2015 Top 20 Semiconductor Forecast

[ShintaiDK]

These numbers are essentially how the final result will end. Lots of cannibalizing, overall 0%(close to -1%) growth for the year. Qualcomms deep problems gets exposed.

There is a slight note to remember, US based companies have it harder due to the strong dollar throughout the year.
Yen changed -13% vs the dollar. Euro -16%, Won -7% and TWD -4%.

Also Avago and Broadcom is now the same company. Same with NXP buying Freescale.

2016 will most likely give more mergers and buyouts (Micron already bought out one).

 

[videogames101]

With all this Intel vs ARM stuff, it's funny to see that ARM's revenue doesn't even break the top 20.

And also that despite their relatively small size, they're shaping the roadmap for many of the companies inside the top 20.

What does the future look like?

 

[ShintaiDK]

ARM doesn't manufactor chips. Its a design company.

The future seems to be stagnant. But with more consolidation as we currently see. Mergers and buyouts is the road to getting big enough to survive. Micron for example bought out another memory maker Inotera.
http://investors.micron.com/releasedetail.cfm?ReleaseID=946777

 

[Nothingness]

With all this Intel vs ARM stuff, it's funny to see that ARM's revenue doesn't even break the top 20.

As Shintai wrote, ARM doesn't sell any chip. It's like comparing the revenues of a car designer versus Ford or GM.

IIRC ARM revenues are slightly above $1B.

And also that despite their relatively small size, they're shaping the roadmap for many of the companies inside the top 20.

Indeed, and most of these top 20 companies are ARM licensees, including Intel.

 

[Phynaz]

Where's AMD in that list?

 

[dark zero]

AMD is deader than dead... unless is Sony who gets the sales

 

[Arachnotronic]

Where's AMD in that list?

AMD is on track to do a bit under $4 billion in sales this year. Since NXPI + Freescale and Broadcom + Avago are merging next year, AMD has a chance to get back on the list due to fewer # of competitors.

 

[ShintaiDK]

AMD is on track to do a bit under $4 billion in sales this year. Since NXPI + Freescale and Broadcom + Avago are merging next year, AMD has a chance to get back on the list due to fewer # of competitors.

Unless their sales continues to drop at the same rate and end at 3B

 

[Systems analyst]

The future may be all SoCs in retail products, using ARM IP. INTEL said that they are no longer a PC company.
ARM are involved in production, via their Physical IP.
In 2020 most laptops may be ARM powered.
ARM are supporting TSMC, UMC and Global Foundries with physical IP. This yields diversity.

 

[ShintaiDK]

The future may be all SoCs in retail products, using ARM IP. INTEL said that they are no longer a PC company.
ARM are involved in production, via their Physical IP.
In 2020 most laptops may be ARM powered.
ARM are supporting TSMC, UMC and Global Foundries with physical IP. This yields diversity.

Are you the brother of Nosta? LOL!

 

[Arachnotronic]

The future may be all SoCs in retail products, using ARM IP. INTEL said that they are no longer a PC company.
ARM are involved in production, via their Physical IP.
In 2020 most laptops may be ARM powered.
ARM are supporting TSMC, UMC and Global Foundries with physical IP. This yields diversity.

Lol, what are you smoking, bro? In 2020 I fully expect most PCs to be Intel powered. ARM in PCs is a no go.

To be fair by 2020 Intel will probably have single digit share of the smartphone market at best. 2% would be a real achievement for these guys; they (Intel mobile) make AMD look like a semiconductor rockstar.

 

[ShintaiDK]

Microsoft would have to be pretty much gone as well for that fantasy dream to happen. Including all legacy needs. Maybe in another universe, but not this one.

 

[ShintaiDK]

To be fair by 2020 Intel will probably have single digit share of the smartphone market at best. 2% would be a real achievement for these guys; they (Intel mobile) make AMD look like a semiconductor rockstar.

Yep. The big question tho may be what the smartphone shipments and revenue will be. We already saw tablets hit the peak and go backwards again. We already see declines in established markets. Gartner for example only estimate a 5% shipment increase from 2015 to 2017. PC shipment including tablet is expected to go up 6% in the same timeframe.

 

[jpiniero]

Yep. The big question tho may be what the smartphone shipments and revenue will be. We already saw tablets hit the peak and go backwards again.

Tablets are sliding but PCs are sliding faster.... and Smartphones have the look of being at a top. It's not going to be easy for anyone, especially since everyone is also having fab problems.

 

[ShintaiDK]

Tablets are sliding but PCs are sliding faster.... and Smartphones have the look of being at a top. It's not going to be easy for anyone, especially since everyone is also having fab problems.

That's how I see it as well. The free lunch is over. This is also why we see so aggressive buyouts and mergers. You buy volume/customers to consolidate your position.

 

[Systems analyst]

Lol, what are you smoking, bro? In 2020 I fully expect most PCs to be Intel powered. ARM in PCs is a no go.

To be fair by 2020 Intel will probably have single digit share of the smartphone market at best. 2% would be a real achievement for these guys; they (Intel mobile) make AMD look like a semiconductor rockstar.

Of course, Arachno, that is why you have shares in Intel.

