Qualcomm CEO Steve Mollenkopf Shares Thoughts on Apple Deal but Declines to Give Specific Details

[StralyanPithecus]

Qualcomm will always be needed, they own the house aka patents and all. I don't know why people like you keep failing to realize that. Nobody on the planet except maybe Huawei is free from QC's clutches.
Edit: Not to mention the time and money it'll take for Apple to design it's own modem (they'll need someone with experience to actually build it) they're at least 7+ years out from a product and by that time QC's will be so much better I doubt it'll be able to compete.
The last line is just my opinion but I doubt their going to pop up with the superior product

The only way is to create something completely different using another way of network, something similar to the old Intel WiMax technology network, that don’t use the Qualcomm patents, or wait long years for those patens to expire...Right now Qualcomm has the upper hand here.

 

[JetTester]

People making nice and playing well together to create new technologies and make groovy products from them. It's a good thing.

 

[whyamihere]

Of course Mollenkopf is excited, it's entirely possible he might have ended up out of a job if Intel hadn't failed to produce 5G modems. He should send a fruit basket to 2200 Mission College Blvd.

 

[Degerz]

It's presumptuous to think they won't. Two years is a LONG time in chip development.

Not long enough apparently for Intel who participated with the 3GPP body which also includes Samsung, Qualcomm, and Huawei but out of the four only two of the mentioned members are competitive with the significant other being regarded as a geopolitical rival ...

Are YOU somehow content that Apple has to keep trailing against the likes of the 3GPP which dominates the very standards that Apple themselves currently have NO say in ? Are YOU content with Apple just licensing patents from the 3GPP ?

Those who do NOT write the standards such as Apple will always be behind and are more importantly at the mercy against the likes of Qualcomm or Huawei so you'd best hope for Apple's own sake that they reflect on this mistake when the time comes to write the next generation standards for mobile telecommunications and they should guarantee themselves a seat in the 3GPP body ...

 

[cmaier]

Not long enough apparently for Intel who participated with the 3GPP body which also includes Samsung, Qualcomm, and Huawei but out of the four only two of the mentioned members are competitive with the significant other being regarded as a geopolitical rival ...

Are YOU somehow content that Apple has to keep trailing against the likes of the 3GPP which dominates the very standards that Apple themselves currently have NO say in ? Are YOU content with Apple just licensing patents from the 3GPP ?

Those who do NOT write the standards such as Apple will always be behind and are more importantly at the mercy against the likes of Qualcomm or Huawei so you'd best hope for Apple's own sake that they reflect on this mistake when the time comes to write the next generation standards for mobile telecommunications and they should guarantee themselves a seat in the 3GPP body ...

Intel can’t even design CPUs anymore. Apple is much better at it.

 

[Kabeyun]

How is it a win for Intel? lol

That would’ve been a fourth win, heh.

 

[techwhiz]

It's presumptuous to think they won't. Two years is a LONG time in chip development.

That is true. It's 2019 and to have a product ready for 2020, means they need the first silicon stepping taped out by the end of the year. They need to do carrier testing and certification of the modem and possibly do a second stepping and start production by July of 2020.

At 7nm wafer process time is about .8 days per layer and somewhere at 80+ layers means 64 days to fab a wafer. If they start in July they will start seeing production wafers at the beginning of Sept.
They still need to do characterization, skew lots, etc. Now they could take the second stepping and presume that it will be production and start risk wafers early. But 2020 seems way aggressive for this.

Also a 2020 Apple chip means they weren't planning on Intel.
If they weren't planning on Intel, that begs the question, why did they sign a chip deal with Qualcomm?
I won't say it's impossible but I wouldn't want to be on that silicon team.

 

[Degerz]

Intel can’t even design CPUs anymore. Apple is much better at it.

I don't care about Intel's ability to design CPUs. I care more about who CAN write the rules and if Apple aren't interested then they'd best be prepared bend over towards the likes of 3GPP which includes undesirable actors like Qualcomm ...

Those who do NOT write the rules will always be behind so Apple will have to learn to be content with being behind ...

Apple hasn't designed ANY modems so far compared to Intel who have at least designed LTE modems so I'll be extremely surprised if Apple can roll out a 5G modem in as little as two years from scratch ...

