Intel Invests In ASML To Boost Extreme UV Lithography, 450mm Wafers

MrSeb writes "When Intel goes looking for new chip manufacturing technology to invest in, the company doesn't play for pennies. Chipzilla has announced a major investment and partial purchase of lithography equipment developer ASML. Intel has agreed to invest €829 million (~$1B USD) in ASML's R&D programs for EUV and 450mm wafer deployment, to purchase €1.7B worth of ASML shares ($2.1B USD, or roughly 10% of the total shares available) and to invest general R&D funds totaling €3.3B (~$4.1B USD). The goal is to bring 450mm wafer technology and extreme ultraviolet lithography (EUVL) within reach despite the challenges facing both deployments. Moving to 450mm wafers is a transition Intel and TSMC have backed for years, while smaller foundries (including GlobalFoundries, UMC, and Chartered, when it existed as a separate entity) have dug in their heels against the shift — mostly because the shift costs an insane amount of money. It's effectively impossible to retrofit 300mm equipment for 450mm wafers, which makes shifting from one to the other extremely expensive. EUVL is a technology that's been percolating in the background for years, but the deployment time frame has slipped steadily outwards as problems stubbornly refused to roll over and solve themselves. Basically, this investment is a signal from Intel that it intends to push its technological advantage over TSMC, GloFo, UMC, and Samsung, even further."

Comments

good slashdot

[Anonymous Coward]

now this is the slashdot kind of post i remember!

Re:good slashdot

[ifiwereasculptor]

Yes, actual relevant news about improvements in technology. And it's seeing way less comments (and traffic, one assumes) than the more recent article about San Francisco not buying Apple products. It's a sad state of affairs.

Re:good slashdot

[Burz]

It's a sad state of affairs.

In your mind, that is. Today more people recognize that sustainability has to be part and parcel of high tech, or else it is merely "so-called high-tech".

Re:good slashdot

[Anonymous Coward]

fine. but the op is talking about an article on the *politics*
of green computing, not the *technical* aspsects.

Re:good slashdot

[kelemvor4]

It's a sad state of affairs.

In your mind, that is. Today more people recognize that sustainability has to be part and parcel of high tech, or else it is merely "so-called high-tech".

People like to complain and argue. What's to complain or argue about here? The only negative I could think of is the potential money loss an investor would have if they had money invested in one of the competing fabs.

Re:good slashdot

[Anonymous Coward]

Yes, actual relevant news about improvements in technology.

nanometer < millimeter

ASML?

[grouchomarxist]

Who else was thinking "yet another damn markup language"?

Re:ASML?

[Bob535]

At least a few of us.

Re:ASML?

[ChristW]

They're about 5km from where I currently am, so, no :-)

Re:ASML?

[Black Parrot]

Who else was thinking "yet another damn markup language"?

YADML? You use it to mark up additional damns?

Re:ASML?

[symbolset]

Sadly, both YADML.com and YAFML.com appear to be taken.

Re:ASML?

[Anonymous Coward]

Did you also know that Holland is het capital of Denmark? You must be one of them smart dudes.

Re:ASML?

[Razgorov Prikazka]

Yes, and another thing: Denmark is the capital of Yurp! So they ARE smart!

Re:ASML?

[Errol backfiring]

I think he meant "Another Stupid Markup Language". Thanks for providing us with a new ETLA (= extended three letter abbreviation).

Re:ASML?

[Anonymous Coward]

Was thinking something a little more low-level, myself. After all, it has ASM in the acronym.

Re:ASML?

[Sulphur]

Who else was thinking "yet another damn markup language"?

Asynchronous Structured Markup Language : Implements a large decision table.

Monopoly

[voxner]

This is an ominous sign of things to come. Intel already has significant advantages in the foundry business. These could be leveraged further to give its x86 chips a boost vis-a-vis ARM. The other players need to pull their act together & pool resources to counter this. If there is no level-playing field because the foundries can't keep up we could well be facing an x86 monopoly in the low-power chip market too.

Re:Monopoly

[voxner]

Posted wrong link - intel-3d-transistors

Re:Monopoly

[symbolset]

I'm a HUGE Intel fan. Really, I am. But they have a rather serious Windows dependency they need to be quit of before they're ready to take on ARM, no matter what their process technology is, nor how fabulous their fabs. They are in serious danger of losing the plot.

