Intel Will Exit 5G Phone Modem Business, Hours After Apple and Qualcomm Settle Licensing Dispute (cnet.com) 114
Intel announced Tuesday afternoon that it will no longer be working on 5G chips for smartphones, leaving Apple with only one supplier for its iPhones, Qualcomm -- the same company that it was battling in court until midday Tuesday. CNET reports: Intel late Tuesday said it plans to exit the 5G smartphone modem business. It had been working on a processor for Apple, with the chip expected to be in iPhones in 2020. Lately there have been worries the chip wouldn't be ready until iPhones released in 2021. "The company will continue to meet current customer commitments for its existing 4G smartphone modem product line, but does not expect to launch 5G modem products in the smartphone space, including those originally planned for launches in 2020," Intel said in a press release. Its only customer in modems is Apple.
Intel added that it will "complete an assessment of the opportunities for 4G and 5G modems in PCs, internet of things devices and other data-centric devices." It also said it will "continue to invest in its 5G network infrastructure business." "We are very excited about the opportunity in 5G and the 'cloudification' of the network, but in the smartphone modem business it has become apparent that there is no clear path to profitability and positive returns," Intel CEO Bob Swan said in a statement. The announcement comes hours after Apple and Qualcomm announced that they had reached a settlement in their multi-year battling over licensing royalties.
Intel added that it will "complete an assessment of the opportunities for 4G and 5G modems in PCs, internet of things devices and other data-centric devices." It also said it will "continue to invest in its 5G network infrastructure business." "We are very excited about the opportunity in 5G and the 'cloudification' of the network, but in the smartphone modem business it has become apparent that there is no clear path to profitability and positive returns," Intel CEO Bob Swan said in a statement. The announcement comes hours after Apple and Qualcomm announced that they had reached a settlement in their multi-year battling over licensing royalties.
Urg (Score:5, Insightful)
I hate to see qualcom's sleazy Frand practices win by default.
Re:Urg (Score:5, Insightful)
This might explain the cryptic announcement in November by Qualcom that they felt that they and apple were close to negotiating an agreement. Apple responded by saying "huh? we haven't talked with Qualcom in 6 months".
How could Qualcom know apple would be agreeing if they were not talking? Maybe they somehow had inside info and knew Intel wasn't going to deliver on 5G so apple would me coming home to mama qualcom asking for forgiveness.
Re: (Score:2)
Or maybe they were just talking out of their ass.
Re:Urg (Score:4, Informative)
Maybe they somehow had inside info and knew Intel wasn't going to deliver on 5G
Anyone who has looked into the sad history of Intel side projects, especially those trying to get a foothold into the mobile / low power world had that "inside info".
Re: (Score:2)
Anyone who has looked into the sad history of Intel side projects, especially those trying to get a foothold into the mobile / low power world had that "inside info".
Quoting for truth.
Re: (Score:2)
Apple can always lower their prices if they want to give Qualcom less.
Re: Urg (Score:1, Interesting)
What's unreasonable about a percent? Apple is using Qualcomm inventions to extort people. And yes, they are real inventions, not apple's rounded corners.
Re: (Score:1)
Go learn about FRAND before you post.
Re: Urg (Score:4, Informative)
Please point us to the part in the definition of FRAND that says that you cannot set the license fee based on the price of the final product. I read the definitions of FRAND several times and I just can't find that part.
I'll just leave this here (Score:1)
Unlocking a device by performing gestures on an unlock image
https://patents.google.com/pat... [google.com]
Re: (Score:3)
What's unreasonable about a percent? Apple is using Qualcomm inventions to extort people. And yes, they are real inventions, not apple's rounded corners.
Ignoring the "rounded-corners" trollbait, the honest answer to your question is that Qualcomm's patent enables the chips, which are just one component of the phone. If Apple adds a better screen, better camera, better CPU/GPU, and more RAM, why should that cause Qualcomm to get more money for one of the chips?
Apple's position has been that the chip suppliers should themselves retain Qualcomm licenses and sell Apple the chips wholesale, incorporating the price of the licensing. So it's not like the chips the
Re: (Score:3)
If Apple adds a better screen, better camera, better CPU/GPU, and more RAM, why should that cause Qualcomm to get more money for one of the chips?
Apple takes a 30% cut (AFAIK) for all apps sold on the Apple Store. No matter whether the app costs $0.99 or $9.99. Why does Apple set a rate based on the final price and not a fixed price per license?
The same (once again: AFAIK) for music, videos, books, news papers etc. sold through Apple. Why should Apple get more money per issue sold from a high-price high-quality news paper than from a cheap rag? The service that Apple provides for both is basically the same.
