Hi Peter - if this plays out, that is, if TSMC does in fact build a fab "next" to GF in upstate NY, this would be the epitome of competitor intimidation. I guess all is fair in "love" and semiconductor customer competition, especially in the foundry business.
unlike state like Texas and California, New York economy is so bad they are just giving huge cash away to Globalfoundries to make their fab. I think same situation for another company...just more business friendly in New York right now. But yes, end result may be intimidating to global
Yes -- especially compared to California. I don't think there is any way a company would consider a fab in Calif. right now.
The author doesn't mention it, but geographical diversity would help TSMC -- ensure stable production.
This is a little off topic, so let me say My family has more than 100 year history living in New York. We know New York economy is not good and are happy at the current efforts.
That said, not sure what metric is being looked at to say NY is better than California, but here is a good one...total state debt. California is horrible, 2x the #2 state, which not coincidentally is NY. So, yes, NY is better than California, but its still worse than all other states in the country. Check the link.
But unlike California, NY wants to spend their way out of debt by encouraging certain big business...like semiconductor. Bill Owens didn't visit TSMC to tell them that NY have a lot of talented workers, he visited to give Morris Chang a gift of land and tax breaks to build a fab.
Another reason to keep fab away from California is environment regulation, it's just harder to have a modern big fab in California, the air regulations just don't allow it. That's one of many resson why there aren't any modern fans in California.
What I'm interested in is opinion if there are any other logical candidate except TSMC for this site selection...I say not samsung, intel, global foundries, or UMC.
I wondered when this news would hit EE Times. This is big news for USA.
In 2010 there was big rumor that Apple forced Samsung to open logic fab in US to make the exclusive apple sales contract. If apple will move away from samsung in the future, they again may make US manufacturing requirement.
TSMC new fab in US and specifically new york makes a logical choice, but maybe only if Apple is requiring it. Otherwise just make another fab in Taiwan, or elsewhere in Asia for current customer.
Also intel doubling size of D1X fab in Oregon must have some big customer plan...either of these new fab could supply dedicated chip output to apple. And still samsung is making big expansion of their US fab for apple. So it could be 3-way race in US for apple contract between TSMC, Intel and samsung.
It's known that Apple ( and Qualcomm as well) attempted to convince TSMC to give them all of their production, but that TSMC refused. It's also known that Apple does deals with their manufacturers where they give them very large loans in the form of outright purchasing of machinery, financing for the building of manufactures, training of workers, etc.
I wouldn't be surprised if this move was inspired by Apple, and is even being partly financed by them. If any company can afford that, Apple can. It's said that a new 22nm plant can cost $8 billion. Even Intel is having problems coming up with that cash by themselves. But that's less than a quarter's cash flow for Apple.
This could be a very interesting development.
They are already expanding the foundry to have the capability to produce the 450mm wafer and the city just approved the construction of a new 3 story, 565,000sqft manufacturing research center. Combine that with Global Foundries 300,000sqft clean room and that is expanding as well i think they are beefing up quite well.
GF future plan for existing site as I know is 3 fabs, and they have dedicated future fab plan team already. So site selection for new land in Albany is not necessary for GF I think.
Intel, Samsung, and this other company (TSMC) have what seems more real and urgent plan. GF would have to act quick to make it happen in same time but has capability but maybe not money, and preferred plan was fab in middle east (crazy plan but that was where the money was)
Who would have think possible foundry capacity race in US??
Albany/capital region New York would be a wise choice for TSMC's next plant. With a region that is 400 years old & deep in industry & innovation from Thomas Edison's lab to the creation of General Electric. Combine that with the high quality of life the region offers, ranging from the beautiful Adirondack mountains where the Olympics were held twice at Lake Placid. To the home of the 4th best engineering school in the world Rensselaer Polytechnic Institute and also the home of Albany College of Nanoscale Science and Engineering which is the leader and pioneer in nanotechnology education, research and economic outreach. Also Albany is known for one of the most highly educated work forces in the world and is located to some of the most unique & innovative cities in the world like New York City, Boston and Montreal.
I believe it is just rumor and will not be true because it conflicts TSMC's cluster manufacturing concept which has worked for a decade. If being true, it only would be Apple urged and co-worked with TSMC to build Fabs in the United States. Location would be near Wafer Tech in Washington State. NY, Texas are not good choice.
Because the cluster methodology will work out really well in the event of a large seismic event in Thailand... it is only a matter of time before TSMC's customers demand some protection from that type of disruption. Plus as you mention, many customers in the US may be pushing TSMC to do it and not just Apple, don't forget qualcomm.
New York is actually a very good choice. There is excellent transportation, and facilities. A highly educated workforce (believe it or not), and a history of technology, with RIT being one of the most highly regarded institutions in the area. The U of Buffalo Is well respected in technology.
The fact that so many other high tech firms have located here is evidence of its compatibility.
Besides the actual building cost, there is also the operating/maintaining cost. Samsung and TSMC already have US locations for comparison, why add a different US location? NY one-time subsidy doesn't make up for subsequent operation/resource costs of 10 nm fab, which are very location-dependent.
TSMC doesn't have any benefit having fab in NY. They don't develop technology in IBM camp, so they have to transfer technology from Taiwan, which is not efficient. In addition, Taiwan and China have more flexible and cheaper workforce. Apple asking is non-sense, as Apple and Samsung are simply one negotiation away from getting deal done. Too risky for TSMC to build a fab in NY just for one particular customer.. Maybe joint venture (i.e., different legal entity) with one of larger customers (Apple or Qualcomm) to build a dedicated fab in US may make more sense.
- They can poach talent from IBM and GE
- Foreign Investor tax break
- Whatever deals they make with New York State and the US government. RPI is close by too!
- New York State
- New York State
and most of all
- New York State
Why not Apple?
Can Samsung keep up with production demands for Apple? Could GloFo keep up production demands? Is TSMC and Samsung providing the technology yields necessary to compete? So...highly doubtful Intel would go the NY fab route with the current investment in AZ and OR. What product is driving the industry that needs lower node size technology and higher production capability?
It is obvious where the fab is going to wind up: it will be stuck very tightly to whichever local government puts forth the biggest and juiciest taxpayer's-hard-earned-money-filled teats.
There was a wonderful article in the NYT a few weeks back about how local and state governments blow $80 billion a year poaching jobs from one another in this stupid, zero-sum manner. The solution is obviously a federal issue, so you can be sure nothing will get done because one party clearly has no incentive to stop this garbage.
A Book For All Reasons Bernard Cole3 comments Robert Oshana's recent book "Software Engineering for Embedded Systems (Newnes/Elsevier)," written and edited with Mark Kraeling, is a 'book for all reasons.' At almost 1,200 pages, it ...