 

[cmaier]

That is true. It's 2019 and to have a product ready for 2020, means they need the first silicon stepping taped out by the end of the year. They need to do carrier testing and certification of the modem and possibly do a second stepping and start production by July of 2020.

At 7nm wafer process time is about .8 days per layer and somewhere at 80+ layers means 64 days to fab a wafer. If they start in July they will start seeing production wafers at the beginning of Sept.
They still need to do characterization, skew lots, etc. Now they could take the second stepping and presume that it will be production and start risk wafers early. But 2020 seems way aggressive for this.

Also a 2020 Apple chip means they weren't planning on Intel.
If they weren't planning on Intel, that begs the question, why did they sign a chip deal with Qualcomm?
I won't say it's impossible but I wouldn't want to be on that silicon team.

I said 2021, not 2020, for what it’s worth (for many of the reasons you’ve mentioned). In other words, 2 more years from now (and, of course, they’ve been working on this for quite awhile).

 

[Mo398]

Honestly this is a good thing. I don't know what the difference is between Qualcomm and Intel chips at the lower levels, but something isn't right with the intel modems when on Verizon. I have two iPhone XS Max's one on AT&T and one on Verizon, the Verizon one has to be restarted two-three times a day.

If I drive into a parking garage and lose LTE and then move up to a higher floor, my LTE will come back on the AT&T iPhone however the Verizon iPhone will show connected but nothing will lode until I restart it. Also when moving into a roaming area, the earlier iPhones on Verizon that had the Qualcomm chip would drop a call but the LTE would come back as roming and would work. Now with the intel chip on Verizon, my phone will still say its connected to LTE but again nothing will load until I reboot the iPhone.

Following yesterday's surprise announcement of a settlement between Qualcomm and Apple, Qualcomm CEO Steve Mollenkopf sat down with CNBC to share a few more details about the new agreement between the two companies.

According to Mollenkopf, after "a lot of talking" both between teams and with Apple CEO Tim Cook, Apple and Qualcomm came to an agreement that "both companies like." Qualcomm and Apple are now focusing on getting new products out, such as the 5G iPhone coming in 2020 that Qualcomm will supply chips for.

Apple and Qualcomm have established a "very broad deal" across all of Qualcomm's technologies, which Mollenkopf says is the first direct license that Qualcomm has had with Apple rather than contract manufacturers.

Each side "found something that was useful" in the deal, and according to Mollenkopf, Apple and Qualcomm "want to work together on products," as evidenced by the multiyear product deal the two signed as part of the settlement.

Part of the agreement between the two companies included a payment from Apple to Qualcomm, but Mollenkopf declined to provide further details on the size of the payment. He also refused to reveal how much Apple is paying Qualcomm per phone.

On the topic of 5G chips for future iPhones, Mollenkopf said that Qualcomm is "excited" and has the "entire team" working to support Apple. Unsurprisingly, no details were given on Apple's product plans or launch timelines for 5G connectivity.

While Apple settled with Qualcomm, Qualcomm continues to face an FTC inquiry into anticompetitive business tactics, which Mollenkopf says is still a risk to Qualcomm. He doesn't believe the Apple decision will impact what the FTC decides.Mollenkopf's full interview, which includes additional details about Qualcomm's relationship with Apple and Qualcomm's goals moving forward, can be watched on CNBC's website.

Article Link: Qualcomm CEO Steve Mollenkopf Shares Thoughts on Apple Deal but Declines to Give Specific Details

 

[techwhiz]

I said 2021, not 2020, for what it’s worth (for many of the reasons you’ve mentioned). In other words, 2 more years from now (and, of course, they’ve been working on this for quite awhile).

My bad you did say two more years. That is 2021 and since Apple releases end of year, it's more like 2.5.
I don't know how long they have been working on it. I do know they recently started a design center in San Diego.
The presumptive reason would be to get Qualcomm people that have modem experience to defect and not have to relocate. SEG is big and they have groups all over the place.

I still wouldn't want to be on that silicon team.
It's still pretty aggressive for having never done a modem.
The validation process is much more difficult than just needing to verify functionality.
They need to verify interoperability with carriers all over the world.
They also need to support 3G, LTE and 5G; probably not 2G.

