Why can't Intel break into the mobile market?

[holden j caufield]

Are there any mainstream phones than have or will use intel chipsets? Or are they not power efficient enough to get into the mobile market?

 

[razel]

It all about business politics. Apple has recently been designing their own mobile chips. Samsung obviously will not be making Intel chips. So they have to try to get design wins from LG, Motorola, HTC or Huawai. Intel can also design their own, but that is not their strength. HTC needs to make a profit and Intel tends to be expensive and Hauwai ... this will be controversial is a bit too close to the Chinese government.

I think their best bet is aligning with Google for the next Nexus phones. LG isn't a bad partner either, they made the Nexus 4 and 5.

 

[senseamp]

Zenfone 2 will be Intel. I am waiting for it. Not because it's Intel, mind you. But it's all about value. It could have an ARM chip at that price point, and I'd be waiting for it too.

 

[pm]

The Lenovo P90 has an Intel CPU in it
http://www.gsmarena.com/lenovo_p90-6920.php

Also the Asus Zenfone 5
http://www.gsmarena.com/asus_zenfone_5-5952.php

And then the Zenfone 2 as well as Senseamp mentioned (odd naming strategy by Asus...)
http://www.gsmarena.com/asus_zenfone_2_prices_unveiled_in_taiwan-news-11470.php

 

[poofyhairguy]

The baseband and radios. From what I understand Intel doesn't make a full solution that would work in the American market. Any Intel phone would still need Qualcomm radios to get on say Verizon's network, so why not only approve a Qualcomm SoC if you are Verizon to make things easy? Only Apple and Samsung have the strength to stand against Verizon if such a rule is made.

That is why you see Intel more on wifi tablets. There price matters more, and Intel is giving those chips away.

 

[ericlp]

It's because ARM and Qualcomm (snapdragon) have cornered the market. There is basically no room for hardly any Profit after retooling, research, and market. Nvidia is probably getting out of the market as well. Nexus 7 Quad-Core with a 5th Battery-Saver Core from nvidia. R/D is very expensive and competition is brutal.

It's not just a battle of AMD and Intel. Intel will get their ass's handed to them time after time... When they do finally get it right (if ever) and manage to sell a million or two chips, the profits probably wouldn't even scratch the surface... AMD was even thinking of taking the fusion CPU's to the mobile market, but it's really hard to beat out an ARM CPU. QualComm can do it, because they have China behind them. I see them beating out just about everyone and giving ARM a serious run for it's money. It's great since, I think they will keep pushing the envelope.

When the Cortex-A57 8 core 2.0+ Ghz 64 Bit CPU ... I envision one day, we will all be running our machines on these energy sipping processors with built in super fast GPU's. It will be interesting times. Hopefully we'll see 20 core and beyond cpu's. Times will be interesting for both AMD and Intel, as they will be left out in the dust.

Windows better watch out as well, as go forward, most people will ditch the clunking OS's from Apple, Windows flavors in favor of a connected OS. Somehow we will all gravitate towards a cross between chrome and android.

Have you noticed MicroSoft is scrambling to give away it's products for free just to stay ahead of the curve. But, even this tactic won't slow down the fast paced drum of of change. It's coming.... Will you be on board?

 

[ericlp]

The baseband and radios. From what I understand Intel doesn't make a full solution that would work in the American market.

I was also going to state, this, but that's opening up a whole nother can of worms!

Not only have the packed in 8 cores at 2GHz on a 64 bit CPU, With a high end graphics chip, but the Inputs and Outputs of connectivity are AMAZING! Think about it! With each chip the built in radio's of not just all the Cellular network traffic of LTE...

But, Bluetooth 4.1, 8.211 AC, Near Field Communication, and GPS (IZat Gen8C)... Hard to compete with that? What's next? A VR input port? Or somehow a input into the brain itself.

I've always said America was on top of the ball, now it's going to get squashed if we don't figure out a way to get half a million more engineers in college today. But when they want to pay peanuts for an engineer or a math major... I guess we can just sit back and watch it happen. Until we wake up one day and start doing what we should have done 10-20 years ago.

 

[kpkp]

It's not just a battle of AMD and Intel. Intel will get their ass's handed to them time after time... When they do finally get it right (if ever) and manage to sell a million or two chips, the profits probably wouldn't even scratch the surface...

ASUS alone sold (if my memory isn't playing tricks on me) 8 million Zenfones last year. This year they are planing on making 1-3 million Zenfone 2 a month, depending on demand.

Just to put a bit of context to the made up numbers you are writing.

 

[corkyg]

My sense is that it is not their core business, and that market's nargins are too slim. It's more of a matter of "won"t" rather than "Can't."

 

[poofyhairguy]

My sense is that it is not their core business, and that market's amrgins are too slim. Iy's more of a matter of "won"t" rather than "Can't."

They have been pretty much giving their mobile chips away. On cost they compete with the Mediateks and Rockchips of the world.

What they "won't" do is give us Core IPC at ARM prices. Eventually the market will force them to, with Apple and Samsung moving up the IPC ladder. The days of selling mid range chips for over $100 are done.

 

[Roland00Address]

Qualcomm is far more integrated vertically than intel has been on the things that matter for SOCs and Radios. Vertical integration means more profit for Qualcomm while keeping the costs low to the end user for they can undercut all their competitors.

