Intel Opens Doors To Rivals, Maybe

← Back to Stories (view on slashdot.org)

Intel Opens Doors To Rivals, Maybe

Posted by Unknown on Saturday November 23, 2013 @07:15AM from the one-foundry-to-rule-them-all dept.

Rambo Tribble writes "In what appears to be a major reversal of policy, Intel's new president, Renee James, has indicated that Intel will be open to manufacturing chips based on rivals' designs. While the language is a bit tentative, this appears to open an opportunity for such as ARM to benefit from Intel's manufacturing expertise and technology." From the article: "James said Intel will evaluate prospective foundry clients on a 'deal by deal basis, not on an architecture by architecture basis.' That applies, James said, 'even in areas where there may be some competition with businesses that we’re in.'" Intel is already manufacturing FPGAs for Altera that include 64-bit ARM cores.

40 of 59 comments (clear)

Min score:

Reason:

Sort:

Probably Apple by valerio · 2013-11-23 07:21 · Score: 1

Will be among their first clients...
1. Re:Probably Apple by jcr · 2013-11-23 07:30 · Score: 4, Interesting
  
  I would hope so. Friends in the chip design world have told me that Apple could save about 20% of the power draw on their A-series CPUs if they had access to intel's fab process.
  -jcr
  
  --
  The only title of honor that a tyrant can grant is "Enemy of the State."
2. Re:Probably Apple by Rockoon · 2013-11-23 07:50 · Score: 4, Interesting
  
  There is no chance that Intel would rent out its best fabs.
  
  The reality is that Intel has a growing number of out-dated fabs that are not utilizing full capacity, and it wants to sell time on those.
  
  --
  "His name was James Damore."
3. Re:Probably Apple by Anonymous Coward · 2013-11-23 08:18 · Score: 1
  
  Apple could buy Intel, at least in theory. They have cash reserves of $147 Billion - Intel's market cap is only $118 Billion.
4. Re:Probably Apple by binarylarry · 2013-11-23 08:32 · Score: 1
  
  Computing would enter the dark ages.
  
  --
  Mod me down, my New Earth Global Warmingist friends!
5. Re:Probably Apple by Kjella · 2013-11-23 08:34 · Score: 1
  
  Apple could buy Intel, at least in theory. They have cash reserves of $147 Billion - Intel's market cap is only $118 Billion.
  They could, but it'd be pointless for the same reason nobody wants to gobble up AMD for 2-3 billion. All those cross licensing deals would become invalid (like for example x86-64 that AMD invented) and they'd have a nightmare trying to relicense it.
  
  --
  Live today, because you never know what tomorrow brings
6. Re:Probably Apple by ebno-10db · 2013-11-23 08:56 · Score: 2
  
  Apple could buy Intel, at least in theory. They have cash reserves of $147 Billion - Intel's market cap is only $118 Billion.
  That's a sad commentary on the state of the "tech" world. Designing cell phones and the like is not trivial, but it's much less sophisticated than designing and building the components that almost all cell phone manufacturers buy. I believe Apple designs some of their digital chips, but are they redesigning every detail of the ARM for higher performance like Qualcomm does w/ Snapdragon? By far and away the most sophisticated and difficult to design chips in a cell phone (smart or dumb) are the mixed-signal/RF chips that handle the actual wireless implementation, from RF down to baseband processing. That usually also includes any necessary DSP and wireless protocol software, which is not simple. After that the wireless stuff is mostly just a black box that your "main" CPU talks to via a simple interface. Does Apple do any of that work? I think they just buy from Qualcomm or Infineon like everybody else. Have they even switched to OLED's yet? That's a sophisticated tech - that Apple will buy essentially off-the-shelf. In other words, an iPhone is basically just a systems integration job, with the sophisticated core tech bought from elsewhere.
  By contrast Intel is still one of the leading CPU design companies in the world, and almost always has the most bleeding edge fabs. What's that worth?
  A joke you'll get only if you've seen the Curious George movie: the real money is in parking garages.
7. Re:Probably Apple by kjc197 · 2013-11-23 09:21 · Score: 2
  
