Exactly. The risk in doing that however is that when your in-house team fails to deliver for whatever reason you've shot yourself in both feet. Then the attempt to have a better part at a reduced costs turns into potentially having no end product or at best a significantly delayed one while it's redesigned to use a different part. The result is massive losses.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple's Recent Hires from Broadcom Boost Rumors of In-House Baseband Chip Development
- Thread starter MacRumors
- Start date
- Sort by reaction score
Exactly. The risk in doing that however is that when your in-house team fails to deliver for whatever reason you've shot yourself in both feet. Then the attempt to have a better part at a reduced costs turns into potentially having no end product or at best a significantly delayed one while it's redesigned to use a different part. The result is massive losses.
Good point, however it's not plausible that they've hired 30 and are looking for 80 more for that purpose.
I remember that talk. The only thing I would be interested in knowing is if that will still allow easy phone swaps, without having to talk to the carrier.
Great. This will provide Apple with multiple benefits.
Size/power savings
The Qualcomm chip is a complex thing. It is so complex that it has its own ARM processor inside running its own little operating system. It takes quite a lot of space on the board:
Some of the code Qualcomm’s CPU is running could probably run on Apple’s main CPU. Apple’s CPU cache could also be reused. The remaining baseband-specific stuff could be implemented as separate blocks inside Apple’s chip package, just like Imagination’s GPU. That would cut down on space.
The power savings could be provided by elimination of redundant hardware across multiple chips, Apple’s P.A. Semi chip power-saving tricks, and Apple’s software better managing baseband power states.
Cost savings
Take a look at this table:
The Qualcomm chip costs about $32. It is one of the most expensive components. Apple could probably make their own wireless chip for $15–20. Or maybe integrate it into their next An chip so that the total cost is about $25–30. Of course, there are some tricky patent licensing costs involved, but it still may be more cost-efficient for Apple to do its own stuff.
Plus, if they’re going to do their own LTE chip, they’re surely going to do their own Wi-Fi/BT/GPS as well. That’s another $4 off.
Innovation opportunities
Apple would be in control of its wireless R&D schedule, aligning it to its own product roadmap.
They could start by moving much more aggressively along LTE Releases 10/11/12, partnering with select carriers around the world to do that. Sort of like Samsung has partnered with SK Telecom to launch 150Mbps LTE-A in South Korea.
They could also implement new LTE features that do not require carrier network upgrades, but do require control of all components; for example, “in-device co-existing”, which eliminates interference between LTE/Wi-Fi/BT radios by co-ordinating their transmission.
After a couple of R&D cycles, they could attempt innovations that no other companies would do. Maybe some wild transceiver/antenna design that would let them get rid of plastic “windows” in cases. Or maybe software-defined radios that could be upgraded in software to work with newly vacant radio frequencies.
Size/power savings
The Qualcomm chip is a complex thing. It is so complex that it has its own ARM processor inside running its own little operating system. It takes quite a lot of space on the board:
Some of the code Qualcomm’s CPU is running could probably run on Apple’s main CPU. Apple’s CPU cache could also be reused. The remaining baseband-specific stuff could be implemented as separate blocks inside Apple’s chip package, just like Imagination’s GPU. That would cut down on space.
The power savings could be provided by elimination of redundant hardware across multiple chips, Apple’s P.A. Semi chip power-saving tricks, and Apple’s software better managing baseband power states.
Cost savings
Take a look at this table:
The Qualcomm chip costs about $32. It is one of the most expensive components. Apple could probably make their own wireless chip for $15–20. Or maybe integrate it into their next An chip so that the total cost is about $25–30. Of course, there are some tricky patent licensing costs involved, but it still may be more cost-efficient for Apple to do its own stuff.
Plus, if they’re going to do their own LTE chip, they’re surely going to do their own Wi-Fi/BT/GPS as well. That’s another $4 off.
Innovation opportunities
Apple would be in control of its wireless R&D schedule, aligning it to its own product roadmap.
They could start by moving much more aggressively along LTE Releases 10/11/12, partnering with select carriers around the world to do that. Sort of like Samsung has partnered with SK Telecom to launch 150Mbps LTE-A in South Korea.
They could also implement new LTE features that do not require carrier network upgrades, but do require control of all components; for example, “in-device co-existing”, which eliminates interference between LTE/Wi-Fi/BT radios by co-ordinating their transmission.
