Info Intel Blames 5G Exit On Apple's Settlement With Qualcomm

Intels failure in the mobile market was an attempt in the futile. The general idea was to cement Intels position because compiling for multiple platforms is a pain in the duckuss and you only code for one (usually) In this case it's Intel or ARM based solutions. (Although MIPS is still hanging on, not in the mobile market) As Intel saw that everyone was moving to mobile devices, that threatened their long term dominance. Competing on the ARM level would mean low margins and marginalization in the marketplace full of ARM competitors. Hence Intel tried the "Give it to them for free and then make them addicted" approach for Atom for Android.

The problem being even with Intel's advanced process nodes, Atom was a power hog. CISC overhead dealing with memory addressing was a disaster of legacy support for low powered systems. It just had too many transistors causing it to loose in the efficiency/watt. And Intel spent I don't know how much money just rewriting the android OS to support it.

Add to this a number of Atom chips that had the memory interface degradation issues that caused them to burn out. This made Intel lose a fortune.

Intels problem is they want high margins. They are trying to protect stock price. And this has caused Intel to paint themselves into a corner where they are desperately trying to figure out how to deal with mind-share to stay relevant. They are getting attacked on all traditional strongholds. (Consumer desktop (AMD), HEDT (AMD), Mobile (ARM), Servers (AMD Epyc & NVIDIA AI & Amazon & Google. All these companies are making competing chips which deliver better value))

The whole reason Intel likely got into the 5G modem business is because there wasn't much in terms of a competitive landscape and they saw fat potential margins with Apple given the dispute. They said the 5G will live on for other markets (ie: Wireless cards for laptops). And I think this is a viable plan. They will continue to lose money, but the competitive landscape for 5G SOC modems is still small. What will bite intel in the tail is their limited 10nm production capability. Do they support low powered mobile first? Or do they push the 5G SOC modem biz? If there's a reasonable 10nm successor to Gimini Lake anytime soon will be a huge shock and surprise.

If I were Intel I would start supporting 5G tower infrastructure including hardware to handle the data increases. That way they could have a claim of End to End Support. And the 5G infrastructure market will be huge. (Even if divided up)

The cutback on the SOC 5G Modem is another example where Intel is ditching funding on another division where they could make inroads, but at great cost. They have killed so many potential products. I guess they are looking at their Atom initiative for mobile and are afraid of losing more money. But at this point, do they have an option with declining market-share?

I don't even think Intel would be pushing graphics so hard (very low margin) if it weren't for the fact that I think APU's and Data centers will deliver much better value than Intel's offerings in a few years because of the lack of MIMD/Matrix operation raw GFLOP throughput. These operations are critical for not only games, but AI as well.

If I did individual stocks, I would short Intel over the next 5 years. I might even short NVIDIA. AMD is in a great position in the next couple of years in terms of APU and AI performance based on a combined ecosystem (Synergy) of APU + CPU/GPU. NVIDIA doesn't have much in terms of CPU performance. NVIDIA is offering a "Me too" ARM processor. Intel is playing catch up on graphics.
 
Last edited:

mlee 2500

Honorable
Oct 20, 2014
298
6
10,785
I like Intels chances in the Graphics space allot better then the odds they faced trying to get into mobile. To add to your litany of shortcomings, they were also LATE to the mobile game. Like many old school PC reliant semiconductor shops, their arrogance led themt to believe mobile was just a fad, and then when it became obvious (with tablets) that it was more then that, that it would "never displace PC's" (they were only part right...PC's stuck around but lost a chunk of marketshare to tablet and phone screens). Intel even divested themselves of some of their own competitive pre-existing ARM IP, products and other mobile efforts early on, they were so sure. When they finally did get serious it was too late...ARM based platforms already had an insurmountable lead.

Given that failure was in the books, they were actually pretry smart to then pivot and focus on the commerical enterprise and datacenter market, which had more then enough new growth and higher profit margin oppertunities to more then make up for what they lost surrendering so much of the previously PC based business.

GPU's are different though...it's basis lies within their core strength (x86 platforms) and it's market spans it's existing customer base since GPU's have a play both in the Datacenter as well as being one of the few remaining higher margin components in an otherwise highly commoditized PC and workstation market.

If they stay serious I see them very quickly reaching product parity and becoming extremely competitive with NVIDIA and AMD in the GPU arena. Don't forget, by virtue of owning half or more of the CPU market, they have natural leverage, especially if part of their line is an SOC based combination to win the lower and mid-tier segments...which will then drive platform support and popularity on the higher discreet end of GPU products as well.


Intels failure in the mobile market was an attempt in the futile. The general idea was to cement Intels position because compiling for multiple platforms is a pain in the duckuss and you only code for one (usually) In this case it's Intel or ARM based solutions. (Although MIPS is still hanging on, not in the mobile market) As Intel saw that everyone was moving to mobile devices, that threatened their long term dominance. Competing on the ARM level would mean low margins and marginalization in the marketplace full of ARM competitors. Hence Intel tried the "Give it to them for free and then make them addicted" approach for Atom for Android.

The problem being even with Intel's advanced process nodes, Atom was a power hog. CISC overhead dealing with memory addressing was a disaster of legacy support for low powered systems. It just had too many transistors causing it to loose in the efficiency/watt. And Intel spent I don't know how much money just rewriting the android OS to support it.

Add to this a number of Atom chips that had the memory interface degradation issues that caused them to burn out. This made Intel lose a fortune.

Intels problem is they want high margins. They are trying to protect stock price. And this has caused Intel to paint themselves into a corner where they are desperately trying to figure out how to deal with mind-share to stay relevant. They are getting attacked on all traditional strongholds. (Consumer desktop (AMD), HEDT (AMD), Mobile (ARM), Servers (AMD Epyc & NVIDIA AI & Amazon & Google. All these companies are making competing chips which deliver better value))

The whole reason Intel likely got into the 5G modem business is because there wasn't much in terms of a competitive landscape and they saw fat potential margins with Apple given the dispute. They said the 5G will live on for other markets (ie: Wireless cards for laptops). And I think this is a viable plan. They will continue to lose money, but the competitive landscape for 5G SOC modems is still small. What will bite intel in the tail is their limited 10nm production capability. Do they support low powered mobile first? Or do they push the 5G SOC modem biz? If there's a reasonable 10nm successor to Gimini Lake anytime soon will be a huge shock and surprise.

If I were Intel I would start supporting 5G tower infrastructure including hardware to handle the data increases. That way they could have a claim of End to End Support. And the 5G infrastructure market will be huge. (Even if divided up)

The cutback on the SOC 5G Modem is another example where Intel is ditching funding on another division where they could make inroads, but at great cost. They have killed so many potential products. I guess they are looking at their Atom initiative for mobile and are afraid of losing more money. But at this point, do they have an option with declining market-share?

I don't even think Intel would be pushing graphics so hard (very low margin) if it weren't for the fact that I think APU's and Data centers will deliver much better value than Intel's offerings in a few years because of the lack of MIMD/Matrix operation raw GFLOP throughput. These operations are critical for not only games, but AI as well.

If I did individual stocks, I would short Intel over the next 5 years. I might even short NVIDIA. AMD is in a great position in the next couple of years in terms of APU and AI performance based on a combined ecosystem (Synergy) of APU + CPU/GPU. NVIDIA doesn't have much in terms of CPU performance. NVIDIA is offering a "Me too" ARM processor. Intel is playing catch up on graphics.
 
Last edited:
  • Like
Reactions: digitalgriffin