  There is no chance that Intel would rent out its best fabs.
  The reality is that Intel has a growing number of out-dated fabs that are not utilizing full capacity, and it wants to sell time on those.
  Not really... The Altera announcment was on 14nm Tri-Gate, which is not exactly out-dated, more cutting edge.
8. Re:Probably Apple by Luckyo · 2013-11-23 10:24 · Score: 1
  
  Intel likely has a poison pill and its stock price would explode the moment anyone big would try for such a move.
  Not to even begin talking about trying to clear such a deal with monopoly watchdogs. Good luck with that.
9. Re:Probably Apple by shentino · 2013-11-23 10:44 · Score: 1
  
  would the DOJ allow it on antitrust grounds?
10. Re:Probably Apple by whistlingtony · 2013-11-23 11:03 · Score: 4, Interesting
  
  Intel has a "copy exactly" policy. Every fab is Intel's "best fab", mainly because maintaining different processes across sites would be a logistical nightmare. Also, I'm sitting in Intel's "best fab", and we run this stuff. So. Nope.
11. Re:Probably Apple by jcr · 2013-11-23 11:28 · Score: 1
  
  There is no chance that Intel would rent out its best fabs.
  What's your next guess?
  Apple is a very important customer to Intel.
  -jcr
  
  --
  The only title of honor that a tyrant can grant is "Enemy of the State."
12. Re:Probably Apple by willy_me · 2013-11-23 11:31 · Score: 2
  
  but are they redesigning every detail of the ARM for higher performance like Qualcomm does w/ Snapdragon?
  yes. Apple has purchased several companies that specialize in ASIC design and the latest A6 CPU is the fruit of their labor. It is very different from other ARM processors on the market so this should not be much of a surprise.
  
  an iPhone is basically just a systems integration job, with the sophisticated core tech bought from elsewhere
  To a certain point, yes. But using that same logic, all other cell phone manufacturers are even worse. And there is nothing wrong with contracting out design when a company requires a unique part. There are times you want to do it in-house and times when contracting out is the preferred solution. Apple performs more in-house design then any other cell phone manufacturer. Only Samsung comes close - but their displays are developed a separate division and sold to everyone so it really doesn't count.
  My point is that you are greatly under-appreciating the difficulty and importance of integrating / specifying the various different system components. Combining the different hardware and software to produce an efficient final product is typically the hardest part - assuming it's done right. And this is precisely why the market for premium cell phones exists. If it was easy, everyone would do it.
  
  By contrast Intel is still one of the leading CPU design companies in the world, and almost always has the most bleeding edge fabs. What's that worth?
  Based on sales - less then iPhones. Don't forget the Intel is not the only designer and manufacturer of CPUs. IBM has their Power7 series which outperform Intel - if you're willing to pay for it. Numerous low cost, highly efficient, low power CPUs exist from other companies. So while Intel is impressive, their value is still limited by the market in which they operate - just like every other company. It's why they're trying so hard to break into the mobile market.
13. Re:Probably Apple by Guy+Harris · 2013-11-23 11:57 · Score: 2
  
  No, the white ages!
  No, the aluminum ages.
14. Re:Probably Apple by Guy+Harris · 2013-11-23 12:01 · Score: 1
  
  but are they redesigning every detail of the ARM for higher performance like Qualcomm does w/ Snapdragon?
  yes. Apple has purchased several companies that specialize in ASIC design and the latest A6 CPU is the fruit of their labor. It is very different from other ARM processors on the market so this should not be much of a surprise.
  The latest CPU is the A7, which is also different from other ARM processors on the market (for one thing, by implementing the ARMv8-A 64-bit architecture; I think it's the only currently-shipping 64-bit ARM processor).
15. Re:Probably Apple by Paradise+Pete · 2013-11-23 12:44 · Score: 1
  
  When taking over big companies, a 25%-50% premium is standard.
  Motorola had a market cap of about $6B when Google bought it for 12.5B. Admittedly, Google got suckered big time by Icahn and the boys.
16. Re:Probably Apple by ebno-10db · 2013-11-23 13:24 · Score: 1
  
  using that same logic, all other cell phone manufacturers are even worse
  I mentioned Apple in particular only because somebody pointed out their cash hoard is bigger than Intel's market cap.
  
  there is nothing wrong with contracting out design when a company requires a unique part
  Of course not - where did so much as suggest there was? Though in many cases they're looking for a customization of otherwise OTS stuff, which is stretching the implication of the word "unique".
  