Some eight years ago a laptop and desktop came to have the capabilities almost anybody needs. The innovation should have turned on that day to making the thing thinner, lighter and smaller; to making it run all day - but it didn't. Instead Windows became more bloated (as it always has) to drive new product sales for Intel and GPU vendors to make ever more powerful systems to give us more beautiful chrome. That worked for a while. It was great for sales and margins back in the day.

And then Apple came and reminded us that the purpose of the widget and the OS is not to sell more OS and more widget. It's to serve people in ever-evolving ways. To enable and empower people to do what they want to do, and get out of the way the rest of the time. To connect us to the things and people we care about. They came out with the iPhone, and then the iPad. They gave us what we had long craved.

Right about seven years ago ARM systems became "good enough" to do this and Apple released the iPod Touch - an innovative product that struck a chord with us. In 2007 came the iPhone. In 2008 Android. Ever since 2007 Intel has fiddled while Rome burned, producing "mobile" chips that burn multiple watts.

In 2005 the talk was about "the next billion users". It was always obvious that the next billion users wouldn't have watts. Well, Apple and Google have found that next billion users even faster than predicted. They're (we're) mobile. Between Android and iOS, they've sold nearly a billion devices - by the end of this year they'll get there - and now by ignoring the needs and wants of people Intel is in for a hell of a fight.

Even now their Windows pal is abandoning them, developing a new version of Windows without the chrome that requires their power-hungry CPUs, slimming it down to the point where a 7-year-old system is more than adequate and pricing it at a spot that's going to give legs to legacy systems and also building ARM-based systems under their own brand. That's going to kill new unit sales in every possible way for Intel. They had a good stretch where they got to milk that special relationship, but it's over now and then need to think about what next to do.

There are trust issues here that are very delicate. Buyers are not going to want to buy gear that leads back to the Bad Old Way where progress was slow.

I hope Intel figures this out. Really I do. But in the meantime I'm going to buy the kids, and Mom, Nexus 7 tabs for Christmas. My youngest son is almost old enough to teach how to build Android apps.

Re:Monopoly

[hankwang]

These could be leveraged further to give its x86 chips a boost vis-a-vis ARM. The other players need to pull their act together & pool resources to counter this.

Not necessarily. Once ASML has developed these technologies, they will be sold to all customers on equal terms. Moreover, unlike normal shareholders, Intel will not have voting rights and can therefore not easily influence the strategy of ASML. ASML's only obligation is that the R&D investment is allocated to development of said technologies. Other ASML customers (Intel competitors) are welcome to take a share in ASML on similar terms, so similar announcements from the competition may come during the next few months. You may want to read the official press releases.

You may be interested to know that ASML has 82% of the lithography market (by revenue), with equipment installed at most if not all manufacturers of CPUs and flash/DRAM memory. The semiconductor industry is driven by Moore's law; in a way, they are dependent on how fast ASML can develop equipment to produce ever smaller features. The interest of the ASML customers in this customer co-investment program is not so much in a competive advantage against each other, but rather to keep up with Moore's law.

Disclaimer: I work for ASML (in R&D), but the views above are my own, etc.

Re:Monopoly

[Teun]

There are other promising technologies waiting for investment:

http://www.mapperlithography.com/

Re:Monopoly

[ifiwereasculptor]

Intel is already out with a x86 phone that runs Android, is it not? Battery life and performance seems to be competitive with ARM offerings and I read talks about roughly 75% compatibility with existing Market apps. The thing about being Chipzilla is you can get blindsided by change and still come out on top, with a superior product, by throwing more money at R&D than the sum of all your competitors' assets is worth.

Re:Monopoly

[serviscope_minor]

I'm a HUGE Intel fan. Really, I am.

Why?

Do you like the way they abused their monopoly and forced AMD out of the market when AMD were very much superior? Or do you like the way they got off with a paltry $1.5bn fine? ($1.5bn offset against the profits and keeping AMD from getting a better foothold and holding up their R&D seems like a very good deal to me).

Their Core i7 chips are excellent (fastest per-thread by a wide margin) as are their graphics. By excellent, I mean just works out of the box and very stable and very well supported in Linux, which fits the bill perfectly for me.

But a fan? Why?

Re:Monopoly

[gl4ss]

pc hit the "what anyone needs" already 20 years ago.

but then there's things that it enables that are just nice to have, like better graphics creation, better games, better vr-applications, faster video encoding, video at higher rates, higher quality image editing faster..

