank god they didn’t
[doublepost=1556309759][/doublepost]
The original iMac shipped with 4gb of storage. Needs change over time. Sure 5G might not be “necessary” now, but who knows what our needs will be 5 years from now. Maybe I’ll want to watch live tv in 4K on my tablet. Sure it’s not possible today, but you never know about the future.

I quite agree. Along with GPS, fast connectivity can lead to autonomy in our mobile lifestyle. I am thinking more along the lines of transportation.

 

[Solomani]

But what’s the point in investing in all that just to still pay hefty licensing fees in the end .... go all they and develop a superior chip

Nothing is preventing Apple from pursuing their own chip design in the future. Maybe they will throw their weight into designing whatever comes after 5G. For now, I simply see Apple biding their time. Apple is simply not ready to be a contender in the modem chip arena, and they know it. But Apple has more cash reserves than most nations on Planet Earth. They can continue R&D for a future modem chip. But in the meantime, they will buy their time and rely on Qualcomm as a supplier.

 

[twolf2919]

Qualcomm was "hassling" apple because apple forced Foxconn and others to stop paying license fees to qualcomm.

...which is, of course, because QC was double-dipping on their IP fees.

 

[Tech198]

Selling the modem business would also prevent one other thing from ever happening:

an iPhone baring the "Intel inside" chip ? *shocked*

Would Apple be happy about that i wonder..

 

[paul4339]

additionally I don't think Apple has ever been over-joyed with Intel's delivery of their CPU roadmap. Intel was frequently late and delayed CPU and chipset releases, throwing a wrench in Apple's laptop plans (I think).

If Intel did the same with their cellphone modem roadmap its be a giant headache for Apple.

 

[Freida]

Thank god they didn’t
[doublepost=1556309759][/doublepost]
The original iMac shipped with 4gb of storage. Needs change over time. Sure 5G might not be “necessary” now, but who knows what our needs will be 5 years from now. Maybe I’ll want to watch live tv in 4K on my tablet. Sure it’s not possible today, but you never know about the future.

Oh dear, not this again.

Future is future and I'm sure 5G will be default but there is really no rush and obsession now. Its just specs chase but thats about it. Most places are not covered fully with LTE and you need 5G asap? Ha! Hilarious as usual
[doublepost=1556314334][/doublepost]

Thank god they didn’t
[doublepost=1556309759][/doublepost]
The original iMac shipped with 4gb of storage. Needs change over time. Sure 5G might not be “necessary” now, but who knows what our needs will be 5 years from now. Maybe I’ll want to watch live tv in 4K on my tablet. Sure it’s not possible today, but you never know about the future.

p.s.: Really no need to watch 4K content on a tablet on mobile data. I have 65" tv for that and even if I ignore the idiocy of the proposal I still can safely tell you that 5G next year has little to no value for customers. That value is at least 4 years away if not more.
We could also bring the debate whether or not you would be able to see a difference between 4K or FHD on an iPad. But that would be for another topic and I'm sure you would not be honest with me

 

[tridley68]

This would have been a good move on Apples part one step closer to becoming self sufficient

 

[Wanted797]

Just tested my XS Max (inside our large steel framed warehouse) and got 22ms ping, 45Mbps download and 26 Mbps upload.

I also looked at my history over the years (nice that the Speedtest App keeps a record of all your tests).

On 3G iPhones I had results from 1.5-4.5Mbps.
On 4G iPhones I had results from 5-20Mbps for my earlier devices (5S and 6).
On my XS Max I get between 40-80Mbps consistently.

The difference from 3G to 4G was significant. The difference from 4G to 5G might look good on paper, but will have practically no impact whatsoever on people using their devices. Getting over 40Mbps is plenty fast for any tasks I could reasonably do on a mobile device. It's also plenty fast enough to use as a mobile hotspot for my MBP - I never see slow downs or delays while browsing the Internet or editing documents stored online. It's not like I'll ever need to download multi-GB files when I'm mobile.

5G is just a check mark on a feature list that companies will try to make seem far more important than it really is so they have an excuse to sell people shiny new devices.

You have the think of the big picture. 5G isn’t going to be big for phones it will be big for everything else internet connected.

I live in Australia where most the population lives on the coast line. Or in small isolated towns in the middle. But most the middle isn’t populated. This makes wireless connectivity far easier to implement then rolling out hundreds and hundreds of miles of fibre optic cable to give the small towns decent internet speeds.

We already have an awesome phone network that already covers theses areas. An upgrade to the tower means that town gets fast speeds, not for phones but for internet on their computers, remote hospitals etc.

Also think of the application with drones flying and relaying information back. With more data steam more information can be sent back to a remote location.

5G isn’t for phones. It’s building a network for more than just phones.

 

[Gutwrench]

Getting over 40Mbps is plenty fast for any tasks I could reasonably do on a mobile device.

I agree (I really do), until it’s not enough. Speed is like memory.

Question: "I read in a newspaper that in 1981 you said '640K of memory should be enough for anybody.' What did you mean when you said this?"

Gates: "I've said some stupid things and some wrong things, but not that. No one involved in computers would ever say that a certain amount of memory is enough for all time."

 

[elvisimprsntr]

Intel was likely not completely forth coming about its 5G plans and issues it was having, thus difficult for Apple to commit to a 5G modem supplier with less than all the information. So Apple suggests an acquisition opportunity, which it knows will take years of negotiation, and had no real intention of an acquiring. Once Apple gets access to the real technical and financial details of the business, internally they make a decision it is cheaper to settle with Qualcom than to try to acquire and fund their own 5G modem development and manufacturing.