 

[cmaier]

My bad you did say two more years. That is 2021 and since Apple releases end of year, it's more like 2.5.
I don't know how long they have been working on it. I do know they recently started a design center in San Diego.
The presumptive reason would be to get Qualcomm people that have modem experience to defect and not have to relocate. SEG is big and they have groups all over the place.

I still wouldn't want to be on that silicon team.
It's still pretty aggressive for having never done a modem.
The validation process is much more difficult than just needing to verify functionality.
They need to verify interoperability with carriers all over the world.
They also need to support 3G, LTE and 5G; probably not 2G.

I would love to be on that team. Nothing like doing the first one. Also, keep in mind that while "apple" has never done it before, many of the employees have. They also have the benefit of having pretty good infrastructure (circuit designers, EDA flows, etc) already in place. As for carriers it's probably the case that they need to test against a bunch of infrastructure providers, but there are fewer of those than there are carriers. And perhaps they are going to pick up resources/IP from Intel, that would kickstart backwards compatibility with LTE and 3G.

I firmly believe it can be done in time, especially given that they started at least a year ago.

 

[thekev]

Intel no longer has to direct resources at trying to develop a 5G wireless chip that can achieve Apple's "high standards"

Intel can now focus those resources at developing better PC processors to avoid Apple dropping them in favor of Apple's own custom ARM processors.

Those are different divisions entirely, so it's only really a choice if Intel is constrained by cash flow or people to manage that division.

On Apple's end, they're likely to do whatever they think is good enough with processors. If they decided to make their own, Intel's performance is unlikely to impact this decision. Processor designs aren't going to see triple digit percentage gains with respect to performance. Even if Intel's solutions were 50% faster, I don't believe this would deter Apple.

Typically you need at least a 50% throughput gain in terms of max GFLOPs to make something actually noticeable in things that don't involve a continuous stream of processing. This figure also holds pretty well across instruction set architectures, because most of the common ones can only get close to maximal throughput if it involves a long stream fused multiply add operations and continuous chunks of memory that fit entirely within L2 cache without banking conflicts.

 

[techwhiz]

I would love to be on that team. Nothing like doing the first one. Also, keep in mind that while "apple" has never done it before, many of the employees have. They also have the benefit of having pretty good infrastructure (circuit designers, EDA flows, etc) already in place. As for carriers it's probably the case that they need to test against a bunch of infrastructure providers, but there are fewer of those than there are carriers. And perhaps they are going to pick up resources/IP from Intel, that would kickstart backwards compatibility with LTE and 3G.

I firmly believe it can be done in time, especially given that they started at least a year ago.

I've been on teams doing it (whatever it is) first.
It can be a blast, but it can be a nightmare.
The question is did they get the right modem people?
The RTL coding people would not be my worry.
Whether they got the people that understand how the modem works is what I want to know?
It's all conjecture and speculation.

It will be interesting in the next couple of years to see how it pans out.

In the meantime, I just need people to buy whatever widgets that increase Apple's stock price.

 

[cmaier]

I've been on teams doing it (whatever it is) first.
It can be a blast, but it can be a nightmare.
The question is did they get the right modem people?
The RTL coding people would not be my worry.
Whether they got the people that understand how the modem works is what I want to know?
It's all conjecture and speculation.

It will be interesting in the next couple of years to see how it pans out.

In the meantime, I just need people to buy whatever widgets that increase Apple's stock price.

I was on the first amd64 chip. I had a blast. And there were only around 20 of us.

With this baseband chip stuff half of the design work is matlab. It’s pretty wild. I’ve seen a lot of the design data because of my “new” career. I’ve spent a lot of time sitting with engineers trying to figure out whether they infringe, if there are workarounds, etc. the turbo coding stuff took me quite awhile to understand, but once you get it it’s not too difficult, at least to the point where I’m confident I could implement it. So if you have one or two geniuses at that stuff, the rest of the team just needs to plug away.