I'd hate to have the innovation on need for speed to disappear, because frankly you can already buy thin & light machines that do communications duties but are shit for general computing. you could always buy them, even a decade ago, even x86. hell, first communicators from nokia ran on x86.

you're not seriously going to teach your son how to code android apps _on_ android are you? and to use image editing tools on android to use in those android apps? and to run raytracer on android? while all of that can be done it's pretty painful.

Re:Monopoly

[TheRaven64]

While interesting from a technical standpoint, it's just more of the same from a business standpoint. Intel's main advantage for the last 20+ years has been being consistently one generation ahead of all competitors in process technology. This let them survive the times (e.g. P4) when their designed turned out to be inferior to the competition, by either ramping up clock speed or using higher yields to lower prices.

The other players largely do pool their resources. This is why AMD no longer has fabs. Spinning them off into Global Foundries meant that GF could get business from other customers (e.g. ARM licensees) and increase the volume to closer to the point where they can invest as much in R&D as Intel. Intel is in a good position because they sell a large volume of relatively high margin chips. If Intel invested 10% of their operating income in R&D, AMD would not be able to match them, even if they invested 100% of theirs.

Re:Monopoly

[JasperKlewer]

Don't worry, ASML is offering equity stake to TSMC and Samsung as well: http://www.eetimes.com/electronics-news/4389932/ASML-in-talks-with-Samsung-TSMC-on-equity

Re:Monopoly

[bertok]

But they have a rather serious Windows dependency

You've got that the wrong way around. Microsoft has a dependency on Intel! Last time I checked, Linux runs just fine on Intel processors, and that combination powers a big chunk of the web. Some of the most important network appliances are BSD based, and run on Intel processors. Intel processors are also commonly used in systems like SAN and NAS arrays.

Some eight years ago a laptop and desktop came to have the capabilities almost anybody needs.

Citation needed.

I heard the exact same quote when I purchased my 20Mhz 486 SX back in the day.

For one, a typical desktop PC from 2004 probably can't play back 1080p HD video without GPU acceleration.

The innovation should have turned on that day to making the thing thinner, lighter and smaller; to making it run all day - but it didn't.

That's exactly what happened. You have to realize that "performance" and "battery life" are interchangeable. Increased performance at the top end allows underclocked low-voltage processors that still perform OK but draw a fraction of the power. Most of the last decade of transistor development has been about operations-per-watt. Either you get more operations per second at 100W, or it lets you stay at a constant level of operations per second while reducing watts.

The laptop that practically re-defined what it means to be light-weight and thin is the Apple MacBook Air, which is... wait for it... Intel based.

Instead Windows became more bloated (as it always has) to drive new product sales for Intel and GPU vendors to make ever more powerful systems to give us more beautiful chrome. That worked for a while. It was great for sales and margins back in the day.

On the contrary. While Windows comes with a larger installer package these days, that's mostly frameworks and drivers that aren't actually in use most of the time. Both Windows 7 and Windows 8 can outperform Windows XP on the same hardware in many cases!

You have to understand that the kernel is still pretty much the same thing, except that later versions have finer-grained locks, smarter schedulers, and revised driver models that allow more parallelism. None of this is "bloat". For example, "win32k.sys" on my Windows 7 SP1 64-bit operating system is just 3 MB in size. The closest comparison is Windows 2003/XP 64-bit, which has a 4.5 MB kernel. Hence, if anything, it's been shrinking!

They came out with the iPhone, and then the iPad. They gave us what we had long craved.

Walled gardens that don't even have a use accessible filesystem. Now, don't get me wrong, I have an iPhone and an iPad, but you're going to have to pry my PC from my cold dead hands.

The iPhone is great to have in my pocket, but I'm never going to sit at my desk pecking away at that thing when I could use a PC instead.

Right about seven years ago ARM systems became "good enough" to do this and Apple released the iPod Touch - an innovative product that struck a chord with us. In 2007 came the iPhone. In 2008 Android. Ever since 2007 Intel has fiddled while Rome burned, producing "mobile" chips that burn multiple watts.

What enabled ARM to do that is not some magic non-Intel or non-Windows approach, but reduced transistor sizes. Intel has been reducing transistor sizes too, and they're far better at it than the competition. The reason that Intel hasn't previously concentrated on the embedded market is not because they don't have the technology -- they do -- but because they saw it as a low-profit market that wasn't worth their trouble when they can be selling chips in the server market for $2,000 each. ARM's board would probably sell some of their limbs (hah!) for that market, which is why you've been seeing so many articles on Slashdot recently about ARM making inroads into the server space.

Ig

Re:Monopoly

[Anonymous Coward]

No, Intel really only took the lead ain the 2004-2006 timeframe. IBM was first with copper over a decade ago, and now the battle is Intel vs. the rest of the industry, which is finally realizing that they have to work together to stay not to far behind Chipzilla. This said, the end of process geometry shrinking is in sight (a bit less than a decade in my opinion) and new generations are going to come at a slower pace (the trend is already here, albeit very small). This will end up levelling the playing field between Intel and the other foundries.

Re:Monopoly

[Anonymous Coward]

Well, you are wrong on many accounts.

IF Intel sticks with it - CPU monopoly is gonna damage EVERYONE and literally their dog.
I dont see x86 issue on mobile, Windows/Linux/BSD/Darwin(on which Mac are built upon, though not sure about IOS) are all compatible with x86

Re:Monopoly

[unixisc]

Well, they were confronted by the FTC, and did settle w/ AMD and all their other competitors.

But this above story is a great reason to be a fan of Intel. Is it Intel's fault that AMD never figured out how to run its fabs, and went into a totally fabless model? Yeah, other chipmakers (like DEC at one time) did have their own fabs, and pretty good ones @ that, which due to larger mismanagement @ the company, they had to get rid of. How does one explain that HP, even in the late 90s, was fabbing the PA-RISC from Intel?

Intel, by contrast, had a fine manufacturing model from the beginning, and has built on it. Each of their fabs are exactly cloned fabs, due to which the process variations that one sees b/w different fabs of the same company is something one doesn't see @ Intel. It spends its money really wisely in being @ the cutting edge of foundry technology, putting it years ahead of its rivals.

I'm not a fan of the x86 architecture, but in this area, Intel is a victim of its own past success. While RISC was definitely superior, x86, being embraced like it was by Windows and its software, quickly became the standard. Intel itself had at least 3 unsuccessful home grown attempts to replace it - the i960, the i860 and the Itanium, not to mention its StrongArm acquisition as well as the fact that they could have embraced PA-RISC or even Alpha (after Compaq sold all IP to them), but the problem is that x86 was so well entrenched that when AMD took the somewhat counter-intuitive, yet simple, strategy of just extending x86 to 64 bits and were out way ahead of Intel, they looked like eating Intel's lunch. In fact, that was what ended the Intel-AMD wars for good w/ the cross-licensing agreement, and since then, if AMD had fallen behind, that's due to a combination of their own simplistic design paradigms and operational shortcomings. Incidentally, the only thing I was disgusted by was an unproven and inadequate microprocessor like the Itanium 1 being the cause of the deaths of the PA-RISC and the Alpha - both far superior CPUs - but for that, it's HP and Compaq that are to be blamed, not Intel. Intel never asked either of these companies to kill their RISC platforms - in fact, it was happily fabbing at least one of them.

I would have loved for there to have been another company to be like Intel and challenge it. I would have loved to see a RISC processor (other than ARM) dislodge the x86. But I'm not going to hate Intel for the fact that neither of these happened - it's not Intel's job to fall on its face so that others can compete.

Re:Monopoly

[Anonymous Coward]

No, the device is no good.

It is a single core CPU, which is not very good power draw, and just about competitive with the old crop of ARM devices in single thread performance (throughput on multi threaded work will be worse). Several vendors have the Cortex A15 based devices coming out very soon, which will really make the atom eat shit.

Its GPU performance is not good.

And it's not even ARM compatible, and is far more restrictive in terms of what a manufacturer may do with it. In order for a newcomer to make headway into an entrenched market, it really needs to provide some pretty compelling advantages. It's not enough to be just good enough (or in this case, "not good enough"). The one or two devices which came out with x86 are basically bought and paid for by Intel. Same thing happened last time around with their last failure.

Re:Monopoly

[serviscope_minor]

Well, they were confronted by the FTC, and did settle w/ AMD and all their other competitors.

They were. They got off with a 1.5bn settlement. I personally don't believe for a moment that the settlement came even close to covering the crime.

But this above story is a great reason to be a fan of Intel. Is it Intel's fault that AMD never figured out how to run its fabs, and went into a totally fabless model?

Quite possibly. AMD were stomping all over intel in the P4 era. The Opteron and Athlon were vastly superior products in almost every benchmark and cheaper to boot. However Intel maintained an 80% dominance through criminal activities.

During that time, Intel got to dump that ill-gotten profit into R&D to develop better chips and fabs, and AMD didn't. If AMD had got that enormous amount of money at that time, then history would have been very different. They would almost certainly ahve better fabs and better chips, for a start. Perhaps not quite as good as intel on the fab front, but much closer.

So yes, it probably is intels fault. That is why monopoly abuse is exceptionally damagind and why I think it's a traversty that intel got off scott free. Actually more so. It was almost certainly a net benefit to Intel wven with the 1.5bn settlement.

it's not Intel's job to fall on its face so that others can compete.

That's not the problem. The problem was illegal kickbacks] and bribes.

Re:Monopoly

[Anonymous Coward]

And then Apple came and reminded us that the purpose of the widget and the OS is not to sell more OS, but more widgets.

There. Fixed that for you.

Re:Monopoly

[Anonymous Coward]

Spot on sir, spot on.

Re:Monopoly

[Anonymous Coward]

I stopped reading at "But they have a rather serious Windows dependency"

they have a stranglehold on the low end and high end server market, and their making huge inroads into the mainframe market.

Re:Monopoly

[Anonymous Coward]

It's "slow" only when running emulated arm code. Slow meaning it's a little slower than the fastest ARM SOC's on the market. Look at the performance of native Intel binaries. Look at how the sun spider javascript benchmark is about 10x faster than the fastest arm based device.

The dirty secret is that ARM cpus are not actually very fast.. They're just the only modern cpus that operate in the very lean power sipping envelope that smart phones demand. They seem remarkable, but I doubt they're actually all that well engineered. We'll see a shakeout in a few years because there are far too many ARM SOC makers right now.

I feel Intel has a shot at bringing the "clunky" x86 arch to a level where it can beat arm handily in the mobile field. Don't underestimate Intel's engineers. Remember that Intel is a full 2 generations ahead of all of it's competitors when it comes to chip fabrication.

Re:Monopoly

[Anonymous Coward]

Intel didn't force AMD anywhere. AMD just phails at research and development, and their software sucks too.

Re:Monopoly

[Anonymous Coward]

I think that Microsoft needs Intel more than Intel needs MS. For many years now I been reading about the ever increasing amount of software coming out of Intel. It is no exaggeration to say they are one of the largest contributors to linux, not just in kernel drivers but userspace too. Everyday for years you can go to Phoronix and read about meego, wayland, cairo, clutter or contributions to xorg. Although defunct for the time being they did come seriously close to producing their own OS in meego.

It gave me the impression they were hedging their bets in regards to their relationship with MS. From this perspective the Microsoft nearly-but-not-really-acquisition of Nokia was like some kind of cold war fight with Intel.

At the end of the day no-one can produce or r&d cpus quite like chipzilla. When it comes to operating systems though there is much choice, including taking existing Free OSs and using/improving them to your advantage.

Re:Monopoly

[ChrisMaple]

Yes, Intel did bad and illegal stuff. The P4 generation was inferior due to the twin blunders of netburst and RAMBUS. However, Intel maintained its process advantage, and AMD sold all the chips it could make (so that even if Intel hadn't used illegal and immoral techniques, there wouldn't have been much of a change in the results.) When Intel reversed its architectural blunders, AMD couldn't even match Intel's architecture, let alone make up for its process technology lag.

The semiconductor industry is littered with the corpses of companies that thought they were better than Intel, and were wrong..

Re:Monopoly

[Anonymous Coward]

Right about seven years ago ARM systems became "good enough" to do this and Apple released the iPod Touch - an innovative product that struck a chord with us. In 2007 came the iPhone.

iPhone came first, iPod Touch considerably later.

Re:Monopoly

[Anonymous Coward]

No that's incorrect. It is not 10x faster than modern fast ARMs, that's ridiculous. If you find a benchmark where that is the case, it's a misconfiguration problem or issue in the JIT or JVM. Medfield Atom is below par performance with latest ARM CPUs (native vs native), and will be destroyed by A15 based ones that come out later this year.

They are not the only modern CPUs that operate in the smartphone envelope either. Apart from Atom (which only just scrapes in with Medfield), there are a lot of other embedded low power and high performance CPUs out there. Look at MIPS, powerpc, to name a couple of classes. And in fact you can go down to vastly smaller and less power consumption than these (which are high end ARMs).

