Apple hires 8 year Qualcomm Engineering VP to its wireless SoC team

AppleInsider

Apple has hired Qualcomm Engineering VP Esin Terzioglu as a wireless "System on a Chip" lead, offering additional evidence that the company may plan to expand its internal chip development into broadband processors working as mobile modems.

Esin Terzioglu, LinkedIn profile photo

As noted in a tweet by Neil Shah of Counterpoint Research, Terzioglu posted the news to his LinkedIn account, stating, "After an amazing ~8 years at Qualcomm, it is time for me to move on to my next adventure.

"It has been my honor and privilege to have worked with so many talented and dedicated individuals at Qualcomm where we accomplished great feats as a team (10nm bring up was a doozy and the team did an amazing job bringing the first product to market!!!). I feel privileged for the opportunity to continue my career at Apple. Stay in touch."

Terzioglu's profile notes that he began working at Qualcomm in August 2009, where he lead the company's QCT [Qualcomm CDMA Technologies] Central Engineering organization, defining its technology roadmap.

He is has a Ph.D. in Electrical Engineering with a Ph.D. minor in Computer Science, both completed at Stanford University.

Apple and Qualcomm

Apple and Qualcomm are involved in a series of legal disputes related to the complex system of royalty rates (depicted below) that calculates how Apple pays for its use of standards essential patents related to wireless networking, specifically in mobile networks that use CDMA and LTE, both standards built upon technologies developed by Qualcomm.

Apple is also a major customer of Qualcomm, which has built most of its iPhone and iPad 3G/LTE baseband processor modems used since the Verizon/CMDA iPhone 4 was introduced in early 2011. Prior to that shift, Apple bought GSM-only modems from Infineon.

Intel and alternatives

Intel has since acquired Infineon, and the chipmaker recently introduced Baseband Processors that support a wider variety of global networks, licensing the required IP from Qualcomm to do so.

Apple began using a mix of Intel and Qualcomm baseband processors last year with the introduction of iPhone 7, with Intel chips specifically tasked for models designed for use with AT&T and T-Mobile, where CDMA support is not necessary. The newest Intel baseband chips could enable Apple to shift entirely to Intel in the future.

In 2014, Apple recruited at least 30 mid and senior-level RF engineers from Broadcom and Qualcomm, suggesting an intent to begin working on its own internal baseband chips.

Apple has developed a series of custom ARM-based Application Processors since the 2010 A4, but none of these incorporate an integrated baseband, which works as an independent ARM-based computer running its own specialized operating system to manage all network traffic between carriers' mobile networks and Apple's system running iOS.

Qualcomm's Snapdragon and Samsung's Exynos both have integrated SoCs that pair an Application Processor and Baseband Processor on the same chip package.

ericthehalfbee

LOL.  Sounds like this guy was pretty high up at Qualcomm, and now gets poached by Apple.

melgross

This doesn't mean that Apple is developing their own cell radios. All major companies have teams for areas in which their devices function. When they design a product, they need people with expertise in all areas. They also need people to liaison with the actual parts and module designers at other manufacturers. This is particularly true when they have special requirements.

it would be interesting if Apple were looking to design their own radios, but why bother? This isn't as important as their SoC. What I can see is the possibility of Apple gaining a concession from Intel that they couldn't get from Qualcomm to put the radio on the SoC silicon. If that were so, they'd need people with more expertise than they likely had, though Intel would work on that with Apple. 

qwwera

Apple is rather brash.....considering the lawsuits and isht..

cali

I believe an Mobile would work perfectly.

Go to the Apple Store. Buy iPhone. Turn it on. Simple as that. It seems carriers have been a slight bottleneck for Apple products. Not a big deal but would be awesome to have Apple as my carrier and fix all the damn problems carriers aren't willing to fix.

Every iPhone open to flashing. No data limit on Siri and Apple services. No SIM card needed. So many ideas.

tht

melgross said:

This doesn't mean that Apple is developing their own cell radios. All major companies have teams for areas in which their devices function. When they design a product, they need people with expertise in all areas. They also need people to liaison with the actual parts and module designers at other manufacturers. This is particularly true when they have special requirements.

it would be interesting if Apple were looking to design their own radios, but why bother? This isn't as important as their SoC. What I can see is the possibility of Apple gaining a concession from Intel that they couldn't get from Qualcomm to put the radio on the SoC silicon. If that were so, they'd need people with more expertise than they likely had, though Intel would work on that with Apple. 

There are some good reasons for Apple to make their own cellular modems:

1. It's cheaper. Designing their own cellular modem that's integrated into the SoC and licensing the FRAND IP should be cheaper than buying a 3rd party chip and licensing the FRAND IP. "10 nm" chips will have a LOT of transistors. Devoting some of them for cellular connectivity would be a good thing. 

2. Integrating cellular modems into the SoC will save board space. This would make room for some other component, more battery or some other feature. It could be a double whammy for instance: the cost for cellular connectivity will be cheaper and say they can use 2 lower density NAND chips instead of 1 high density NAND chip. 

3. They can design cellular modems specifically for their hardware. An LTE modem in a Watch? In Airpods? In some other wearable? Their own laptops? For the Watch, it doesn't need to be a multi aggregating band monster. Just something highly targeted and therefore could be optimized for lower power consumption. 

I do do agree that this guy could simply be hired as a VP for Apple's cellular integration work, but there are good reasons for Apple to design their own cellular chips. Cellular bandwidth is at the point of diminishing returns. Driving down its cost in all phases would be a good thing, and making their own would be one way to do it. 

gmgravytrain

Apple needs to build a wide a moat as possible by taking responsibility for as many areas as they can to personally control. Put as many functions on a SoC as possible. Build hooks into all their hardware for the Apple Neural Engine processor to control. I have no idea if it's easy for Apple to design its own cellular modems but I'll just have to trust Apple knows what it's doing. I'm sure there will be a lot of naysayers who have no faith in Apple's decisions, but that's their problem. One thing for certain, the more functions Apple can pack on a SoC, the more efficient it will be and any space saved can be filled with a larger battery.

Apple needs to make it as hard as possible for rivals to copy them.  As it is, I don't know how other companies manage to keep up with Apple. How efficient can Android OS be having to work with so many hardware variants?

edited May 2017

shapetables

Tie

sflocal

It makes complete sense for Apple to build their own baseband processor.  Qualcomm's shot itself in the foot with Apple and now they are paying the price.  Apple eventually creates their own, and does not need to share whatever improvements back to Qualcomm, and thus, its competitors.

Eventually Qualcomm will just be a generic baseband processor for Android, and nothing else... if that.

If this does indeed result in Apple making their own chip, I wonder what the folks over at Intel are thinking right now?

fatman

Integrating as many functions as possible in the main SOC at a small die size (<10nm) = Less dependency on vendors, higher efficiency in space, speed and cost. The increased complexity will require engineers like this guy. Apple's goal: make the teardown boring -- just one chip that handles everything. I know there are cons to this - but fun to imagine - especially for the iEye ... the product I estimate will be the next best thing.

radarthekat

cali said:

I believe an Mobile would work perfectly.

Go to the Apple Store. Buy iPhone. Turn it on. Simple as that. It seems carriers have been a slight bottleneck for Apple products. Not a big deal but would be awesome to have Apple as my carrier and fix all the damn problems carriers aren't willing to fix.

Every iPhone open to flashing. No data limit on Siri and Apple services. No SIM card needed. So many ideas.

There are close to 1000 carriers around the world.  It's be a big effort to cover a meaningful amount of the planet.  And without doing that, the effort, and iPhones that support it if there would need to be anything different about them, would be severely limited in geographic region where they could be sold.  I think this would be a distraction to Apple.  Don't do what others can do unless you can add some significant capability.  This is why Apple didn't/won't get into the TV screen market.  

strangedays

qwwera said:

Apple is rather brash.....considering the lawsuits and isht..

Brash?

janiceandrusty

ericthehalfbee said:

LOL.  Sounds like this guy was pretty high up at Qualcomm, and now gets poached by Apple.

I doubt he'll make it through the second reorg.  He was hired most likely for the PR and press release.

melgross

tht said:

melgross said:

This doesn't mean that Apple is developing their own cell radios. All major companies have teams for areas in which their devices function. When they design a product, they need people with expertise in all areas. They also need people to liaison with the actual parts and module designers at other manufacturers. This is particularly true when they have special requirements.

it would be interesting if Apple were looking to design their own radios, but why bother? This isn't as important as their SoC. What I can see is the possibility of Apple gaining a concession from Intel that they couldn't get from Qualcomm to put the radio on the SoC silicon. If that were so, they'd need people with more expertise than they likely had, though Intel would work on that with Apple. 

There are some good reasons for Apple to make their own cellular modems:

1. It's cheaper. Designing their own cellular modem that's integrated into the SoC and licensing the FRAND IP should be cheaper than buying a 3rd party chip and licensing the FRAND IP. "10 nm" chips will have a LOT of transistors. Devoting some of them for cellular connectivity would be a good thing. 

2. Integrating cellular modems into the SoC will save board space. This would make room for some other component, more battery or some other feature. It could be a double whammy for instance: the cost for cellular connectivity will be cheaper and say they can use 2 lower density NAND chips instead of 1 high density NAND chip. 

3. They can design cellular modems specifically for their hardware. An LTE modem in a Watch? In Airpods? In some other wearable? Their own laptops? For the Watch, it doesn't need to be a multi aggregating band monster. Just something highly targeted and therefore could be optimized for lower power consumption. 

I do do agree that this guy could simply be hired as a VP for Apple's cellular integration work, but there are good reasons for Apple to design their own cellular chips. Cellular bandwidth is at the point of diminishing returns. Driving down its cost in all phases would be a good thing, and making their own would be one way to do it. 

As I said, it's much more likely that Apple will get Intel to allow the cell radio (it's not a modem, folks) to be integrated into the core. For all we know, they've already agreed to that and are waiting for the right radio to do it with. Apple doesn't need to design their own radios for that. While we don't what Apple is doing in their labs, it's very expensive to design flagship radios. It takes years to get good, state of the art designs. It may not pay.

in addition, Intel has shown, over the years, that they are very open to modifying their designs specifically for Apple. They really want their business. Apple has shown that they are willing to move their manufacturing to different companies. Ever since it was admitted that it was of the biggest mistakes Intel made when they turned down the opportunity to manufacture Apple's SoC originally, it's very likely that now that they are willing to be a foundry for others, that they are angling to get Apple as a customer for the SoC.

there's no guarantee that Apple will remain with TSMC forever.

melgross

radarthekat said:

cali said:

I believe an Mobile would work perfectly.

Go to the Apple Store. Buy iPhone. Turn it on. Simple as that. It seems carriers have been a slight bottleneck for Apple products. Not a big deal but would be awesome to have Apple as my carrier and fix all the damn problems carriers aren't willing to fix.

Every iPhone open to flashing. No data limit on Siri and Apple services. No SIM card needed. So many ideas.

There are close to 1000 carriers around the world.  It's be a big effort to cover a meaningful amount of the planet.  And without doing that, the effort, and iPhones that support it if there would need to be anything different about them, would be severely limited in geographic region where they could be sold.  I think this would be a distraction to Apple.  Don't do what others can do unless you can add some significant capability.  This is why Apple didn't/won't get into the TV screen market.  

Exactly! I don't know why people think that this is all easy. They forget that it took Apple about six years from when they first began with small modifications to their SoC to where they are now. And they were able to do that because the SoC market had been lazy, not concerned with significant performance or efficiency improvements.

that's not so for cell radios. Apple would need to come out of the gate running level with first place manufacturers. It takes years to do that. Look at Intel's first try, even after having bought a radio company. Now, they've bought another with CDMA designs and patents, so, hopefully, this year, in addition to more equal performance to Qualcomm (which, in the real world wasn't an actual problem) they should, hopefully, include CDMA.

so,where is Apple in all this? How many patents do they have? What leading edge designs? Likely, very little.