  Apple performs more in-house design then any other cell phone manufacturer.
  
  My point was that that's not saying much.
  
  Only Samsung comes close - but their displays are developed a separate division and sold to everyone so it really doesn't count.
  How does that not count? It's Samsung. They design and build OLED displays. Actually having serious tech means they make more money selling the stuff on the open market rather than keeping it as an "exclusive" just for their cell phones.
  
  Combining the different hardware and software to produce an efficient final product is typically the hardest part - assuming it's done right.
  No it's not. I speak from experience, as I've worked both sides of the street.
  
  And this is precisely why the market for premium cell phones exists. If it was easy, everyone would do it.
  Lots of companies do it. The cell/smart phone business is like the fashion business - up one year with the hot product and going down the toilet the next. Remember when Nokia was king of the hill? How about Blackberry? The list goes on.
  Maybe that's why Qualcomm got out of the cell phone business and now does the RF/modem chipsets and software. It has a much greater barrier to entry because it's more difficult, requires greater expertise, and takes years of refinement of a product line to get it to the point where it's truly competitive. AFAIK their only real competition is Infineon, and they had a very hard time getting there. Once established in that business though, it's more stable because of the barriers to entry, and the fact that all cell phone manufacturers need you product.
  
  Based on sales - less then iPhones.
  Of course - that was my point. I wonder what the parking garage business is worth. You know how much the garages in Manhattan alone rake in?
  
  Don't forget the Intel is not the only designer and manufacturer of CPUs.
  No, but for many years they've consistently had the most bleeding edge digital fab processes. That's a wee bit more difficult than insisting that rectangles should have rounded corners, though apparently not as profitable.
17. Re:Probably Apple by ebno-10db · 2013-11-23 13:26 · Score: 1
  
  Good to know there's some actual tech being done at Apple.
18. Re:Probably Apple by MysteriousPreacher · 2013-11-23 13:32 · Score: 4, Funny
  
  Apple could buy Intel, at least in theory. They have cash reserves of $147 Billion - Intel's market cap is only $118 Billion.
  That's exactly how it works.
  Scene 1: The stockbroker's office
  Tim Cook enters the office of his stockbroker, carrying a briefcase and looking determined.
  Cook: Good morrow, old chap.
  Stockbroker: Ah, Mr. Cook. Do please take a seat and explain how I may be of service?
  Cook: I'd like to buy Intel shares.
  Stockbroker: Excellent choice, sir. What quantity did you have in mind?
  Cook: All of them
  Stockbroker: Right away, Mr. Cook. Cash or credit?
  Cook opens a suitcase to reveal $118 billion in billion dollar notes.
  Cook: Cash, of course. Jeeves, ready my stealth jet.
  Jeeves: Very good, sir. Where to?
  Cook: Santa Clara.
  Jeeves: Yes, sir.
  Scene 2: The Intel boardroom
  Tim Cook strides in to the boardroom, smoking a cigar and eyeing the decor.
  Krzanich: Cook, you can't come barging in here!
  Cook: I think not, old friend. For I just bought Intel!
  Krzanich: How?
  Cook: I read an informative post on Slashdot in which it was explained how it was as easy as having money equal to the market cap. I'll need to ask you to leave.
  Krzanich: Damn it! I didn't think anyone knew this!
  Jeeves: Shall I begin construction of your Tony Stark-like secret labs under the building, where you may commence work on powered armour or something similar, sir?
  Cook: Yes, Jeeves. That would be awesome - Awesome to the max!
  