As long as the percentage that Qualcomm asks for per unit isn't unreasonably high and as long as the percentage is the same for all customers of Qualcomm (taking into account the question of whether the customer themselves has patents that are part of the standard) I don't see how Qualcomm violates the FRAND rules.
Re: (Score:2)
If Apple adds a better screen, better camera, better CPU/GPU, and more RAM, why should that cause Qualcomm to get more money for one of the chips?
Because their expensive luxury phone wouldn't be an expensive luxury phone if it didn't have a top spec Qualcomm chip in it. The fact that they tried to get away with giving Qualcomm trade secrets to Intel so they could improve their modems tells you how important that chip is to the iPhone.
Also note that Apple clearly doesn't think a percentage is unfair, given that it takes 30% on all sales on its platforms.
When did Intel become such a quitter? (Score:4, Insightful)
Oh yeah it was when they first dropped out of the Graphics Card Market in the late 90s, then followed up by dropping all those 'low margin' embedded systems and memory parts whose market they used to dominate.
Intel stopped being competitive when the MBAs decided they should focus on their 'core market' of desktop and network chips and ignored the fact that that BREADTH of ecosystem benefitted them in less tangible ways, like providing less complicated products to test agains, and to get the R&D and engineering experience from a wide variety of problems, optimizations or test procedures of which might have benefits in other aspects of their business... like when they had that SATA controller failure a few years back. Intel had dealt with similar issues on other process technologies in the past on their embedded systems processors. Having access to all that extra test data from even higher volume products with simpler and easier to debug logic has untold benefits when shifting process technologies. In the past Intel did that across their product lines, before focusing so much on their CPUs that they compromised their own design and testing infrastructure by having CPUs become the leader of process technologies instead of trialling it on other parts first (I think they still do flash on the latest processes, but in the past it could have been flash, ram, or certain high margin embedded controllers where debugging intermittent failures and getting clearer feedback for modifications/updates to process models was easier and clearer to do. I imagine a lot of the 10nm shortcomings are related to exactly this change in technlogical leadership. Intel's time came and went, and unless someone absolutely slaughters the board and sensior leadership to reinstate engineers with some management competence, instead of MBAs with no engineering competence, Intel will continue to falter and eventually fall, no matter how much business they seem to have today.
Re: (Score:2)
Intel were never leaders in any of those markets though, just also-rans who were able to leverage their existing relationships.
Intel GPUs have always, and probably will always be shit. They are okay for basic integrated graphics but they will never be competitive on performance or low power.
Intel CPUs and chipsets were only "low power" back in the early 2000s when the Pentium Mobile was impressive, nowadays they are not in the same league as ARM and their ecosystem.
Intel tried, and Intel failed. All they ar
Re:Urg (Score:5, Informative)
.
Read Demerjian's whole piece here for a more complete picture:
https://www.semiaccurate.com/2... [semiaccurate.com]
Re: (Score:2)
Apple have always been dicks when it comes to FRAND stuff, because their own patent portfolio is shit and no-one wants to licence it. The usual way FRAND works is both parties agree to cross-licence patents and no money changes hands, but who wants Apple's patents on rounded corners?
Their other problem is that they want at least two sources for all components. In this case they noticed that Intel modems were inferior to Qualcomm ones, and decided to help Intel out by handing over some Qualcomm trade secrets
TLDR, reader's digest version : (Score:1)
Intel sucks at making new chips and they're giving up.
Opens door for AMD or Huawei or Nokia? (Score:3)
Apple and Samsung should insist that Qualcom license their in-name-only FRAND patents to AMD as a second source. How did these major companies manage to make thensleves dependent on Qualcom? THe only other leverage now will be either the EU or China. I could see China insiting Qualcom lend their patents to Huawei and I cold see the EU insisting Qualcom also allow second sourcing.
Re: (Score:3)
Huawei (Score:3)
Huawei holds more 5G patents than Qualcomm and many of their patents are critical to the 5G standard. Qualcomm and Huawei have crosslicensed.
Re: (Score:1)
Why should Samsung care? They have their own 5G modem - Exynos Modem 5100. Why should Huawei care? They have their own 5G modem - Balong 5000. Why should Nokia care? The real Nokia is making cell towers, so they have even bigger 5G modems too. The spin-off for handsets doesn't want to make their own chips.
Re: (Score:2)
It sure helps businesses innovate! /s
(Germany was once called the land of thinkers and poets. That was during a time where it had no copyright-like laws. While the UK, who had them already, fell into an information dark age. Lesson learned, right?)
The phrase was, in German "Das Land der Dichter und Denker".
In the middle of the 20th Century this was transmuted into "Das Land der Richter und Henker" (the land of judges and executioners).
What's wrong with Intel? (Score:3)
What's wrong with Intel? They had a great streak after that Netburst fiasco (which took some very illegal anti-competitive practices from their part to survive pretty much unscathed), but it's been several years now that they seem to be struggling with new things. AMD finally caught up with them, their new fab process didn't pan out, they've been trying to enter the smartphone business in various ways and it seems they are always failing...
Re: (Score:3)
Too many bean counter CEOs.
Their CEOs used to be EEs.
Now they are financial "geniuses" who know jack shit about the business they are supposedly running.
Re: (Score:3)
It is the whole battle of profit versus profitability. So say you invest 1 million dollars in a business and make 20% profit, bean counters will argue it is more profitable to only invest 100,000 dollars in the business and borrow the rest at 10% interest because they will calculate the profitability against the 100,000, even though you are now making less having to pay off a 900,000 dollar loan. The economist high priest will make the claim you can invest your 900,000 dollars in 9 other business and borrow
Re: (Score:2)
It's also the surety of securities rate of return vs. the risk of loss on product innovation.
If you have $10 billion in capital, you can invest in securities and get 10% or you can risk some or all of it and possibly lose it all if your innovation efforts fail. Even if they succeed, you may only partly succeed and your innovation may still produce a rate of return lower than the securities market.
So you wind up being better off not innovating at all, and just investing your capital in securities markets.
Re: (Score:2)
People are buying AMD and they are selling well. It's partially due to the new competitiveness of their chips. But I think the main reason is the global chip shortage from Intel. Currently only the big three SIs, Dell, HP and Lenovo can seem to get all the chips they need. Everyone else has had to start buying AMD to meet their production numbers.
And I do think it is inevitable that as a company grows in size and wealth it will eventually get into the financial market. We already see it with Apple and thei
Re: (Score:2)
I wonder how well they are selling in corporate servers. It's a clusterfuck to mix AMD and Intel CPUs under VMware.
Re: (Score:1)
That sounds a little like letting the head of Purchasing run the company.
Re: (Score:2)
AMD has their own problems ;(
I just bought (and had to return) an amd rizen 1700 system.
make -j16 with either silent errors or sigsegv.
no known cure, and bug was reported to amd back in 2017!
they suggest disabling ALL smt threads (-j16 becomes -j8) and disabling many other things. its fucked.
I tried using AMD and its useless for a build server. sorry, but I have to return to intel even though it pains me so.
I'll check back again in a year or two, but for now, I do NOT trust AMD for compute/build servers.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
When you think about it, Intel has never really been that competitive. That's why they cheat so much.
It's been that way since the 286 at least.
Re: (Score:2)
Re: (Score:2)
Intel cheated on some benchmarks, by making their compiler disable the fastest code paths when non-Intel chips were detected.
I miss 68k and PPC.
Re: (Score:2)
Re: (Score:2)
They made the compiler produce multiple code paths for some sections, e.g. one that used SSE instructions and one that didn't. But the SSE code was not used on CPUs that had SSE but were not made by Intel, artificially reducing their competitiveness in those benchmarks.
I seem to recall they justified it with some bullshit about SSE being a proprietary Intel thing and not testing other CPUs for compliance so they couldn't guarantee correct behaviour blah blah.
Re: (Score:2)
Re: (Score:2)
I'm not sure we are on the same page here. Their compiler was not being benchmarked, the compiled code was. The compiled code checks the CPU for supported instructions and manufacturer during start-up, and from then on the overhead is going to be at or very near zero to select the right code path. I'm not sure how they do it, store a flag somewhere or just re-write the call instructions.
Anyway, it was shown that if you killed the manufacturer check and performance on rival CPUs was similar to Intel ones and
Re: (Score:2)
Re: (Score:2, Troll)
Okay, I see your point. I don't agree though, if the CPU says it supports those instructions they should be used, and if something goes wrong it's the CPU's fault for lying.
Re: (Score:2)
The 6502 was the processor that people who thought the M6800 was too expensive bought.
The Z80 was the processor that people who thought the 8080 was too expensive bought.
I own QCOM stock (Score:1)
QCOM, INTC, and AAPL all up (Score:2)
QCOM is up 12.25%, but INTC is also up 3.26%. AAPL up 1.95%.
If you want a phone that fit in a trailer (Score:2)
These company nowadays mostly make infrastructure equipment (cell towers, etc.)
So yes, they have modem chips.
That fit in a trailer, require at least a small generator to run and can cover a whole city block with thousands simultaneous conditions.
Not exactly what Apple is looking to put inside their next upcoming smartphone.
Re: (Score:2)
Except when the prices of already too expensive phones rise because there's no competition in the design/fab space.
Re: (Score:3)
The patent fees were always a percentage of retail price, being the core technology that makes a "cellphone" a "cellphone".
The increased phone price wasn't caused by 5G patent fees.
