Intel Q1 2015 earnings, still losing ~1B per Quarter from Mobile

[witeken]

60+% target margins for low-end tablets? But good post.

 

[dahorns]

60+% target margins for low-end tablets? But good post.

There should have been a question mark on the end. I don't know what Intel targets for the low-end of its product stack. I assume it is somewhere between 50% and 65%.

Don't Qualcomm and the like operate around 50% for their mobile processors?

 

[DrMrLordX]

I doubt we will see Asus Atom X5 or Atom x7 (formerly Cherry Trail) tablets at $70. I will be surprised if we do not see Atom X3 (formerly SoPHIA) tablets at $70. Baytrail-T generated negative gross margin dollars. Atom X3, X5 and X7 are expected to generate positive gross margin dollars without contra revenue.

I see a potential problem there. Intel's shipments are expanding on the basis of price incentive for the consumer (and the OEM). Right now, people are getting a Z3745-equipped tablet at stupid-low prices. If they drop SoFIA into the same price bracket, consumers and/or OEMs may balk at that value proposition. They'll want Atom X5 or X7 at that price, not SoFIA/X3. When they don't get it, one of two things will need to happen:

1). Intel will have to continue with subsidies to keep prices low or
2). Overall shipments will drop off and market penetration will stall.

If Intel can actually eliminate the BoM differential, it will make good margins on the products. Further, pricing of end products shouldn't change as the total costs to Intel's customers remain the same.

See above. They're fooling themselves if they think people really want SoFIA at that price point. As has been already-stated in this thread that ARM is entrenched. I do think Intel can hurt the ARMy badly if they can get OEMs to push X5 Cherry Trails for ~$70 (or maybe even for a few dollars more).

Just to clarify, though, are we talking the SoFIA dual-core announced last July, or are we talking about the SoFIA quad?

 

[witeken]

There should have been a question mark on the end.

There's already one, isn't there?

 

[dahorns]

There's already one, isn't there?

On my post, not yours.

 

[Guest1]

I see a potential problem there. Intel's shipments are expanding on the basis of price incentive for the consumer (and the OEM). Right now, people are getting a Z3745-equipped tablet at stupid-low prices. If they drop SoFIA into the same price bracket, consumers and/or OEMs may balk at that value proposition. They'll want Atom X5 or X7 at that price, not SoFIA/X3. When they don't get it, one of two things will need to happen:

1). Intel will have to continue with subsidies to keep prices low or
2). Overall shipments will drop off and market penetration will stall.



See above. They're fooling themselves if they think people really want SoFIA at that price point. As has been already-stated in this thread that ARM is entrenched. I do think Intel can hurt the ARMy badly if they can get OEMs to push X5 Cherry Trails for ~$70 (or maybe even for a few dollars more).

Just to clarify, though, are we talking the SoFIA dual-core announced last July, or are we talking about the SoFIA quad?

There is a SoFIA Mid coming which should provide more than the X3 variant due out soon. The ARMy may be entrenched now but with what we are seeing with Samsung fauxteen nm it is only a matter of time before Intel's process lead is too far ahead. GloFo is reportedly only getting 30% on the tech sharing agreement 14 nm which means Samsung's yields can't be that much higher. This does not bode well for the future of the ARMy because everyone in that ecosystem depends on Samsung or TSM and if the best is only delivering 30% yields at 14 nm there are going to be huge supply constraints. What happens at 10 nm and below? I can see huge defections to x86 since Intel has no problems supplying any vendor that wants their product. They have an unfinished Fab 42 in AZ that could handle all of the chip volume that the ARMy could throw at them as well as all Apple volume! If Morganfield turns out to be the beast it seems to be Qualcomm is in big trouble especially after the 810 Fiasco. They really helped Intel buy some time with that one.

 

[DrMrLordX]

We'll just have to see where OEMs price the SoFIA quad tablets.

I'm curious about your source for GF only hitting 30% yields on Samsung's 14nm process. Got any links for that?

 

[Nothingness]

Wasn't there a link recently posted that explained that GF isn't using Samsung 14nm process, but their own 14nm based on more reuse of existing tools?

 

[IntelUser2000]

Are we going to see 7" Cherry Trail tablets from Asus going for such low prices? If Intel can pull that off and actually generate some positive revenue in the process, then maybe they set the bar exactly where it needs to be for their future product lines. If not, then consumers are going to see a snap-back once all the subsidized Bay Trail hardware clears out of the channels.

Two things can happen:

Intel is really telling the truth:

If the contra-revenue program is what they are saying it is and due to "platform cost differences" that they are merely paying until they can equalize the prices, then when they can equalize the prices they won't have to pay contra-revenue, platform costs stay same, and Tablets stay cheap as ever.

Intel is really lying:

If the contra-revenue program is artificial subsidiy like some people suspect than prices on the platform and therefore Tablets, will rise immediately, or Intel will lose same amount of money forever.

My opinion is that they are doing both. They aren't lying about contra-revenue itself, but they are paying subsidies for companies to use their chips, help design devices based on it. The contra-revenue part should disappear and losses will decrease somewhat. But subsidies won't.

 

[Idontcare]

There is a SoFIA Mid coming which should provide more than the X3 variant due out soon. The ARMy may be entrenched now but with what we are seeing with Samsung fauxteen nm it is only a matter of time before Intel's process lead is too far ahead. GloFo is reportedly only getting 30% on the tech sharing agreement 14 nm which means Samsung's yields can't be that much higher. This does not bode well for the future of the ARMy because everyone in that ecosystem depends on Samsung or TSM and if the best is only delivering 30% yields at 14 nm there are going to be huge supply constraints. What happens at 10 nm and below? I can see huge defections to x86 since Intel has no problems supplying any vendor that wants their product. They have an unfinished Fab 42 in AZ that could handle all of the chip volume that the ARMy could throw at them as well as all Apple volume! If Morganfield turns out to be the beast it seems to be Qualcomm is in big trouble especially after the 810 Fiasco. They really helped Intel buy some time with that one.

Hey, we can look down upon GF and Samsung till the cows come home over their 14nm yield situation, but is Intel's any better?

When was the last time you saw a BW-K available on Newegg? None there you say? Must not be back-to-school time (2014) just yet...

Personally I don't care whether Samsung's 14nm == Intel's 14nm. What I care about is whether or not I can buy product from Samsung now that is markedly superior to what I could purchase from Samsung one year ago.

The answer there is yes, yes I can. I replace my S5 with something that is a generation superior.

Can I replace my desktop Haswell yet? No? Why not, hasn't Intel super superior process lead been out for nearly a year at this point? Why aren't the desktop chips available? Oh, that's right, the yields are too craptastic, even Intel admits it.

So lets not get too high-and-mighty about where GF and Samsung stand on their 14nm efforts. At least they have something to show for their efforts. Intel is kinda missing in action right now when it comes to 14nm.

 

[IntelUser2000]

Did you know that the two most disappointing process launches by Intel were those that they hyped the most?

22nm.

And 14nm.

Everyone who marketed their process must be kicking themselves of how they promoted their 22nm and 14nm processes to be so "ahead" of the competition yet it brought the least gains per generation ever.

Really, this could be the start of real process losses for Intel. Watch them lose being first for 10nm process, and technological advantage at 7nm.

 

[witeken]

Oh, that's right, the yields are too craptastic, even Intel admits it.

Craptastic is of course relative. I'm sure Global Foundries / AMD and possibly even SS would love to have Intel's 14nm yields.

 

[videogames101]

It's not that I'd worried about the financial state of the company. Intel is still making money hand over fist, no doubt. But if I was an investor my concern would be whether or not Intel will remain king of the hill in the CPU space (across all TDPs). I'd be worried the execs are too scared to move away from the recipe that got them to the top in the first place, which may quickly be becoming obsolete. I just don't see any vision from the Intel execs. They're strategy seems to be "ride Tick-Tock into the ground". While that's not necessarily a losing strategy, it's hardly visionary or a vehicle for growth. Their mobile "strategy" is apparently nothing more than chasing the existing market, trying to retrofit their products to fit into current devices rather than developing products which are compelling enough to design devices around. To me it simply reeks of a company with nobody at the helm, just retreading what has worked in the past. BUT that retreading is still going to make gobs of cash for them and investors, so meh.

Did you know that the two most disappointing process launches by Intel were those that they hyped the most?

22nm.

And 14nm.

Everyone who marketed their process must be kicking themselves of how they promoted their 22nm and 14nm processes to be so "ahead" of the competition yet it brought the least gains per generation ever.

Really, this could be the start of real process losses for Intel. Watch them lose being first for 10nm process, and technological advantage at 7nm.

Given their R&D expenditures, this outcome seems unlikely. If it did happen, it would raise serious questions about the health of Intel's internal/technical management.

 

[Idontcare]

Wasn't there a link recently posted that explained that GF isn't using Samsung 14nm process, but their own 14nm based on more reuse of existing tools?

Uh, I don't know if there is a link to that or not, but the answer is no. That is not what GF is doing.

 

[ShintaiDK]

Did you know that the two most disappointing process launches by Intel were those that they hyped the most?

22nm.

And 14nm.

Everyone who marketed their process must be kicking themselves of how they promoted their 22nm and 14nm processes to be so "ahead" of the competition yet it brought the least gains per generation ever.

Really, this could be the start of real process losses for Intel. Watch them lose being first for 10nm process, and technological advantage at 7nm.

You gotta be kidding. Did you notice what those 2 nodes did for mobile and server? We now have 18 cores servers and 4.5W mobile Core chips.

 

[witeken]

Happy 50th anniversary, Moore's Law.

 

[Arachnotronic]

Can I replace my desktop Haswell yet? No? Why not, hasn't Intel super superior process lead been out for nearly a year at this point? Why aren't the desktop chips available? Oh, that's right, the yields are too craptastic, even Intel admits it.

So lets not get too high-and-mighty about where GF and Samsung stand on their 14nm efforts. At least they have something to show for their efforts. Intel is kinda missing in action right now when it comes to 14nm.

Low-power notebooks have by-and-large transitioned to 14nm at this point, to be fair to Intel. Core M, and both 15W and 28W Core i-series SKUs are shipping in volume quantities to major PC customers.

Intel's also shipping 14nm Pentiums/Celerons (i.e. Braswell) to OEMs in sizable quantities now.

That said, on bigger dice Intel may be facing a much tougher yield situation; BDW-H (high perf notebooks) isn't scheduled to show up until "Mid-2015" and Skylake-DT looks like an August volume launch (which implies PRQ some time before then).

However, I wonder how much of the Broadwell-H/Skylake-DT delay has to do with 14nm process rather than just Intel's channel partners wanting to draw down their own inventories of Haswell-based product in light of weak desktop (and potentially DT-replacement notebooks) sales?

 

[DrMrLordX]

Question: Is Intel moving all future Pentium and Celeron lines to Braswell and away from Broadwell/Skylake?

Two things can happen:

Intel is really telling the truth:

If the contra-revenue program is what they are saying it is and due to "platform cost differences" that they are merely paying until they can equalize the prices, then when they can equalize the prices they won't have to pay contra-revenue, platform costs stay same, and Tablets stay cheap as ever.

Intel is really lying:

If the contra-revenue program is artificial subsidiy like some people suspect than prices on the platform and therefore Tablets, will rise immediately, or Intel will lose same amount of money forever.

My opinion is that they are doing both. They aren't lying about contra-revenue itself, but they are paying subsidies for companies to use their chips, help design devices based on it. The contra-revenue part should disappear and losses will decrease somewhat. But subsidies won't.

You may be right. If Guest1 is correct and Intel starts trying to peddle SoFIA dual-core X3s in the same space that they were selling Bay Trail quads, Intel is going to get smacked around. They'll need continued subsidy if they want their mobile shipments to rise.

Of course that's going to hurt everyone else as I articulated above, but hey, that's just life. Me and my irrelevant desktop feel no sympathy for the mobile majority.

 

[jpiniero]

Question: Is Intel moving all future Pentium and Celeron lines to Braswell and away from Broadwell/Skylake?

Unknown, but they are selling 15 W Pentium and Celeron Broadwell U parts.

 

[ShintaiDK]

Question: Is Intel moving all future Pentium and Celeron lines to Braswell and away from Broadwell/Skylake?

Very unlikely. Just as previous.

Braswell is no where near Pentium and Celeron Core performance. While you can get away selling some special Pentium/Celeron options with Atom chips. You cant stop selling Core based Celeron and Pentium chips. The market is simply too massive and the performance delta too huge.