  --
  -- Using the preview button since 2005
19. Re:Probably Apple by Anonymous Coward · 2013-11-23 13:59 · Score: 1
  
  No they didn't, Motorola's tax writeoffs alone were worth $6.6B to Google. They also sold their STB business to Arris for $2.4B, and paid for the entire deal with offshore money that would have to be repatriated at 35%.
  That's a purchase price of $12.5B*.65 = $8.1B, minus $700m per year till 2019 in capital write-offs, minus an immediate write-off of $1.7B on Motorola's previous losses. In the end they got the business for minus -$900M and free patents out of the bargain.
20. Re:Probably Apple by ebno-10db · 2013-11-23 14:21 · Score: 1
  
  Wow - quite a bargain then. Seriously.
21. Re:Probably Apple by tlhIngan · 2013-11-23 19:25 · Score: 1
  
  The latest CPU is the A7, which is also different from other ARM processors on the market (for one thing, by implementing the ARMv8-A 64-bit architecture; I think it's the only currently-shipping 64-bit ARM processor).
  It is, actually - every other 64-bit solution aims to ship in 2014, and Android I think is only scheduled to get 64-bit support in late 2014 as well.
  And you want to know why the A7 is so fast? The ARMv8 architecture changes up how ARMs work so they're much more efficient - and faster. The A7 only runs 32 bit code a little faster than the A6, but when you run native 64-bit code, it speeds up seriously because of the way ARMv8 works.
  If you're an Android fan, I would suggest buying nothing until Android gets 64-bit native, because it's going to get the advantages we're seeing with the A7 - yes, right now the A7 is keeping up with SOCs with twice the cores (A7 - dual, Snapdragon - quad) and over 50% higher clock speeds (1.4GHz vs. 2.2GHz). Imagine how fast Android will be with those advantages - it'll be 4-6 times faster than what we have today. Doubling the cores and bumping the speed up 50%, that is.
  And I'm sure Intel's worried - the latest Bay Trail Atoms are basically even with the A7 in performance. Sure it's WIndows vs. iOS, but on the benchmarks, if ARM is catching up to Atom, then Intel's Atom advantage is shrinking fast. Intel has positioned the mobile Atoms as way better performers than ARMs, after all.
  Once the quad core 2.2GHz 64-bit processors come out, Intel could very well be left in the dust.
22. Re:Probably Apple by RightSaidFred99 · 2013-11-23 21:48 · Score: 1
  
  This is simply not true. Intel has copy-exactly but that does not mean every fab manufacturers products on every process. It means P1272 is the same everywhere, not that every fab has P1274 or whatever the latest is.
23. Re:Probably Apple by RightSaidFred99 · 2013-11-23 21:49 · Score: 1
  
  Rubbish. The margins on manufacturing Apple's chips would be less than manufacturing their own. Apple would be lucky to get 22nm.
24. Re:Probably Apple by michael_cain · 2013-11-24 04:12 · Score: 1
  
  There were a whole series of stories in the EE Times about 18 months ago when Intel started selling capacity on their then leading-edge fab line. The reality at the time was that Intel couldn't sell enough parts to keep the line full, and were going to eventually have to take big write-downs unless they found a way for the line to generate more revenue. That continues to be true. The really interesting event over the last 18 months has been the announcements by a number of Far East foundry companies that they can't afford to build fab lines that go below 20-nm. There seems to be a growing body of evidence that Rock's Law is starting to bite, and that as components shrink past 14-nm, there will be very, very few places in the world where you can get such parts made.
25. Re:Probably Apple by Ottibus · 2013-11-24 21:51 · Score: 1
  
  Android I think is only scheduled to get 64-bit support in late 2014 as well.
  I suspect that Android will get 64-bit support when there is a phone that needs it, but it feels like that will be closer to the middle of 2014 rather than the end.
  
  I'm sure Intel's worried - the latest Bay Trail Atoms are basically even with the A7 in performance.
  
  And Bay Trail is on a newer 22nm process compared to A7's 28nm. I don't know the mind of Intel, but they have to be concerned that their "process advantage" is still not delivering concrete benefits.
Good by symbolset · 2013-11-23 07:23 · Score: 1

If you are going to run a silicon foundry business then telling your customers what not to invent is not a strong position.

--
Help stamp out iliturcy.
1. Re:Good by SpzToid · 2013-11-23 07:34 · Score: 1
  
  True. But only when/if Intel chooses not to be in the simple foundry business. When they choose to be market leaders, they've also chosen to give themselves enough wiggle-room to opt out of less-desirable deals. Intel wants to do both engineering of, *and* commodity production of, silicon.
  
