Apple's Acquisition of Intel's Smartphone Modem Business Completed, Intel Admits 'Multi-Billion Dollar Loss'

[MacRumors]

Intel today announced it has completed the sale of the majority of its smartphone modem business to Apple for $1 billion following regulatory approval. The transaction was first announced in July and includes intellectual property, equipment, and approximately 2,200 Intel employees joining Apple.

The deal sees Apple acquire a large portfolio of wireless patents from Intel. Apple now holds over 17,000 wireless technology patents, ranging from protocols for cellular standards to modem architecture and modem operation.

Intel will retain the ability to develop modems for non-smartphone applications, such as PCs, internet-of-things devices, and autonomous vehicles.

Last week, Intel admitted that it sold its smartphone modem business to Apple at "a multi-billion dollar loss," according to court documents unearthed by Reuters. Intel added that rival chipmaker Qualcomm's patent licensing practices "strangled competition" and effectively forced it to exit the market.

Apple is expected to use Qualcomm modems for its first 5G-enabled iPhones next year, as part of a six-year licensing agreement between the companies. Farther down the road, multiple reports have claimed that Apple plans to develop its own modems for iPhones by 2022-23, and this Intel deal would certainly help those efforts.

Article Link: Apple's Acquisition of Intel's Smartphone Modem Business Completed, Intel Admits 'Multi-Billion Dollar Loss'

 

[Razeus]

Still want Qualcomm chips though.

 

[guzhogi]

I'm curious to see what kind of performance an Apple-made modem would have, assuming Apple makes one.

 

[Bandaman]

I'm curious to see what kind of performance an Apple-made modem would have, assuming Apple makes one.

Hopefully it's as good as their A-series chips. Can't be any worse than Intel's offerings.

 

[jz0309]

No surprises there.
I also think Apple designed modems will be several years out

 

[fokmik]

So now, starting with 2021 probably we will have macs with 5G built in

 

[mag33]

Finally every 5G fanatic can finally shut up

 

[Plutonius]

At least Intel makes better processors than AMD so this mistake should not hurt them that much (i.e. their computer chip business will make up the loss).

 

[ghanwani]

Hopefully it's as good as their A-series chips. Can't be any worse than Intel's offerings.

I don't think this is a valid comparison. The A-series chip is something completely hidden from the external world. On the other hand, building a cutting edge modem requires making calculated calls on where the industry will move and having deep knowledge of the relevant standards, and any non-standard enhancements, that would allow interoperability with carrier equipment. Since Apple doesn't build carrier equipment, they will likely always be trailing the competition.

 

[Wide opeN]

I'm sure like most things Apple will be conservative in any Modem roll outs en masse on their phones.

 

[Rigby]

I wouldn't be surprised if we saw Apple using Intel-derived modems in a two-chip solution as early as 2020 or 2021, although it will take them longer to integrate a modem with their CPU and other components in a SoC. From what I heard the Intel 5G modem wasn't too far from completion when the Apple acquisition was announced.

In any case, Qualcomm has solidified its stranglehold on the market for now. Will be interesting to see what the big players other than Apple will do (primarily Huawei and Samsung).

 

[gank41]

With the inclusion of Intel’s mobile chip IP in the purchase, does this mean there could be potential improvements with existing/older iPhones powered by Intel chips?

 

[mannyvel]

On the other hand, building a cutting edge modem requires making calculated calls on where the industry will move and having deep knowledge of the relevant standards, and any non-standard enhancements, that would allow interoperability with carrier equipment. Since Apple doesn't build carrier equipment, they will likely always be trailing the competition.

The wireless standards track is well-known by all the players, and there aren't really that many choices you have to make. And Apple has some experience being on the cutting edge of wireless: they were the first vendor to ship 802.11. And in any case the standards are, well, standards. No equipment vendor in their right mind wouldn't allow Apple into their interoperability labs.

The real question, though, is if it's really worth it for Apple to be in the modem business at all. That business relies on economies of scale and a lot of engineering and wireless expertise.

I'd guess the goal is to do what it did with Bluetooth and the W1, making the modem low-power. Radios are a big power draw, and I doubt Qualcomm or Intel were willing to make a power/performance tradeoff. At some point that could become yet another competitive advantage for Apple.

I mean, they could probably save a bunch of power just by optimizing the modem's software stack. Most of those modems are driven using the Hayes AT command set, which means you need some kind of serial emulator. Apple could rip out that interface layer and go direct to the modem, which would make it more efficient, which should lead to better power/performance.

If they did it for BT/WiFi they should be able to do it for LTE/5G/6G and beyond.

 

[fmcshan]

Great! Now let’s make sure that we get our 5G in the 2020 iPhone 🙂

 

[Hazmat401]

At least Intel makes better processors than AMD so this mistake should not hurt them that much (i.e. their computer chip business will make up the loss).

What rock have you been stuck under.... AMD is crushing intel!

the only warfront AMD is struggling against is Nvidia

 

[0924487]

I don't think this is a valid comparison. The A-series chip is something completely hidden from the external world. On the other hand, building a cutting edge modem requires making calculated calls on where the industry will move and having deep knowledge of the relevant standards, and any non-standard enhancements, that would allow interoperability with carrier equipment. Since Apple doesn't build carrier equipment, they will likely always be trailing the competition.

They will collaborate with HUAWEI
[automerge]1575307748[/automerge]

The wireless standards track is well-known by all the players, and there aren't really that many choices you have to make. And Apple has some experience being on the cutting edge of wireless: they were the first vendor to ship 802.11. And in any case the standards are, well, standards. No equipment vendor in their right mind wouldn't allow Apple into their interoperability labs.

The real question, though, is if it's really worth it for Apple to be in the modem business at all. That business relies on economies of scale and a lot of engineering and wireless expertise.

I'd guess the goal is to do what it did with Bluetooth and the W1, making the modem low-power. Radios are a big power draw, and I doubt Qualcomm or Intel were willing to make a power/performance tradeoff. At some point that could become yet another competitive advantage for Apple.

I mean, they could probably save a bunch of power just by optimizing the modem's software stack. Most of those modems are driven using the Hayes AT command set, which means you need some kind of serial emulator. Apple could rip out that interface layer and go direct to the modem, which would make it more efficient, which should lead to better power/performance.

If they did it for BT/WiFi they should be able to do it for LTE/5G/6G and beyond.

Yes it does, communication silicons are integral to everything to come in the future as connected things become more and more distributed. Even for AirPods’ W1 and H1 chips, they are using the same engineering foundations as the modern technology, which is just a manifestation of a domain of skills Apple can greatly improve on.

 

[JPack]

The wireless standards track is well-known by all the players, and there aren't really that many choices you have to make. And Apple has some experience being on the cutting edge of wireless: they were the first vendor to ship 802.11. And in any case the standards are, well, standards. No equipment vendor in their right mind wouldn't allow Apple into their interoperability labs.

The real question, though, is if it's really worth it for Apple to be in the modem business at all. That business relies on economies of scale and a lot of engineering and wireless expertise.

I'd guess the goal is to do what it did with Bluetooth and the W1, making the modem low-power. Radios are a big power draw, and I doubt Qualcomm or Intel were willing to make a power/performance tradeoff. At some point that could become yet another competitive advantage for Apple.

I mean, they could probably save a bunch of power just by optimizing the modem's software stack. Most of those modems are driven using the Hayes AT command set, which means you need some kind of serial emulator. Apple could rip out that interface layer and go direct to the modem, which would make it more efficient, which should lead to better power/performance.

If they did it for BT/WiFi they should be able to do it for LTE/5G/6G and beyond.

It's largely about the implementation. If it were as easy as following standards, everyone would be as far ahead in 5G as Huawei. And MediaTek and Samsung's 5G baseband would be equal to Qualcomm's.

Intel chose to implement their LTE and 5G basebands using DSPs and x86 cores instead of being hardwired. It's one of the reasons why their modems use more power compared to Qualcomm.
[automerge]1575308302[/automerge]

I don't think this is a valid comparison. The A-series chip is something completely hidden from the external world. On the other hand, building a cutting edge modem requires making calculated calls on where the industry will move and having deep knowledge of the relevant standards, and any non-standard enhancements, that would allow interoperability with carrier equipment. Since Apple doesn't build carrier equipment, they will likely always be trailing the competition.

Agreed. A-series chips are built on the backbone of existing ARM developments. More importantly, Apple can customize features for iOS and remove backwards compatibility, such as 32-bit blocks. They can't do the same for a modem and remove 3G, for example.

 

[Breaking Good]

The wireless standards track is well-known by all the players, and there aren't really that many choices you have to make. And Apple has some experience being on the cutting edge of wireless: they were the first vendor to ship 802.11. And in any case the standards are, well, standards. No equipment vendor in their right mind wouldn't allow Apple into their interoperability labs.

The real question, though, is if it's really worth it for Apple to be in the modem business at all. That business relies on economies of scale and a lot of engineering and wireless expertise.

I'd guess the goal is to do what it did with Bluetooth and the W1, making the modem low-power. Radios are a big power draw, and I doubt Qualcomm or Intel were willing to make a power/performance tradeoff. At some point that could become yet another competitive advantage for Apple.

I mean, they could probably save a bunch of power just by optimizing the modem's software stack. Most of those modems are driven using the Hayes AT command set, which means you need some kind of serial emulator. Apple could rip out that interface layer and go direct to the modem, which would make it more efficient, which should lead to better power/performance.

If they did it for BT/WiFi they should be able to do it for LTE/5G/6G and beyond.

mannyvel, you seem to have a very good grasp of the technology involved. Could you expand on your post above for us lay people to better understand?

 

[konqerror]

You have a 10 year old view of the smartphone modem.

The real question, though, is if it's really worth it for Apple to be in the modem business at all. That business relies on economies of scale and a lot of engineering and wireless expertise.

Apple has to if they want to integrate the baseband into the app processor, like Qualcomm and Samsung are doing. This saves not only board space and cost, it makes a impact to battery life particularly because separate basebands require their own memory or a power-hungry high-speed link to system RAM.

Currently, Apple has to put down 3 chips, where Qualcomm integrates cellular, Wi-Fi and app processor into one.

I mean, they could probably save a bunch of power just by optimizing the modem's software stack. Most of those modems are driven using the Hayes AT command set, which means you need some kind of serial emulator. Apple could rip out that interface layer and go direct to the modem, which would make it more efficient, which should lead to better power/performance.

Hasn't been that way physically or logically for more than 5 years. They use shared memory interfaces, just like an Ethernet NIC. I think the latest ones are logically PCIe over MIPI M-PHY. Before that I recall they often used an Ethernet-over-USB standard and ran USB over M-PHY or a reduced power version of USB (HSIC).

The wireless standards track is well-known by all the players, and there aren't really that many choices you have to make. And Apple has some experience being on the cutting edge of wireless: they were the first vendor to ship 802.11.

Actually no. The motivation for companies in the standards process is to force everybody to use their own patented technology, or else keep the standard as non-specific as possible so they can have a competitive advantage.

To take Wi-Fi as an example, the spec tells you how to communicate what transmit speed and how many MIMO channels you use (MCS), however, the exact algorithm to determine those parameters is unspecified. Each vendor has proprietary technology to evaluate the radio channel and select the right parameters, and you can see this in that two devices communicating will select different rates.

Intel chose to implement their LTE and 5G basebands using DSPs and x86 cores instead of being hardwired. It's one of the reasons why their modems use more power compared to Qualcomm.

Qualcomm does the same. They use ARM and their own DSP platform, Hexagon/QDSP. The iPhone's Intel chips were on ARM until last year, 7 years after they bought the operation from Infineon.

 

[dugbug]

I don't think this is a valid comparison. The A-series chip is something completely hidden from the external world. On the other hand, building a cutting edge modem requires making calculated calls on where the industry will move and having deep knowledge of the relevant standards, and any non-standard enhancements, that would allow interoperability with carrier equipment. Since Apple doesn't build carrier equipment, they will likely always be trailing the competition.

The a series chip is not isolated / hidden. The machine code follows a particular expected behavior exposed to compilers.

Also the advantage apple has with modem technology is to move as much of it as they can into their A-series silicone and get power optimizations

 

[V_Man]

Im just thrilled that my phone will now work in areas with weak cellular signal. That’s my biggest gripe. Service is great in area with strong service

 

[FlyingDutch]

Qualcomm is basically killing the market...
I’m very intrigued by Apple solution in 2022. When they entered the market for CPU and GPU in the past years, they literally annihilated competitors.

 

[MrGIS]

At least Intel makes better processors than AMD so this mistake should not hurt them that much (i.e. their computer chip business will make up the loss).

Umm... No. AMD is killing Intel on the CPU front. To the point you have to wonder when Apple might start putting AMD inside..

 

[sub150]

Umm... No. AMD is killing Intel on the CPU front. To the point you have to wonder when Apple might start putting AMD inside..

AMD doesn't kill Intel when you look at power consumption which is relatively important on a Macbook.

 

[GFLPraxis]

The wireless standards track is well-known by all the players, and there aren't really that many choices you have to make. And Apple has some experience being on the cutting edge of wireless: they were the first vendor to ship 802.11. And in any case the standards are, well, standards. No equipment vendor in their right mind wouldn't allow Apple into their interoperability labs.

The real question, though, is if it's really worth it for Apple to be in the modem business at all. That business relies on economies of scale and a lot of engineering and wireless expertise.

I'd guess the goal is to do what it did with Bluetooth and the W1, making the modem low-power. Radios are a big power draw, and I doubt Qualcomm or Intel were willing to make a power/performance tradeoff. At some point that could become yet another competitive advantage for Apple.

I mean, they could probably save a bunch of power just by optimizing the modem's software stack. Most of those modems are driven using the Hayes AT command set, which means you need some kind of serial emulator. Apple could rip out that interface layer and go direct to the modem, which would make it more efficient, which should lead to better power/performance.

If they did it for BT/WiFi they should be able to do it for LTE/5G/6G and beyond.

I think it absolutely is worth it for Apple, because it becomes a lot easier for them to customize the chips inside their SoC. Remember, Apple has to integrate Qualcom's modem designs into their chips.