After a couple of R&D cycles, they could attempt innovations that no other companies would do. Maybe some wild transceiver/antenna design that would let them get rid of plastic “windows” in cases. Or maybe software-defined radios that could be upgraded in software to work with newly vacant radio frequencies.
Last edited:
Seriously, what is everyone wishfully smoking?
1. The two dudes have been hired in management positions. They are going to contribute jack **** to actual development.
2. With due respect to the two people hired, don't take this personally -- A "principal engineer" at Broadcom is not that senior a position, there are multiple levels of seniority above it. Just go check the profiles on LinkedIn. Chip lead is a senior position and it means the person is responsible enough to be given overall coordination for a chip. And that's more management work than technical work.
3. If Apple were really building their own baseband, they would need an army and would be aggressively going after the technical talent at Qualcomm, Broadcom, MediaTek or Intel. They would be hiring in the hundreds and you would see Apple presence in the various wireless standards organizations.
4. Building a chip isn't an one time thing. You build one, get it fabricated, fix issues, get it fabricated again, hopefully it works, then you push it to production and then go back to the drawing board to build the next one. That's why everyone stopped building their own CPUs and bought or licensed chips or designs from the likes or Intel, AMD and ARM. It's no different for Basebands. Or Wifi chips.
6. It takes 2+ years to take a chip to production. Possibly more if you are starting at the ground level.
I can think of two viable scenarios:
A. Apple may be licensing wireless IP from someone (possibly Broadcom or Qualcomm) and incorporating them in their own CPU dies for efficiency. And they need managers versed in wireless to manage the effort. They have probably hired project managers from various chip vendors before to get the A7 to market.
B. These guys get the fun job of beating up on their previous employers.
1. The two dudes have been hired in management positions. They are going to contribute jack **** to actual development.
2. With due respect to the two people hired, don't take this personally -- A "principal engineer" at Broadcom is not that senior a position, there are multiple levels of seniority above it. Just go check the profiles on LinkedIn. Chip lead is a senior position and it means the person is responsible enough to be given overall coordination for a chip. And that's more management work than technical work.
3. If Apple were really building their own baseband, they would need an army and would be aggressively going after the technical talent at Qualcomm, Broadcom, MediaTek or Intel. They would be hiring in the hundreds and you would see Apple presence in the various wireless standards organizations.
4. Building a chip isn't an one time thing. You build one, get it fabricated, fix issues, get it fabricated again, hopefully it works, then you push it to production and then go back to the drawing board to build the next one. That's why everyone stopped building their own CPUs and bought or licensed chips or designs from the likes or Intel, AMD and ARM. It's no different for Basebands. Or Wifi chips.
6. It takes 2+ years to take a chip to production. Possibly more if you are starting at the ground level.
I can think of two viable scenarios:
A. Apple may be licensing wireless IP from someone (possibly Broadcom or Qualcomm) and incorporating them in their own CPU dies for efficiency. And they need managers versed in wireless to manage the effort. They have probably hired project managers from various chip vendors before to get the A7 to market.
B. These guys get the fun job of beating up on their previous employers.
Not sure why you think that. Best engineering managers in my experience tend to be the ones who comprehend and understand the task so they fully understand what they're delegating. They tend to have extensive technical knowledge and are responsible for reviewing and approving the engineering that does occur.
Technical line managers, yes.
Program managers, which is what the linkedin profile says in this case, not as much. Their focus is entirely on the logistics of getting a product to market in a timely and efficient manner.
I know this thread is getting old but it occurred to me.. This probably isn't just about power savings. That will definitely be essential when they want to make a cellular version of the iWatch. But with reports that the new apple TV will have a wifi router/access point in it, It looks like Apple wants to reduce lag on their wireless display devices. They'll probably be working on improving the latency so the response of a tv will be close enough to real time that they can be used more like monitors. Current gen Apple TV's won't cut it for a good user experience. Imagine having one Mac Mini (or whatever they replace it with) acting as a master for every tv attached to an Apple TV. When you get close to an Apple TV (in the same room) it can detect the distance and transmit the experience of a desktop computer to that TV in your house. There's huge potential for a new experience in personal computing.
They could even use the computer to drive multiple desktop experiences simultaneously throughout the house, serving multiple people like a server.
They could even use the computer to drive multiple desktop experiences simultaneously throughout the house, serving multiple people like a server.
Last edited: