Intel postpones Arizona chip plant opening as demand for PCs wanes

Shawn Knight

Posts: 15,284   +192
Staff member

intel arizona arizona chip factory

A decline in personal computer sales has forced Intel to postpone the opening of a large chip manufacturing plant in Chandler, Arizona. Fab 42, deemed as a $5 billion project late last year, will remain unused for the foreseeable.

The Arizona Republic was the first to report the development which has since been confirmed by Intel spokesperson Chuck Mulloy via Reuters. In a statement on the matter, Mulloy said the newer fab has not been equipped with the capital equipment. Specifically, it has heating and air but it lacks the actual tools to build chips – or as he noted, the “expensive stuff.”

Mulloy said the new construction will remain vacant for now and will be targeted at future technologies although no timetable for a grand opening was mentioned. In the end, it all boils down to better capital utilization, he concluded.

In the meantime, other factories on the same site are being upgraded to produce Intel’s more advanced chips. At present, those facilities build chips on the 22-nanometer manufacturing process. The initial plan was to use the new fab to built 14-nanometer parts but that equipment will now be installed in the other facilities.

Elsewhere, the company is also working to expand a plant in Oregon that will be used to make larger chip wafers. The largest wafers used in the chip industry now measure 300 millimeters across but the Oregon plant will be capable of wafers measuring 450 millimeters.

Permalink to story.

 
Odd, I would have thought after spending the money you might as well use the place. But I guess circumstance have forced their hand and they are now looking to an alternative direction for the place.
 
Odd, I would have thought after spending the money you might as well use the place.
Not really. The building is the least of the issues. The running/labour costs and tooling (which will be used at other fabs) are the big expense items. The other variable is that ASML's NXE:3300B - ostensively being used for the 22nm and 14nm process nodes, is also being validated for the 10nm and 7nm processes. At a likely $200+ million a pop and a production rate of less than one per month (closer to one every three months at present), the lithography tools are a much more highly valued commodity than any refurbed clean room facility, so it's odds on that the tools will go to the D1X fab in Oregon
But I guess circumstance have forced their hand and they are now looking to an alternative direction for the place.
Intel have had a massive overcapacity capability for some time - I've mentioned it myself a number of times in the past. Until Intel weans most of its non-CPU chip production off older processes (many controller chips still use 65nm) there simply isn't enough products to warrant the process capacity that they have. It is also the reason that Intel started courting other companies for their silicon fab needs.
Wake up MS, Intel is starting to hurt! </End Trolling> lol
A victim of their own success....and to think, back in the early days when Intel began fabbing EPROM's and DRAM, their fabrication yields were a joke...now they have what is likely the most sophisticated fab setup producing some of the best yields in the industry.
 
Last edited:
I'd say that not just a lower demand for whole PCs have caused this, but from enthusiasts as well. Anyone with a Sandy Bridge has no real reason to upgrade their CPU right now; it's better to just get a new video card.
 
Make it cheaper, and make it so that we don't have to buy a new motherboard everytime, then more people will buy, and more often.
 
Not really. The building is the least of the issues. The running/labour costs and tooling (which will be used at other fabs) are the big expense items. The other variable is that ASML's NXE:3300B - ostensively being used for the 22nm and 14nm process nodes, is also being validated for the 10nm and 7nm processes. At a likely $200+ million a pop and a production rate of less than one per month (closer to one every three months at present), the lithography tools are a much more highly valued commodity than any refurbed clean room facility, so it's odds on that the tools will go to the D1X fab in Oregon

But your missing one thing, in the article Intel said
The initial plan was to use the new fab to built 14-nanometer parts but that equipment will now be installed in the other facilities
so in other words they are instead of installing it there, just going to push it on their other facilities and replace the components/upgrade them for the 14nm chips. But either way it does not matter, its just odd to waste such a huge facility like that, but I guess it can just become a giant storage facility.
 