They're pretty well engineered. They work in safety critical systems far more than x86 CPUs do, and they ship more devices per year than any other CPU architecture. They've been around for years, long before smartphones.

A15 is a pretty advanced out of order, 3 issue CPU that will clock at 2GHz on 28nm. Medfield is an in order, 2 issue CPU that can fit only half the number of CPUs into the same power envelope as A15, and clocks significantly lower.

Intel has a significant process lead, but that has not helped it break into the smartphone space (no, medfield is not their first failed attempt).

Re:Monopoly

[unixisc]

The next challenge is to clamp the current performance levels, and drop the power consumption. Some of it will happen due to shrinks, since the internal transisters will want VDDs far less than 1.8V (but not below 0.7V). As the VDD drops, the power consumption should drop by a squared factor. Once power consumption is low enough that a single AAA can support 34-48 hours of operation, CPUs would be as close to perfect as possible.

Re:Monopoly

[Anonymous Coward]

Note: when I say "smartphone envelope", it is in response to your statement "They're just the only modern cpus that operate in the very lean power sipping envelope that smart phones demand.". There aren't actually MIPS or powerpc etc used in smartphones that I know of, but they are used in areas with similar power and performance profiles.

Re:Monopoly

[unixisc]

It would seem to me that AMD's lagging Intel in architecture (once Intel reversed its policy of prioritizing on the Itanium) was at least partially a result of their being a generation behind Intel in the process technology. As a result of that, if Intel could offer 4 cores in a CPU, AMD could offer just 3, since they were a full generation behind. Generally, the design engineering and process engineering have to be synergistic in order to work, and it's ideal if one is capable of covering any shortcomings of the other. But if one is a full generation behind a competitor in terms of process shrinks, then having a far superior design is not likely to help (unless the other design is genuinely crippled).

And given that AMD has had no problems being sold out, it's not like Intel has been capping their revenues.

Re:Monopoly

[unixisc]

Well, Intel tried to replace the x86 platform w/ Itanium, but despite being that much bigger than AMD, they couldn't. The reason x86 is not an issue on mobile is that mobile doesn't have the legacy software baggage that the desktop has.

Re:Monopoly

[symbolset]

Quite the interesting comment there. Nicely done. I don't get quite so many from the 'softy fans as I used to back in the day. Maybe I should try to be more polite.

Some eight years ago a laptop and desktop came to have the capabilities almost anybody needs.

Citation needed.

It was plenty. Eight years ago was 2004. We were on Dothan by then, and the last Prescott, which means PCIe and SATA to erase the bandwidth bottleneck. Two years later was even better as we got a major leap just then. These were (are!) killer chips for software that isn't utter crud given a decent GPU. XP on these is still a great experience - especially if you match them with SSD storage. Linux on them is a great environment, and I suspect W8 will be too, though I haven't fully put it through my compatibility lab yet. Vista of course was unsatisfactory, but W7 was somewhat better and the W8 preview practically flies. I'm actually writing this on a laptop from this era - upgraded some of course . I have a more recent one but manufacturers have gone to very poor display resolution because of HDTV economies of scale and I keep that one in reserve for when I need the HDMI out or have to run virtual machines. I have no idea what sort of citation you're looking for here. This isn't some debatable thing that is even possibly in doubt.

You just contradicted your earlier statement about increased bloat in Windows.

No I didn't. For 25 years the saying has been "Intel giveth, Microsoft taketh away." It was working for the two of them. It has changed, and Microsoft has reportedly decided to come out with a thinner, lighter OS with Windows 8. They've killed Aero, for example, which will be fabulous for VDI. While I doubt it goes as far as Linux or BSD in this direction, reports are that it is so much thinner and lighter than W7 that it's like getting a PC upgrade to install it. My own experience seems to agree somewhat, and it seems likely the golden image when it drops will be even better. This is of course not going to help Intel: they would be down to capacitor failure incidence to force hardware upgrades, and capacitors have better lifetimes now than they did a few years ago. I'm not a big fan of W8 for other reasons, but Microsoft appears to at least have discovered a need for performance optimization that is death to Intel's upgrade cycles - and if it isn't a winner Intel isn't going to move a lot of desktop chips anyway as OEM PC sales crash again as they did with Vista before the "downgrade rights" stopgap. The change to this dynamic was quite exactly my point, and you missed it.

I'm not even exaggerating, Intel's revenue is about 50 billion compared to ARM's 500 million!