  --
  You can't be ahead of the curve, if you're stuck in a loop.
2. Re:Good by Trepidity · 2013-11-23 07:44 · Score: 2
  
  Well, what they're doing historically is running a vertically integrated silicon foundry / chip-design business, not simply doing those two things independently. So they usually make sure neither side does anything to jeopardize the other, and if anything actively works to maximize the success of the other half. That can, in some contexts, be more profitable than running them as independent businesses that treat third-party customers neutrally. It's also usually legal unless you have a monopoly. If Intel had a foundry monopoly, they could be found to be misusing it if they refused to do business with companies that competed with one of their other business areas.
  
  --
  10 PRINT CHR$(205.5+RND(1)); : GOTO 10
seems unlikely by Gravis+Zero · 2013-11-23 07:28 · Score: 1

considering ARM is a fabless IC designer, i dont think they are going to jump on this offer.

--
Anons need not reply. Questions end with a question mark.
1. Re:seems unlikely by Anonymous Coward · 2013-11-23 08:01 · Score: 1
  
  Not ARM directly, but for many of the fabless ARM-licensees.
  For example: Apple.
  Which already has a huge partnership with Apple. So if Intel has excess manufacturing capacity, why not lease it out to partners?
2. Re:seems unlikely by Anonymous Coward · 2013-11-23 10:08 · Score: 1
  
  ARM is a fabless IC designer
  
  I think you meant to say, Arm is FABULOUS!!! sweetums!
this would be extremely good for nvidia by etash · 2013-11-23 08:48 · Score: 1

IF intel would allow nvidia to print its chips @ intel (and I say if, because they now are kind of competitors with the xeon phi and co. ).
AMD may benefit by Forever+Wondering · 2013-11-23 09:23 · Score: 2, Interesting

This forces Global Foundries to be more competitive with Intel, which benefits AMD.
GF, TMSC, etc. have been riding the [profitable] curve of being a generation back. That is, Intel is always a generation [or two] ahead, but also incurs significant R&D costs to do so. The competitors could wait and get the same results for far less investment in R&D. They could do this because Intel wasn't competing with them [by producing ASIC's, FPGA's, etc.]
This forces the non-Intel foundries to produce cutting edge stuff sooner. AMD was a big chagrined after spinning off GF and seeing it fall back into the TMSC model [making AMD less competitive against Intel].
The benefit for Intel is trifold:
- More ROI for their expensive fabs. Previously, costs were always recovered because the PC market was always expanding. With this now shrinking, a nextgen Intel fab may need to do piece work to stay profitable.
- Forcing the competition to compete head on [with the increased costs of being first generation], weakening them in the process [pun intended].
- A toe-in-the-water with ARM and mobile space [Atom notwithstanding] as a hedge against x86 arch going the way of the dinosaurs [without the stigma to x86 of a full fledged announcement of direct ARM support].

--
Like a good neighbor, fsck is there ...
1. Re:AMD may benefit by gtall · 2013-11-23 10:34 · Score: 1
  
  Intel cannot make money on ARM. My own opinion is that this is an attempt to knacker ARM development. No company should take Intel up on this.
2. Re:AMD may benefit by Forever+Wondering · 2013-11-23 12:00 · Score: 2, Interesting
  