Until recently Intel did not have any cellular radios so when a laptop or such used a radio they just bought an off the shell part (Wifi is different though.) Intel bought infineon in 2011 but since it takes 2 to 3 years to build an SOC, and porting radio to a new process or competitor process is actually the most difficult thing to make then you will not have an intel part with an integrated modem and an integrated modems make the handset makers lives a lot cheaper, easier, and faster to market.

And Qualcomm has not just the radio (and the radio patents that matter the most), but also the soc, the isp, the dsp, etc.

Until Intel comes anywhere close to that integration than Intel will have to give their chips away or they must have a substantial higher performance like 50%> or more. Well since Intel does not want to give away their core m chips away at phone soc prices (and they are also far more complicated with power delivery and such) than intel must give atom away or not gain any market share.

----

We are supposed to see 28nm tmsc products with intel cpus and infineon 3g radio socs this year in 2015. Dual core atom 28nm tsmc with 3g radio was supposedly shipped to partners in Dec 2014. I do not know if it actually did or not. Intel actually ported the atom IP to tmsc for infineon was already developing 28nm radios in the lab so it was easier to port atom there to get as fast as possible turn around. It is likely that intels smart bracelet with built in 3g is using sofia they showed at ces but we can't be sure until we open it.

28tmsc 3g will come in a dual core and quad core offering but the quad core was never promised in q4 2014 but q1 2015. Recent evidence is it being sometime around march for partners but products will take later

28tmsc 4g lte (which is more complicated to make) is supposed to come out in the 1st half of 2015 but it is looking may

28nm rockchip version of atom with lte may actually come out faster than intel 28nm sofia atom lte chip.

Link to some intel slides from previous conferences

Now Broxton Intel 14nm version of atom which will have an lte model is supposed to come out in 2015 but Intel is hinting it may be early 2016 in the last 6 months or so. Broxton will be using goldmount cpu cores which is not a mere die shrink but a new architecture for atom completely. (Like Nehalem to Sandy or Ivy to Haswell)

 

[podspi]

The Atom refresh is long overdo. Given how well BT has held up so far (I mean that non-sarcastically) I am very excited for next-gen Atom.

 

[VirtualLarry]

The Atom refresh is long overdo. Given how well BT has held up so far (I mean that non-sarcastically) I am very excited for next-gen Atom.

In the under-$100 Win8.1 tablet market, the quad-core 22nm BT-T Atoms have done some amazing things. My HP Stream 7 is great!

 

[poofyhairguy]

In the under-$100 Win8.1 tablet market, the quad-core 22nm BT-T Atoms have done some amazing things. My HP Stream 7 is great!

I agree with that. Though I do wish Microsoft could do some Apple magic and allow their OS to move up the tablet market. X86 Windows easily has the worst DPI scaling of any major OS, as desktop mode on a retina tablet is useless because even with a jacked DPI some elements don't scale. Compare that to Android where every element in the interface scales to the DPI, or an OSX that has a no hassle retina mode that doubles the size of everything, and Windows feels primitive.

I hope they fix that in Windows 10. For now Intel needs Android to move up the ladder.

 

[podspi]

I agree with that. Though I do wish Microsoft could do some Apple magic and allow their OS to move up the tablet market. X86 Windows easily has the worst DPI scaling of any major OS, as desktop mode on a retina tablet is useless because even with a jacked DPI some elements don't scale. Compare that to Android where every element in the interface scales to the DPI, or an OSX that has a no hassle retina mode that doubles the size of everything, and Windows feels primitive.

I hope they fix that in Windows 10. For now Intel needs Android to move up the ladder.

It really is bad. I had no idea. Just purchased my first laptop with a 1080 screen, which really is not that high-res nowadays...

And I see now why Windows laptops are often stuck with low-res screens. Either everything is tiny, or blurry. Granted, programs are getting better (some are fine and look fantastic, some not so much) and all the Metro apps are perfect. Microsoft's problem is that nobody wants to create metro apps.

 

[Achtung!]

Are there any mainstream phones than have or will use intel chipsets? Or are they not power efficient enough to get into the mobile market?

Well who is Intel going to supply?

Apple? Nope.
Samsung? Nope.

Everyone else? Maybe.
Qualcomm has recently become troubled because of Snapdragon 810's heating problems. This may be a chance where Intel could step in but it needs to invest in manufacturing facilities for mobile chips. That's costly and risky business.

This is the exact reason why Apple doesn't invest in manufacturing. It's risky when you're just entering an already advanced market.
It takes decades of investments to manufacture mobile processor chips efficiently and economically.
Developing the technology is one thing, developing precision instruments and techniques to manufacture them efficiently is a whole different technological battle.

 

[grissom]

It really is bad. I had no idea. Just purchased my first laptop with a 1080 screen, which really is not that high-res nowadays...

And I see now why Windows laptops are often stuck with low-res screens. Either everything is tiny, or blurry. Granted, programs are getting better (some are fine and look fantastic, some not so much) and all the Metro apps are perfect. Microsoft's problem is that nobody wants to create metro apps.

I think the problem is that Windows ignored the dpi scaling problem for so long that there is this massive back catalogue of software with one or other of the many kludges that were introduced over the years. Few can be bothered to recode that stuff.

As you say, it's bad enough on a 1080p laptop, let alone anything higher res.