Apple didn't like the fact that Qualcomm got more money when they switched a 128GB NAND chip for a 512GB chip and charged the customer $300 more for what is effectively a $30 SD card.
Re: (Score:2)
A percentage of retail price or of $500 (now lowered to $400) - whichever is lowest. Increasing retail price from $700 to $1000 would not increase the licensing fee.
I believe Apples beef was that Qualcomm requires companies who buys their chips to also license all their patents. It's a kind of double dipping.
Re: (Score:2)
Perhaps they won't care which specific 5G chips are in their phones, but they'll certainly care about the inevitable cost inflation that appears when there isn't sufficient competition among chip manufacturers.
Re: (Score:2)
I wonder what this means... (Score:5, Interesting)
You can certainly slap a 3rd party cell modem card into an x86(it's a standard option on a fair percentage of laptop lines); but that is considerably less compact than the ARM SoC option; which is a minus for space constrained applications. It's also likely to be more expensive and power hungry, since peripheral integration usually ends up being helpful on those counts.
Given that, it seems like Intel is either really pessimistic about their situation, enough so that they don't think they can even justify a pet cell modem aggressively sold along with their chips and wifi/bt silicon(either just because the R&D isn't going so well or because they suspect the patent litigation will be hideous); or they are fairly optimistic about Qualcom being more cooperative in the future and being willing to license modems for integration at rates reasonable enough that it's simply not worth reinventing the wheel.
I'm just not sure which. It doesn't help that Apple's main possible motives point in the same two directions: either a belief that the patent situation is bad enough that they'll get hammered in court/import bans/etc. even if they cultivate a secondary supplier; or a belief that Qualcom's position is weakening and they are likely to be cooperative enough on pricing and not shaking people down on patents that there's no reason to turn down their parts unless something genuinely superior shows up(which, so far, it hasn't).
Any guesses?
Re: (Score:1)
One type of production line, one design.
That allows for everyone to buy one size product and work it into their design over a generation of product.
Who wants to support 2 products in one generation of smart phone?
Everyone just wants the most easy product that gives the needed support for the hardware.
Then build the "new" product around that standard global hardware. Add an OS and GUI.
They just cannot hack it (Score:5, Interesting)
This is a bit surprising not so much because Intel has been having such a good time in phone silicon(since they haven't); but because I would have assumed that Intel would have considered an at least adequate cell modem to be essential
They probably do consider that essential.
But the fact is, they just cannot do as good a job as Qualcomm can. Losing Apple meant that there was no way they could fund the years required for Intel to build up the expertise needed.
Re: (Score:3)
It's no small task to build a 5G cellular modem. Getting it certified world-wide is extremely expensive and time consuming too.
There are patent problems as well. Huawei and other Chinese companies hold a lot of 5G essential patents, so Intel either has to pay them or licence some of its own patents in exchange. So deep cuts into profit margins or licence valuable tech to Chinese competitors.
I get the impression that Intel is giving up on the really low cost, highly integrated, low power IoT side of things.
Re: (Score:1)
It was well known years ago in the inner circles of the industry that Intel couldn't possibly cover the huge R&D costs for modems by making modems just for Apple. While qualcomm charges a lot they are smart enough to not charge so much to allow a competitor to cover the design costs for a modem by picking up a single customer. The people I knew who understood the modem industry were fairly confident that Intel's phone modem would eventually fail. That became almost a sure thing when it was clear more
Re: (Score:2)
but because I would have assumed that Intel would have considered an at least adequate cell modem to be essential for purposes of selling their CPUs and chipsets for 'IoT' and embedded stuff; as well as 'Centrino' style chipset bundling.
I have a guess. Intel like Microsoft are completely incompetent in the world outside their core business. Their attempts at complete integration have mostly failed. Their attempts at IoT have mostly failed. Their attempts at mobile have mostly failed. Their attempts at low power have mostly failed (I'm not talking about laptops here, but really low power).
Making any assumptions which require strategic or management competence here is foolish.
Only one Supplier? (Score:2)
Is that only one supplier that Apple can work with or is there only one company in the world making 5G modems for smartphones?
Re: (Score:2)
Huawei
Re: (Score:2)
That's what happens when you no longer want the cheapest, but only want "American". All the American companies bow out because they can't profit from it.
That's also what happens when Apple screw over their suppliers and go looking for another partner "out of spite"... the other suppliers take you at your word and believe you'll do the same to them, so they ask you to pay upfront and design *EXACTLY* to your specs and then quote you a price for that.
I thought Apple were going to do everything "in-house" any
Real reason (Score:5, Insightful)
Intel only wanted to do it because Apple guaranteed demand and gave them Qualcomm's IP to use.
No one else wants to pay Qualcomm patent license fees just to use an inferior chipset from Intel.
Cloudification of the network (Score:1)
Whatever the fuck that means...