Intel Q2 2014 financial result.

[Arachnotronic]

Intel17, what are you looking forward to as a milestone in Intel's mobile adventure? Is it sill Broxton in mid-2015?

It's actually much simpler than any particular product. From a milestone perspective, I would be looking for Intel to start taking tablet/smartphone share while generating net revenue on a per product basis.

I'm fairly confident that the tablet situation gets a lot more attractive for Intel next year as the contra-revenues fade away (and I assume some volume growth on top of this year's 40 million). However, if you notice what is really dragging down Intel's mobile financials, it's the evaporation of the 2G/3G modem business.

The big open question is the smartphone market. SoFIA could be a very compelling product for the mass-market smartphones, and if Intel can secure mass-market design wins, this should go a long way towards a recovery from the 2G/3G drop.

The really simple answer is: design wins in fairly high volume smartphones without the need for a contra-revenue bill of materials offset. Whether these designs are SoFIA or Broxton isn't that important as right now Intel's goal should be to build share and customer relationships rather than necessarily mountains of profit (those come in about 5-10 years for the remaining vendors, and I expect that it will be Qualcomm + somebody else).

 

[frozentundra123456]

@ Homeless: Exactly. In any case, I don't think AMD is going out of business, but whether they ever become more than a niche competitor is problematic at best. They face the same competition as Intel in x86, and just because they are trying to get into ARM doesn't mean they can compete successfully with the established competitors who have hugely greater resources to draw on. I certainly would not count on some magical chip destined for 2016 at the earliest until a lot more info is available.

 

[tential]

It's actually much simpler than any particular product. From a milestone perspective, I would be looking for Intel to start taking tablet/smartphone share while generating net revenue on a per product basis.

I'm fairly confident that the tablet situation gets a lot more attractive for Intel next year as the contra-revenues fade away (and I assume some volume growth on top of this year's 40 million). However, if you notice what is really dragging down Intel's mobile financials, it's the evaporation of the 2G/3G modem business.

The big open question is the smartphone market. SoFIA could be a very compelling product for the mass-market smartphones, and if Intel can secure mass-market design wins, this should go a long way towards a recovery from the 2G/3G drop.

The really simple answer is: design wins in fairly high volume smartphones without the need for a contra-revenue bill of materials offset. Whether these designs are SoFIA or Broxton isn't that important as right now Intel's goal should be to build share and customer relationships rather than necessarily mountains of profit (those come in about 5-10 years for the remaining vendors, and I expect that it will be Qualcomm + somebody else).

+1000

 

[mrmt]

I'm fairly confident that the tablet situation gets a lot more attractive for Intel next year as the contra-revenues fade away (and I assume some volume growth on top of this year's 40 million). However, if you notice what is really dragging down Intel's mobile financials, it's the evaporation of the 2G/3G modem business.

I would also add Intel LTE modem to this mix. If they can dent Qualcomm dominance here, there's a fairly good chance that they can reach 2016 making very good money.

 

[witeken]

The really simple answer is: design wins in fairly high volume smartphones without the need for a contra-revenue bill of materials offset.

Then I guess Intel just reached their first big mobile milestone!

Joking aside, I'm waiting to see how SoFIA does. Broxton seems like it will only appeal to the high-end because it lacks an integrated modem (although I think its performance and power will be great), and I'm not sure if a 28nm SoFIA will attract the midrange, competing against 20nm SoCs. I think Intel will only have its full smartphone platform in 2016 with 14nm SoFIA. Pretty disappointing. I hope they'll get an important design win like Nexus, that could be a nice catalyst.

 

[Arachnotronic]

Then I guess Intel just reached their first big mobile milestone!

Joking aside, I'm waiting to see how SoFIA does. Broxton seems like it will only appeal to the high-end because it lacks an integrated modem (although I think its performance and power will be great), and I'm not sure if a 28nm SoFIA will attract the midrange, competing against 20nm SoCs. I think Intel will only have its full smartphone platform in 2016 with 14nm SoFIA. Pretty disappointing. I hope they'll get an important design win like Nexus, that could be a nice catalyst.

I agree that Broxton will probably have some pretty strong compute/GPU performance. The jury is out on imaging performance, but at the London Analyst Summit, Josh Walden acknowledged Intel's deficiencies there and indicated that Broxton would be a significant improvement. I am optimistic for Broxton from a compute/performance perspective.

I also agree that it will have limited appeal in the mass-market smartphone space, though I think paired with a good modem it could do well in the mid-range and up. The question is whether it will be more of a S808/S810 or FinFET Qualcomm chip competitor. I'm leaning towards the latter, and in that case, it may not be so cut and dry, but if Intel can get it out well ahead of a FinFET Snapdragon then it could be much more interesting.

But, the part I really care about is 14nm SoFIA. The cost structure on that will probably be fantastic and none of the fabless companies will likely be able to afford to sell a cheap 16nm/14nm chip during that timeframe. Intel will be able to because Broadwell/Skylake/Cherry Trail/etc. will have eaten the brunt of the depreciation costs on those factories, leaving 14nm SoFIA to very profitably fight whatever 28/20nm parts are in the market at that time.

 

[Abwx]

Abwx,

snip...

Thanks for the thorough analysis i17, agree on the whole.

 

[pw257008]

These companies aren't really comparable to Intel. They all sell end product consumer goods. Hell, LG and Samsung are large diversified conglomerates that are in everything: ship buildings, construction, telecommunications, insurance (the list goes on). They are in many respects outgrowths of the South Korean government--heavily subsidized since the end of the Korean war.

just wanted to bump this post to appreciate how right it is on how poor a comparison these companies make to Intel (and its relevance to many threads here, such as those comparing Samsung to Apple).

 

[VirtualLarry]

AMD isn't giving up these markets, consumers have been giving up AMD in droves, that's what is making AMD irrelevant: Nobody wants what they sell.

Brazos is crap, we know that. Too many Wal-Mart special Brazos laptops, and yes, desktops, tarnished AMD as a brand.

AMD used to be known as "competitive, but budget", back in the days of the Athlon XP. (Superior price / performance to PIII and early PIV.)

Now they're just known as "the slow CPU brand".

More crapcat cores aren't going to save them. They need something with real performance.

 

[jpiniero]

More like they need a credible smartphone/tablet chip 2 years ago and/or some corporate share since consumers aren't buying new PCs.

 

[Arachnotronic]

Thanks for the thorough analysis i17, agree on the whole.

You are welcome!

 

[ShintaiDK]

Now they're just known as "the slow CPU brand".

More crapcat cores aren't going to save them. They need something with real performance.

Real performance cost R&D. And that keep declining.

 

[mrmt]

More crapcat cores aren't going to save them. They need something with real performance.

Something with real performance costs real money, and something with real performance doesn't fit their embedded business strategy. Expect future AMD processors to look like a lot more like the cat family than derp family.

 

[VirtualLarry]

Their cat cores wouldn't be so bad, if they could clock like Intel chips. Give me a 4.2Ghz Athlon 5350-equivalent, and I think that I could live with that. Especially if it was 65W or less.

Edit: Sorry for straying a little bit off-topic.

I'm glad that Intel is doing well, well, everywhere except mobile, I guess.