  ARM is starting to encroach on x86 in the server space:
  - lower data center power requirements
  - they're coming out with a 64 bit version
  - ARM has a much smaller die footprint.
  Intel must do ARM to stay in that game.
  ARM would not be Intel's first foray into alternate architectures for x86 [8080, 8085, 8086, 80186, 80286, 80386, 486, 586, 686]. Remember Itanium [;-)] but also the 432. The Itanium and 432 didn't pan out because [the market for] x86 was so strong, but this indicates that while Intel is wedded to x86, it isn't slavish to it either. They care about making chips at a profit more than they do a given processor architecture. x86 has been a great tool to allow them to do that. But, x86 is just a means to an end for them.
  When x86 ceases to be the asset it currently is, Intel will adopt whatever the market demands. The trend for this is ARM (vs. sparc, mips, etc.). At this point, [even] Intel can't kill ARM. ARM has too much demand for it now [it's a better solution for mobile and embedded/hybrid systems and will surpass in the server space in the near future]. Intel is adapting/reacting now, while it has time to do so on its terms instead of waiting 10 years and being forced to do it in a panic.
  Contrast this with MS and Windoze. MS lost the mobile space race because of its insistence on Windoze. Intel won't make the same mistake, if for no other reason, they saw what it did to MS.
  As to MS, most likely, in 10 years, we'll see MS/Office running on OSX/iOS, Linux, and Android with Windows just a fond memory.
  Long term, Intel must become a foundry because it will lose its process generation edge (e.g. 22nm->14nm, but after 6(?)nm there isn't much room left. Others will catch up).
  Intel will make money on this. In the mid 80's, Intel was selling its first generation 386 chip for $750. An Intel engineer told me that the same chip was designed to be profitable even if it sold [had to be sold] for $35.
  
  --
  Like a good neighbor, fsck is there ...
3. Re:AMD may benefit by mikael · 2013-11-23 13:26 · Score: 2
  
  Intel could always do as they do and move to higher performance systems.
  Could this be a way of Intel to learn about ARM's technology and go one better? There are so many ways of optimizing CPU's now - more pipeline stages, lookahead stages, more registers, more cache, custom logic units, custom track layout, internal instruction sets. Intel themselves have said that their CPU's actually convert legacy x86 code into internal RISC instruction set code. So in theory, they could do the same with ARM code.
  
  --
  Vintage computer adverts: http://www.vintageadbrowser.com/computers-and-software-ads
4. Re:AMD may benefit by ebno-10db · 2013-11-23 13:40 · Score: 1
  
  Long term, Intel must become a foundry because it will lose its process generation edge
  Maybe. It sure looks like they're running into some pretty serious problems. EUV doesn't seem to be working out. Someday Moore's law has to end, but I've heard it predicted so many times that I take a wait and see attitude.
  Intel's bleeding edge graphene? Who knows.
5. Re:AMD may benefit by Forever+Wondering · 2013-11-23 16:44 · Score: 1
  
  You're right. Graphene is probably the future. I think so too. If the [zero] bandgap problem can be solved. There was an article on /. recently indicating that somebody had made a breakthrough here. As an interim, graphene might be useful in replacing the copper/aluminum intrachip interconnect.
  A graphene processor running at 100GHz would [probably] outperform a 32 core x86. Initially, a graphene chip would probably have far fewer gates. Thus, a true RISC ISA [like ARM] would be the logical choice. It could run x86 in software emulation (e.g. QEMU) and still beat the pants off the best Intel offering. Yum.
  As to process/die shrinkage, I, too, have seen the announcements about the end of Moore's Law. They always seem to occur at the same interval of ML [2 years]. I've also seen articles about techniques to extend ML for the foreseeable future. No worries.
  The other part of the puzzle is memory bandwidth. Getting a cache miss [no matter how large a cache (e.g. 12 MB)] slows things down to the point where advances in chip computation speed stall on memory transfer.
  For that, we'd need a DRAM replacement like magneto-resistive RAM [or ferroelectric RAM]. MRAM has the same cell size as DRAM, retains data without power, and is at least 10x faster than DRAM. L1 cache is pure static RAM [which has active power draw and large footprint]. MRAM is as fast as L2 cache.
  Hewlett Packard [which has been taking the lead in MRAM development] has a roadmap for MRAM deployment:
  - Replace flash memory [and unlike flash, MRAM doesn't "wear out"]
  - Replace DRAM
  - Replace/eliminate L2 [and L3] cache
  - Put MRAM at the heart of an SOC solution
  HP is willing to license the tech to anybody that wants it. Unfortunately, the last announcement of any real progress was a while back.
  
  --
  Like a good neighbor, fsck is there ...
6. Re:AMD may benefit by LordLimecat · 2013-11-23 21:07 · Score: 2
  
  Every benchmark ive seen suggests that noone comes remotely close to the performance or power-per-performance of a bog standard Xeon.