 

[magicschoolbus]

Well maybe I am in the minority here but I think they should have purchased it. A hardware war is heating up, and considering what Apple has been able to with SoC's and A series chips, I would think that they could do the same with modems.

Another company was literally holding Apples next flagship product hostage - this is not good.

 

[Rigby]

Leading edge ARM designs like Cortex-A76 can be licensed from ARM. It's how the business is run. But nobody is selling modem designs. Everyone in the industry is doing their own implementation. So it's no surprise Apple can produce great ARM products but struggle with modems.

Yes, and the technology is highly complex and parts of the standards were only recently finalized. There are just very few companies that have the engineering capability to master 5G modems at this point.

Current fab technology is working just fine for 5G modems. Qualcomm is using 10nm while Huawei is on 7nm. It's Intel that's far behind.

The XMM8161 baseband is actually based on Intel's 10nm process, which is roughly comparable to TSMC's 7nm (which is what Qualcomm is planning to use eventually).
[doublepost=1556318399][/doublepost]

What is so special about 5G? Sure its faster but LTE is already fast so why is everyone obsessed about 5G? Also, it will take years before the coverage is present so surely there is no rush

It's not only about speed, but also opening up additional spectrum, better spectrum efficiency, latency, better scalability for billions of IoT devices, a more flexible core architecture and a number of other benefits.

 

[mikethemartian]

I’m seeing a 26ms ping on my LTE on my 2017 iPhone X. I just ran Speedtest from inside the office building I work in.

Seems fine, pretty darn low latency really.

Did you expect the ping to be bandwidth critical?

 

[raghu8912]

What is so special about 5G? Sure its faster but LTE is already fast so why is everyone obsessed about 5G? Also, it will take years before the coverage is present so surely there is no rush

Service providers can charge more for same speeds but can claim 5G speeds.
Mobile phone manufacturers can brag about having 5G modems in their hardware.
AT&T can brag about being first 5G to market.
At the end customers would still get same speeds, same crappy signal quality, at least in US.

Similar to how the transition was from 3G to 4G.
I am confident non of the service providers are using 4G to full extent, they are talking about 5G.
Similar to how there not much content in 4K but Samsung is already talking about 8K.
[doublepost=1556319637][/doublepost]

It took QC a couple decades to get to the point where they can develop 5G chips. Why are you baffled Apple can't get a 5G modem chip ready in a couple years? Designing modems isn't just adding a few more transistors on a CPU chip!


This makes no sense. QC, Samsung, Huawei, and a couple other companies already have 5G chips - so they're obviously getting fabricated.

Qualcomm doesn't have Fab either, I think the problem with 5G is not Fab, just design & IP.

 

[mikethemartian]

What is so special about 5G? Sure its faster but LTE is already fast so why is everyone obsessed about 5G? Also, it will take years before the coverage is present so surely there is no rush

Apparently Tim Cook disagrees with you otherwise he wouldn’t have needed to settle.

 

[NickName99]

Did you expect the ping to be bandwidth critical?

Not at all, that was in the context that 5G is all about latency.

 

[JPack]

Yeah, no. I work in automotive engineering. There's no way AI or anything related to autonomous driving is going to be offloaded somewhere and have to rely on a flaky wireless connection in order to work. It will all be done on-vehicle.

Sure there is, it’s called a sensor network. Vehicles in front may detect a crash or other dangerous condition and relay that information to other vehicles.
[doublepost=1556321238][/doublepost]

The XMM8161 baseband is actually based on Intel's 10nm process, which is roughly comparable to TSMC's 7nm (which is what Qualcomm is planning to use eventually).

TSMC’s second gen 7nm using EUV is already in mass production. The 10nm XMM8160 by the time it shipped in smartphones, would be in 2020. That is well into TSMC’s 5nm ramp.

 

[realtuner]

Sure there is, it’s called a sensor network. Vehicles in front may detect a crash or other dangerous condition and relay that information to other vehicles.

Yes, and they'll do that using something like V2V (vehicle to vehicle) communication. Vehicles will talk directly to each other so they can respond instantly to safety hazards (like a collision a few vehicle ahead) which completely removes the issue of relying on a wireless network to "hopefully" get the data to the other vehicles in time.

For other items that are less critical (like accidents further up ahead, road construction, icy roads or other issues) wireless over cellular is plenty good enough. And that type of information doesn't require the speed or low latency of something like 5G.

 

[Rigby]

Sure there is, it’s called a sensor network. Vehicles in front may detect a crash or other dangerous condition and relay that information to other vehicles.
[doublepost=1556321238][/doublepost]

TSMC’s second gen 7nm using EUV is already in mass production.

That may be so, but Qualcomm's current single-mode baseband isn't using it.

The 10nm XMM8160 by the time it shipped in smartphones, would be in 2020. That is well into TSMC’s 5nm ramp.

Again, Qualcomm's modem part at that time would use the 7nm process. Intel was behind by a few months, but it's not by a huge margin. And they may actually have considered outsourcing the fabbing since the modem chips would have blocked more profitable uses of their own fabs.

 

[realtuner]

Apparently Tim Cook disagrees with you otherwise he wouldn’t have needed to settle.

Logical fallacy. Here, let me try another one.

If 5G was so critical, then the 2019 iPhone should come with 5G and Apple shouldn't wait until the 2020 iPhone to add it.