1. There are few industry suppliers of process equipment. My feeling is more and more they are providing process solutions to customers. As this trend continues you get a broadening of technology to all customers...oh yes many GREAT equipment supply companies are not US based and thus have different views on the legalities of sharing tech.
Looks like you're talking about the AMAT model. Granted there are a few equipment suppliers out there trying to dominate the market, but I'm honestly not convinced that the tool vendors will be successful in creating a "solutions-based" approach covering enough of the process to make your model work. The days of simply scaling a process are over. When the same basic building blocks were used by everyone and simply scaled from generation to generation, the tooling was of paramount importance. If you read Semiconductor International or EETimes online (or any of the technical journals), you'll see that the next decade of scaling is going to require new materials (high-k gates, low-k dielectrics, etc.) to keep Moore's law going. This opens the door to smaller firms getting a foothold in the market with a superior solution.
2. The growing revenues base in Asia. Companies want a presence there to better serve the markets, presence means training the population in tech. There are many advantages to overseas ops....cost, talent pool ( yes talent exist outside the US and there is a density of it in ASIA), customer awareness, ect.
Intel knows this, and has a very large presence in Asia.
3, Semicons can't afford new fabs and are Increasing the use of foundries for leading edge products. While you may think that AMD, Phillips, MOTO, whatever can't partner with a foundry and come up with a leading edge process that is state of the art and be right, it is silly to ignore the bottom line. Foundries are being trained to deliver the best and on the best products. The partnerships even if they fail, will succeed in the training aspect.
There I'm going to disagree. Foundry processes are good, but go to the "lowest common denominator" to provide the broadest possible customer base. High-performance chips are not going to come out of foundries.
4. Companies like Intel look at R&D from a business model view point (IMO), and sometimes suggest that others can't compete do to cost and margins. I don't think this holds true for a nation wanting to push forward into the future... they will support (gov) the tech growth with money.
Only so long as there is ROI, and right now the foundries are getting hammered.
5. As you build synergy in the foundries (they will now grow revenues at a greater rate) they will demand the attention of equipment makers. I think Intel has enjoyed being first to get certain litho tools ect in the past due to size and the possible accounts a salesman could land if they got Intel on board...what if Intel is not the biggest fish..or second, third?
That's a pretty big "if" considering it's Intel driving a lot of the litho technology at this point. I take issue with your attempting to evidence your point-of-view with a "what if" scenario.
6. Have you ever heard the term "not invented here"? enough said...I would explain in PM if needed.
... and exactly how is that relevant? Intel has made a living out of taking a basic toolset and increasing its performance through solid process development...
7. Standards...hmmm. Why did SEMATEC ever exist? I would hate to see PACTECH. Standards was a bad term to use. Out of SEMITECH came FOOP boxes and other 300mm standards. What I meant to drive to the point is that process leadership will be lost. Companies don't use SEMATECH to show process leadership. They might share a little knowledge at SEMATECH or use there facilities for small efforts but I bet most of the nitty gritty R&D at Intel has no clue what SEMATECH does or cares. This is what can be lost to ASIA, the fact you have lost talent to foundries since companies need to use them to develop new processes. It's the wafer data that's the learning's, who will control the wafer data in the partnerships?
I thought it was a FOUP...
What's your source for saying "most of the nitty gritty R&D at Intel has no idea what SEMATECH does or cares"??? Intel is a member of SEMATECH...
You start out with one or two reasonable points, but then start moving very far afield quickly. Also your one or two reasonable points fail to take into account certain strong trends in process development. |