ARM's profits are adequate for their tiny staff. On a per-employee basis they make far more than Intel. By licensing the IP they allow a bigger share to their manufacturing partners like Samsung, nVidia and Apple who, if I recall correctly, are doing quite well these days. By leaving room for partner profits they benefit from having many successful partners. Somebody believes in ARM Holdings, Inc. ARMH P/E is 52 - 5x Intel's, which is an expression of the market's general confidence in their prospects for the future. ARM doesn't just design IP for mobile clients though - they do embedded devices, routers and switches, the iLO on the latest HP servers, their own server designs are in wide trials. Far more ARM processors ship today than Intel processors do - and this has always been true. You will be tempted to say, I'm sure, that Samsung is the only profitable one and if I were you I wouldn't go there because it's provably untrue.

I doubt they're worried about ARM.

This was the point of my post. I doubt it too and that makes me concerned because as I said, I am an Intel fan. They should be worried, and they should have started being worried long ago -

Extreme UV Lithography?

[Black Parrot]

Is that a new competition at the Olympics this year?

Re:INTEL RULEZ !!

[Anonymous Coward]

That's what I'm thinking, right? Normally I have to fucking SWALLOW to get this kind of info, but slashdot and Intel are here to bust a nut for me. Thanks!

Re:INTEL RULEZ !!

[Anonymous Coward]

What ever happened to the good old american pussy? Ass this, ass that, ass, ass, ass. Is it cool to have prison sex? I don't think so but what do I know. Where's Hans Reiser on this?

No guarantee of success

[Anonymous Coward]

Intel has the money and the lead right now. If this investment doesn't pan out it just might mean the end of Intel's lead and its money. It's a pretty big bet by Intel and it might pay off. I hope it does or else TSMC will never get to 450mm wafers. Intel might be the only company that can justify 450mm wafers. They have the big chips, the big volumes, and the tiniest features and the development budget. A lot of other people can make money on smaller chips, in smaller volumes with larger features and affordable development costs.

Re:No guarantee of success

[unixisc]

Usually, the first few years of a fab are its most expensive, when it has to both operate @ capacity and be profitable. Essentially, using the ASML and other technologies, Intel would either build new fabs or upgrade some of their existing ones to 45cm technology. Here, they'd get twice as much die as they got on an equivalent 30cm wafer on the same lithography node (e.g. 22nm), but if one factors in that there might be a die shrink involved as well, make that even more. Translated into units, Intel would simply have to sell a whole lot more.

There are a variety of ways that they can do it - building everything from Atoms to Xeons and their complete chipsets, fabbing chips for other chipmakers, and last but not least, when the fabs have depreciated considerably, fab out NAND flash memory that can be used in making their SSDs. The last should really have fabs operating @ full volume. Once they are there, that fab is a pure asset, except that it has to be kept running full time regardless of demand, which is one of the pitfalls that makes many companies consider going fabless.

Re:No guarantee of success

[hankwang]

Intel might be the only company that can justify 450mm wafers.

As another poster mentioned, Intel is not the only one who wants 450 mm wafers. A big part of the cost of wafer processing is proportional to the number of wafers and not on the surface area; that's why a transition to 450 mm will lead to cost reduction. This cost aspect actually doesn't apply for ASML's lithography tools (or so I believe), since the tool throughput (wafers per hour) is roughly inversely proportional to the wafer surface area. The throughput depends largely on how fast you can move wafers around with nanometer-level accuracy (does not get easier with a bigger wafer) and how much light power is available (which doesn't depend on the wafer size).

Disclosure: I work in ASML's R&D department (not on the 450 mm stuff), but the above views are my own.

You lost me with BS about Global and non-450nm

[tyrione]

Do you folks even realize IBM, TSMC, Global Foundries and Samsung announced their 450nm production back in March?

http://blog.timesunion.com/business/tool-makers-waiting-for-clarity-on-450mm-cost-sharing/53301/

Tool makers waiting for “clarity” on 450mm costs
March 28, 2012 at 10:45 am by Larry Rulison

The companies that supply the costly manufacturing equipment to computer chip factories – also known as “tool” makers – are waiting to get “greater clarity” about how much they will be asked to pay for the industry’s transition to using 450 millimeter silicon wafers.

The Times Union reported Tuesday that the tool makers will be asked to foot $450 million of the $1 billion price tag for the first phase of a 450mm transition program that will take place at the University at Albany’s College of Nanoscale Science and Engineering.