But your missing one thing...
Actually I'm not. The article says exactly what I said. The "equipment" largely centres around the lithography tools (see last link in this post) - the NXE:3100 and NXE:3300B. The latter have the supposed ability to produce 10nm and 7nm wafers, why would Intel underutilize them producing 22/14nm parts that can be made at Fab 24, D1D, and D1X ? Especially when A. The market for 22nm parts is limited, and B. the NXE:3300B is as rare a rocking-horse sh**.
so in other words they are instead of installing it there, just going to push it on their other facilities and replace the components/upgrade them for the 14nm chips
Great. You just paraphrased my post- albeit with substantially less information...
The other variable is that ASML's NXE:3300B - ostensively being used for the 22nm and 14nm process nodes, is also being validated for the 10nm and 7nm processes. At a likely $200+ million a pop and a production rate of less than one per month (closer to one every three months at present), the lithography tools are a much more highly valued commodity than any refurbed clean room facility, so it's odds on that the tools will go to the D1X fab in Oregon
Just to make it crystal clear...tools originally earmarked for the Chandler fab, are likely to be diverted to the Hillsboro fab complex - since A. It was completed last year (Fab 42 isn't complete), and B. D1X is the site of Intel's new fab process research, and with Module 2 needing the same ASML litho tools next year that Fab 42 would use...litho tools that will be in very short supply.
 
Great. You just paraphrased my post- albeit with substantially less information...
So you paraphrased the article and added the machine names, because that info was gotten directly from the article were reading as I quoted above. Either way it does not matter, the tools are being moved to another location instead of the new facility end of story.

With that in mind, a 5 billion dollar plan trashed to just re-allocate the tools to other facilities after setting it up is a waste. Only the future can tell what the facility will be used for but it still seems like a waste of something already setup with heating and AC even if these components are not the most expensive parts of the facility because Intel is paying for the facility whether or not they are using it for anything.
 
So you paraphrased the article and added the machine names, because that info was gotten directly from the article were reading as I quoted above.
Well, to be fair, the article is a little short of specifics- and somewhat late, considering the information was being dissected by financial institutions and Intel shareholders over the past few weeks. Also, I was originally replying to your question of why it is better to mothball a plant than have it churning out product no one needs.
With that in mind, a 5 billion dollar plan trashed to just re-allocate the tools to other facilities after setting it up is a waste
Only of you don't know the back story.
Governments/states supply financial rewards and incentives to have large industry set up house on their turf. As an example, Intel receives substantial tax incentive from the Irish government to build and maintain Fab 14 at Leixlip...a tax incentive not available in Arizona, and an incentive not matched by Israel. Safe to say that tax breaks on Intel's other operations, and a quicker ramp of D1X Mod2 help offset keeping Chandler offline.
And BTW, Fab 24 isn't "trashed", just mothballed. The site is also an asset, and likely attracts a tax break for depreciation whilst it remains in Intel's portfolio.

The other alternative is to allow the site to go on line, and to produce chips Intel doesn't need, using labour and fixed overheads Intel has to pay for. Copying AMD's example of stockpiling chips nobody wants and taking a write-down on inventory in addition to production costs doesn't seem like a good example to follow.
 
Last edited:
See what happens when you project your future needs based on an imaginary, unwritten, unprovable, and mostly likely purely speculative law?
 
See what happens when you project your future needs based on an imaginary, unwritten, unprovable, and mostly likely purely speculative law?
Also to be fair, Moore's law is only tangentially involved, since it relies upon transistor density at a given cost. Probably mo(o)re a case of Intel leveraging as many subsidies/properties as possible (playing the semiconductor version of Monopoly) and underestimating the bite that the ultraportables and phones are taking out of the their traditional markets.

At the end of the day, it is still a serviceable foundry once the tooling goes in. If Intel ever decide to go back into producing memory ICs, or diversify further into GPUs or any other integrated circuit , they have the option of a quick ramp with prodigious throughput at hand. In some bizarro parallel universe, Intel are offering AMD and Nvidia 14nm wafer starts and cutting TSMC's 28nm throat.
 
Also to be fair, Moore's law is only tangentially involved, since it relies upon transistor density at a given cost. Probably mo(o)re a case of Intel leveraging as many subsidies/properties as possible (playing the semiconductor version of Monopoly) and underestimating the bite that the ultraportables and phones are taking out of the their traditional markets.
Relax. I was merely extrapolating completely out of context. The chat about Moore's Law Intel had with its stockholders awhile back. seemed to foreshadow stormy times ahead. And verily, lo and behold, here they are.....:D
 
Relax. I was merely extrapolating out of context, the chat about Moore's Law Intel had with its stockholders awhile back. That seemed to foreshadow stormy times ahead. And lo and behold , here they are.....:D
Yep, Intel's overcapacity and the slow down in some of their core businesses have been an issue for a while. The shareholders have had a pretty easy ride for the last few years (at least since 1995) and probably aren't used to bad news. Intel's once-bitten-twice-shy approach (or lack of) to product line diversification makes them look an oil tanker amongst a fleet of America's Cup catamarans, yet if they continue to plot the familiar path they end up doing everyone else's R&D since they have to maintain a clear superiority in a diminishing market.

That said, I don't think anyone at Intel is in any danger of having to apply for food stamps just yet.
 
Any wonder pc sales are slumping.....the lack of competition from AMD at the high end has allowed Intel to slow development a little.....so there is just no real point in upgrading.....give us something worthy to upgrade to and we will give you are money.....atm my 2700k at 4.5Ghz is doing fine.
 
Must be great to spend $5 billion on a building and then not use it. I recommend turning Fab 42 into Fab 42 Apartments, forty-two apartments that are each roughly the size of a 3,000 sq ft home, which come fully furnished with state of the art amenities.
 
Last edited:
Only of you don't know the back story.
Governments/states supply financial rewards and incentives to have large industry set up house on their turf. As an example, Intel receives substantial tax incentive from the Irish government to build and maintain Fab 14 at Leixlip...a tax incentive not available in Arizona, and an incentive not matched by Israel. Safe to say that tax breaks on Intel's other operations, and a quicker ramp of D1X Mod2 help offset keeping Chandler offline.
And BTW, Fab 24 isn't "trashed", just mothballed. The site is also an asset, and likely attracts a tax break for depreciation whilst it remains in Intel's portfolio.
The idea of offering Tax incentives and break among other things is for things like bringing jobs or similar to the area. Its the idea of helping the state and the people or where ever your building it. As it is right now, they just added a massive warehouse and are just putting it on stand by for awhile. They may still have those incentives, but the building is just for now sitting with no use and using electricity and Intel is footing the bill for it but your correct that the incentives are probably more than enough to cover the cost of the warehouse at least in the long run.

Relax. I was merely extrapolating completely out of context. The chat about Moore's Law Intel had with its stockholders awhile back. seemed to foreshadow stormy times ahead. And verily, lo and behold, here they are.....:D

We all are feeling this...the tech industry is taking hits as of late and we are all getting to the point that the desktop market is becoming almost taboo in many respects. I still wonder when the day will come that we will be able to go on newegg or the likes and order parts to build a laptop since mobile and ARM seem to be the king right now.
 
I can see it now, the enthusiast water cooling their phones. lol
I-dont-want-to-live-on-this-planet-anymore.jpg


When that day comes...
 
I still wonder when the day will come that we will be able to go on newegg or the likes and order parts to build a laptop since mobile and ARM seem to be the king right now.
IMO, no.
Desktop has DIY options simply because that's where the personal computer started (Mark-8, Altair 8800, SWTPC, IMSAI etc.), something that has never applied to the portable/mobile markets.
Given the component size, and (dis)assembly methodology, and lack of standardization- along with the rabidly proprietary nature of the markets, I wouldn't hold your breath waiting for IHVs and mobile providers to reach some accord. Of course, if they did I sincerely doubt it would be to the benefit of the consumer.
 