 

[Arachnotronic]

Very unlikely. Just as previous.

Braswell is no where near Pentium and Celeron Core performance. While you can get away selling some special Pentium/Celeron options with Atom chips. You cant stop selling Core based Celeron and Pentium chips. The market is simply too massive and the performance delta too huge.

In traditional desktops (i.e. towers), I don't think we'll see a particularly heavy mix of Atom-class CPUs, but in notebooks the majority (>= 83%) of Pentium/Celeron are already Atom-based.

In AIOs, I can see Atom-class Pentiums/Celerons making up a big chunk of the business.

 

[Guest1]

We'll just have to see where OEMs price the SoFIA quad tablets.

I'm curious about your source for GF only hitting 30% yields on Samsung's 14nm process. Got any links for that?

http://appleinsider.com/articles/15...tsmc-for-30-of-a9-chip-orders-for-next-iphone

 

[Guest1]

Hey, we can look down upon GF and Samsung till the cows come home over their 14nm yield situation, but is Intel's any better?

When was the last time you saw a BW-K available on Newegg? None there you say? Must not be back-to-school time (2014) just yet...

Personally I don't care whether Samsung's 14nm == Intel's 14nm. What I care about is whether or not I can buy product from Samsung now that is markedly superior to what I could purchase from Samsung one year ago.

The answer there is yes, yes I can. I replace my S5 with something that is a generation superior.

Can I replace my desktop Haswell yet? No? Why not, hasn't Intel super superior process lead been out for nearly a year at this point? Why aren't the desktop chips available? Oh, that's right, the yields are too craptastic, even Intel admits it.

So lets not get too high-and-mighty about where GF and Samsung stand on their 14nm efforts. At least they have something to show for their efforts. Intel is kinda missing in action right now when it comes to 14nm.

It's not a matter of high and mighty it's a matter of seeing the problems that are coming down the road. Their 14 nm isn't even as good as Intels and it is already having yield issues. I understand Apple is pushing them to be on the bleeding edge but if they are desperate enough to ship with such terrible yields than what does that say? Does anyone really think Apple went t GloFo first for their 14nm or is that because of the tech sharing with Samsung?

 

[Guest1]

Question: Is Intel moving all future Pentium and Celeron lines to Braswell and away from Broadwell/Skylake?



You may be right. If Guest1 is correct and Intel starts trying to peddle SoFIA dual-core X3s in the same space that they were selling Bay Trail quads, Intel is going to get smacked around. They'll need continued subsidy if they want their mobile shipments to rise.

Of course that's going to hurt everyone else as I articulated above, but hey, that's just life. Me and my irrelevant desktop feel no sympathy for the mobile majority.

There's SoFIA LTE2 and SoFIA MID both going to be 14 nm not 28nm X3. That would be noncompetitive. I am curious about what SoFIA mid is going to be packing. Morganfield will be on Goldmont I would expect SoFIA mid to be between Goldmont and Airmont. So maybe a down clocked Goldmont or an upclocked Airmont? Cherrytrail is already launched with Surface 3. Morganfield is supposed to be after that if I remember the roadmap properly. SoFIA is just a phone AP not for tablets of course that wouldn't stop an OEM from squeezing it in one.

 

[Arachnotronic]

There's SoFIA LTE2 and SoFIA MID both going to be 14 nm not 28nm X3. That would be noncompetitive. I am curious about what SoFIA mid is going to be packing. Morganfield will be on Goldmont I would expect SoFIA mid to be between Goldmont and Airmont. So maybe a down clocked Goldmont or an upclocked Airmont?

I don't think an upclocked Airmont would be able to hold up; Cherry Trail is likely to be outperformed by the Cortex A72/A53 big.LITTLE solutions that the likes of Qualcomm and MediaTek will be shipping later this year.

If SoFIA MID is a 2016 part, and if Intel wants it to be competitive with its ARM contemporaries, it really needs to pack Goldmont CPU cores.

Given that the original SoFIA wasn't on the roadmap ~15 months ago, I'd assume SoFIA MID is a new addition to the family that was defined fairly recently. Intel's product planners should know what they're up against, and if they decide to put in a ~2.7GHz Airmont, SoFIA MID will not hold up well against TSMC 16FF+ A72s.