 

[Degerz]

My bad you did say two more years. That is 2021 and since Apple releases end of year, it's more like 2.5.
I don't know how long they have been working on it. I do know they recently started a design center in San Diego.
The presumptive reason would be to get Qualcomm people that have modem experience to defect and not have to relocate. SEG is big and they have groups all over the place.

I still wouldn't want to be on that silicon team.
It's still pretty aggressive for having never done a modem.
The validation process is much more difficult than just needing to verify functionality.
They need to verify interoperability with carriers all over the world.
They also need to support 3G, LTE and 5G; probably not 2G.

Not to worry Apple anymore since they're already having a problem but 3GPP release 15 conformance specifications has been frozen just nearly a month ago so the validation process has been sorted out. The 3GPP body will take responsibility for the functionality of interoperability ...

Like you said though, I don't think many of the top tier scientists or engineers in that particular field would be thrilled with the idea of joining a new division in a corporation in which the said corporation currently doesn't participate with the standards body on the thing that those scientists/engineers are supposed to create within a very optimistic target ...

TBH, it sounds like a really bad idea for Apple to enter the mobile modem market just to solely supply their own 5G modems without representation in the 3GPP body. If that is their strategy then it's probably not going to get them very far in the long run when the 3GPP decides to write the next generation standards and Apple sees a use case for it but they're effectively locked out because they don't have a coveted seat at 3GPP ...

 

[dampfnudel]

This is the best outcome by far. Waiting for Intel would’ve meant 5G in 2021 at the earliest and probably not as good as Qualcomm. Continuing the lawsuit(s) would’ve been a victory mainly for the attorneys. Maybe ten years from now, Apple will have an in-house solution in place, but there’s no doubt that Qualcomm will be the best choice for the next few iPhone generations.

 

[Tevos]

So Qualcomm settled. If Apple could do that with Qualcomm you know who is next on the 'to do' list.

Now time to make up with nVidia.

You don't have to put nVidia cards in Macs. Just add driver support. Is it so much to ask? If not then indeed Apple fruit is that rotten they don't care about Pro customers anymore.

 

[recoil80]

How is it a win for Intel? lol

The modems business wasn't really profitable for Intel. I wouldn't call it a win of course, but not a huge loss either.

 

[gnasher729]

We need to get Steve Mollenkopf on here for AMA. People want to know if Intel suddenly dropped out of the mobile business since they were faced with impending antitrust case colluding with Apple to stifle Qualcomm and if Qualcomm/FTC will continue to pursue that.

That's the worst nonsense I've read for awhile. Intel was _competing_ with Qualcomm. Which is not only legal, but highly desirable.
[doublepost=1555577970][/doublepost]

Intel can’t even design CPUs anymore. Apple is much better at it.

To be fair, Apple has an advantage there, because whatever they do, their customers will agree (because Apple's only chip customer is Apple). I think Apple managed to get rid of lots of historical ballast (CPUs are 64 bit only), while Intel still supports tons of ancient options. Instruction decoding may not take that much physical space, but is still a lot of design effort that Intel has to do, and Apple doesn't, or much less.

 

[chatin]

That one-time payment has got to be in the billions! What's not to like if you are Qualcomm. I would guess $5 billion for the 5 G's

 

[cmaier]

That's the worst nonsense I've read for awhile. Intel was _competing_ with Qualcomm. Which is not only legal, but highly desirable.
[doublepost=1555577970][/doublepost]
To be fair, Apple has an advantage there, because whatever they do, their customers will agree (because Apple's only chip customer is Apple). I think Apple managed to get rid of lots of historical ballast (CPUs are 64 bit only), while Intel still supports tons of ancient options. Instruction decoding may not take that much physical space, but is still a lot of design effort that Intel has to do, and Apple doesn't, or much less.

Then compare apples to apples (amds to intels). As a former AMD cpu designer, I stand by “intel can’t design chips anymore”

 

[tooltalk]

Where did he say it was a win for intel?

It's a win for Intel's baseband team -- this deal effectively puts that group out of misery.

 

[Wags]

Shouldn’t some of this be public record and especially for investors

 

[veg8]

Shouldn’t some of this be public record and especially for investors

Hints are in this article
https://www.cnbc.com/2019/04/18/apple-paid-5-billion-to-6-billion-to-settle-with-qualcomm-ubs.html