Deborah Geiger, a spokeswoman with SEMI, the San Jose, Calif. trade group that represents the tool makers, said the organization is hosting a forum on April 4 at the NanoCollege that will touch on the issue of how the 450mm program will be structured.

“We are not aware that definitive details and amounts have been established and publicly communicated,” Geiger said. “SEMI members are interested in greater clarity around the program structure and funding, including the cost share scenarios.”

The details included in the Times Union story were included in documents used by the Empire State Development Corp. in its approval of $300 million in funding for the NanoCollege for the 450mm program and another IBM program to shrink chip features nearly in half, down to 14 nanometers.

New York state is providing $150 million in cash and $50 million in cheap power, for $300 million total, toward both programs, which will be located inside the college’s new $365 million NanoFab Xtension building under construction on Washington Avenue Extension.

Five leading chip companies that make up what’s known as the Global 450mm Wafer Development and Deployment Consortium – Intel, IBM, GlobalFoundries, Samsung and TSMC – will each contribute $75 million over five years toward the 450mm program.

Geiger says that a meeting is expected to be held in May in which suppliers to the G450C will be provided with a “more complete communication” on the 450mm program and how involved the tool makers will be.

Computer chips are currently made on wafers that are 300mm, or smaller. But the move to 450mm would save incredible amounts of money for manufacturers since output would roughly double with the larger size wafers.

All the players are in the game:

http://phys.org/news/2012-07-imec-nanophotonics-components-300mm-silicon.html

Re:You lost me with BS about Global and non-450nm

[Anonymous Coward]

Do you realize that mm != nm ?

Successful economy of scale

[roman_mir]

Intel is an economy of scale, they are not a monopoly until government steps in to prevent other companies from investing into microprocessor design and manufacturing.

Of-course I always argue that patents and copyrights DO provide a government created barrier to entry into any market, especially high tech stuff.

Now, if Intel is not providing you with a product you want and you think they are overcharging and you believe there is a FINANCIAL sense in having another designer and manufacturer company in the market and it would be profitable by being able to sell because of whatever efficiencies and better product, better business, then clearly you have an opportunity that you should approach investors with.

There are plenty of millionaires and billionaires that want to invest into a sound business idea and if they are not investing it's because they don't think the idea is sound (that's why the California plan for high speed rail is not done privately - nobody is stupid enough to believe they can make a profit there, only government doesn't care about losing all sorts of tax payer money on nonsense projects that market isn't willing to take on).

So you think you have a good idea? What is stopping you (except the shitty economy of-course, but again, that's gov't for you with all the inflation, laws, taxes). If you can make a better processor and it will be cheaper, don't you think you'd be able to SELL it?

You think you can make a processor better than Intel and sell it cheaper than Intel? You can be a billionaire yourself. Do it, I want a cheaper better processor, go ahead.

Units of length

[unixisc]

Since the wafers are growing larger, unlike the process sizes, which are shrinking, they should use larger units. Previously, they'd use 0.45 microns, but when it shrank considerably more, they started using nanometers to represent the process size. Conversely, previously, they'd call the wafer sizes 200mm or 300mm, but now, since there is the potential of confusion between 0.45 micron process size and 0.45 meter diameter of a wafer, since the former will be 450nm and the latter will be 450mm, they should change the units of the wafer diameter as well. Either call it 45cm or 4.5dm (deci-meters) - I'm assuming that we're not likely to ever see wafer sizes reach 1 meter, since it would complicate the transportation of those things.

Re:Units of length

[Anonymous Coward]

No they shouldn't.

No one ever made chips with a 450nm process. So that was never used, and can never be confused with 450mm.

More still... if 450nm had ever existed, it would have existed in he mid 90's, and be 15 years outdated by now.

Re:Units of length

[butlerm]

You are off by a factor of ten. Process sizes are currently in the 22 - 32 nm range. 0.45 micron (450 nm) would have been close to the state of the art almost twenty years ago.
http://en.wikipedia.org/wiki/22_nanometer

incorrect

[slashmydots]

Basically, this investment is a signal from Intel that it intends to push its technological advantage over TSMC, GloFo, UMC, and Samsung, even further.

Really? Because I love it when the person with all the money puts up a huge bet at the poker table. I can't win unless they do. If they play it ultra conservative, I have no chance.

Re:INTEL RULEZ !!

[Anonymous Coward]

Am I the only one experiencing issues with Javascript events on Slashdot in Chrome? It's been broken for me for like a day now.