The main arguments against your position, in my view, are that:

1. ARM are producing new mobile core designs every year; this is because they and their partners foresee the need to issue new mobile products every year; this includes smart-phones, tablets and laptops/AIOs. This is a relentless pace of development.

2. The price of ARM based SoCs will be an order of magnitude less than the price of Intel equivalents, based on X86; Intel may well be able to compete at the top end, but there they face competition from Apple. I think that Apple will win that fight.

 

[Systems analyst]

Microsoft would have to be pretty much gone as well for that fantasy dream to happen. Including all legacy needs. Maybe in another universe, but not this one.

Microsoft will support whoever wins the laptop/AIO market, in my view.
Presumably some Linux derivative will be an alternative.

 

[Systems analyst]

Tablets are sliding but PCs are sliding faster.... and Smartphones have the look of being at a top. It's not going to be easy for anyone, especially since everyone is also having fab problems.

The total mobile market includes smart-phones, tablets and laptops/AIOs; this market is projected to rise in SoC value over the next five years.
Laptops/AIOs in particular will be the battle ground, I believe.
TSMC do not seem to have fab problems.

 

[jpiniero]

TSMC do not seem to have fab problems.

TSMC (and Samsung really) have big cost problems starting with their fake "16nm"/"14nm" nodes, and yield was bad enough that Apple had to source from both of them. They of course tried to include GloFo, but GloFo failed miserably at doing Samsung's "14nm" node. Yield has improved at TSMC that Apple was able to do the A10 exclusively there but that's basically a 3 year old node at that point and the cost problems aren't going away. Don't be surprised if the A11 is fabbed at Intel's 14 nm.

 

[Systems analyst]

TSMC (and Samsung really) have big cost problems starting with their fake "16nm"/"14nm" nodes, and yield was bad enough that Apple had to source from both of them. They of course tried to include GloFo, but GloFo failed miserably at doing Samsung's "14nm" node. Yield has improved at TSMC that Apple was able to do the A10 exclusively there but that's basically a 3 year old node at that point and the cost problems aren't going away. Don't be surprised if the A11 is fabbed at Intel's 14 nm.

ARM completed their work on 16nm with TSMC, in September last year, IIRC. They announced follow on work on 10nm with TSMC, at that time. They have since implied that 10nm ARM physical IP is available from ARM/TSMC, since a licence has been sold for this.

You may argue that TSMC and the others are behind Intel; we have been told for years that Intel would win in mobile, because of a superior production process; there is just no evidence of the truth of this; in fact, Intel are nowhere in production of mobile SoCs. In my book, that puts them behind TSMC and Samsung.

 

[Arachnotronic]

Of course, Arachno, that is why you have shares in Intel.

I have no problems owning ARMH when it gets cheap Actually, it's starting to get pretty attractive here...

1. ARM are producing new mobile core designs every year; this is because they and their partners foresee the need to issue new mobile products every year; this includes smart-phones, tablets and laptops/AIOs. This is a relentless pace of development.

Um, Intel does the same...

2011 - Sandy Bridge
2012 - Ivy Bridge
2013 - Haswell
2014 - Broadwell
2015 - Skylake
2016 - Kaby Lake + Knights Landing Core
2017 - Skylake Xeon + Cannonlake
2018 - Icelake + Cannonlake Xeon (?)

2. The price of ARM based SoCs will be an order of magnitude less than the price of Intel equivalents, based on X86; Intel may well be able to compete at the top end, but there they face competition from Apple. I think that Apple will win that fight.

Ah yes, Apple is TOTALLY getting into the game of supplying A-series SoCs to other PC vendors. Oh wait, Apple hasn't even put A-series SoCs into its own MacBooks like the Intel doomsayers have been predicting...

In the PC market, a little thing called legacy code is important and creates a very high barrier to entry for other players. That, and the fact that Intel's processors here are world class.

 

[arandomguy]

Isn't the consumer PC market changing? What x86 legacy support does the mainstream consumer really need these days? I'm not saying that ARM will start to massively displace x86 in short order but I don't see the idea of a large segment of consumers not needing a personal x86 PC being far fetched either.

Unless Samsung or Apple get displaced from their positions in the mobile (smart phone/tablet) market I don't see what Intel's long term prospects would be there either. They want to be high margin and in high margin devices but the only two players in that segment with meaningful volume and profits both have their own chips.

 

[AtenRa]

yea, at 14/16nm a quad core 2.5GHz ARM A72 will be more than enough for the vast majority of users.

When you can have a Quad Core ARM A72 Smart Phone with Windows 10 Mobile that can be used as a desktop when connected to a monitor or TV, then entry x86 CPUs are starting to become irrelevant for the vast majority of users.

From H2 2016 and 2017, 14/16nm FF ARM A72 quad core or even 8-Core cheap CPUs will be widely available in sub $300 Smart Phones. Unless x86 will bring something new, im expecting the majority of users to switch to a smartphone only device from 2017 onward to replace entry level PC, Laptop and perhaps even their tablets with a single device.

 

[CakeMonster]

When will the SoC market stagnate? When will there not be a case for a new SoC/phone every year? When will the market catch up to the PC situation?

 

[ShintaiDK]

When will the SoC market stagnate? When will there not be a case for a new SoC/phone every year? When will the market catch up to the PC situation?

It may hit next year.