IMO, no.
Desktop has DIY options simply because that's where the personal computer started (Mark-8, Altair 8800, SWTPC, IMSAI etc.), something that has never applied to the portable/mobile markets.
But you can't help but wonder if the DIY crowd will force change and migrate to smaller form-factors. It is a big crowd and we love our hobby, no matter how small the tech gets.
 
IMO, no.
Desktop has DIY options simply because that's where the personal computer started (Mark-8, Altair 8800, SWTPC, IMSAI etc.), something that has never applied to the portable/mobile markets.
Given the component size, and (dis)assembly methodology, and lack of standardization- along with the rabidly proprietary nature of the markets, I wouldn't hold your breath waiting for IHVs and mobile providers to reach some accord. Of course, if they did I sincerely doubt it would be to the benefit of the consumer.
Meh, maybe not but it would be an interesting and something I would at least appreciate. In all honesty besides the size a laptop would not be all that much different from a desktop. Yes a laptop has many difference from its desktop counterpart, however I feel it could be done if a few companies would standardize a case and motherboard then it could be possible. I doubt it will happen, but there is always a chance.

The closest we have to customizable laptops, would be companies like Sager and MSI who allow us to change GPUs (minus when they bios lock), CPU's, Ram, and HDD's. But that's probably as far as we will get sadly.
 
But you can't help but wonder if the DIY crowd will force change and migrate to smaller form-factors. It is a big crowd and we love our hobby, no matter how small the tech gets.
Possible, but unlikely I would think. Even something as rudimentary as a graphics card change isn't easily effected in a standard sized laptop - an item that has been fairly ubiquitous for thirty years. For an everyday computing product range with a built in market for graphics upgrades, there are very few (r)etailers of MXM modules.

For the DIY thing to happen, you'd need a great many manufacturers to arrive at a common standard for ultra's and mobiles - size, form factor, power consumption, cooling. Theoretically someone could come up with a "barebones" type approach, but whether the userbase is large enough to break the partnership service providers have with hardware vendors is an interesting question to ponder, as is warranty status for a DIY tablet, phone or notepad.

Not sure about the "big crowd" argument either. Consumer computing is a big crowd but the vast percentage of sales are via OEMs and ODMs. If anything, the move seems to be toward embedded systems with fewer upgrade options- less inventory and warranty hassles for the OEMs and IHVs, accelerated built in obsolescence, and a cookie-cutter product line that appeals to the majority of consumers.
 
FWIW, I can't see any practicality, pride of accomplishment, or even a use for building a laptop, tablet, or phone.

It's been said since PCs developed a certain level of potency, that most people who owned them, had very little use for the extraordinary power they posses.

So, "personal computing", has been handed over to a level of utility designed to satisfy the armies of imbeciles that inhabit this planet.

Phones pass for cameras, iPads masquerade as computers. With limited storage available in these mediums, the marketers have in one fell swoop, deprived people of their independence, whilst addicting them to extravagant internet, phone, and media subscription charges.

It really is a sad commentary in this day and age, that children either need, or believe they need, a $600.00 mobile "smart phone", to be able to successfully walk themselves to school.
 
It really is a sad commentary in this day and age, that children either need, or believe they need, a $600.00 mobile "smart phone", to be able to successfully walk themselves to school.
Monkey see. monkey do.

Less about need than about not standing out from the herd. All you have to do is have the celebrity icon for the day shill the product, and add in some strategic product placement. In an era that prides itself on txtspk, an attention span no longer than 140 characters, and a "facebook activist" mentality where the most important issue of the day is superceded every five minutes, you really only need to see your next favourite gadget getting screen time on this weeks primetime reality show to know it will be a "must have" tomorrow.
 
Back