Qualcomm and some other companies are leaving money on the table because they can't get enough 28-nm chips to meet demand out of foundry partner Taiwan Semiconductor Manufacturing Co. Ltd. Horror of horrors some customers are looking for ways to switch other companies' chips into slots previously designed around those vendors' ICs.
Could this be one of the ways Intel gets a leg up in the smartphone and tablet computer markets? Whether or not that happens to any significant degree, the current undersupply situation looks set to trigger refinements to some business models.
But that does not mean the immediate end of the foundry-fabless model that has served and driven the electronics industry for the last couple of decades. More likely it will stimulate a return to a strategy tried before: the funded joint-venture wafer fab. But this time with the additional wrinkle that process R&D will be front and center of the collaboration.
Those with sufficiently long memories will remember that Altera Corp., Analog Devices Inc. and Integrated Silicon Solutions Inc. each took stakes in a joint venture wafer fab back in the mid-1990s. That joint venture was controlled by TSMC and was used by the fabless participants to try and ensure their future supply of chips after a period of undersupply and allocation.
A $1.2 billion leading-edge 200-mm wafer fab was built in Camas, Washington, and became Wafertech. The fact that Wafertech is now a wholly-owned subsidiary of TSMC, illustrates that the story did not go particularly smoothly for all concerned. Altera committed about $140 million to gain an 18 percent share in the joint venture but by the time the plant was up and running the market was in oversupply again. In 1997 TSMC announced it would cut prices due to excessive capacity and fierce competition. So some might say Altera put its money down to no great effect on that occasion.
So why would Qualcomm, or indeed Altera, do the same again now?
For the last decade or so our Fabless Wonders have had a free ride on process R&D done at "credible" IDMs ( in my book after their HKMG fiasco IBM no longer qualifies ) siphoned out through various consortia and tool vendors.
They may now have $$$ on the table but they can't buy time. Intel started finFET work almost 8 years back. Hard to catch up even at 14 nm within 5 years even if the fabless wonders were now able to pry off enough process talent on behalf of their Foundries and went at it using the traditional chinese water torture approach to Physics ( a huge no. of experiments over a wide parameter space rather than targeting on narrow space with the benefit of device simulation ).
I think ultimately that Qualcomm and other chip companies would just as soon continue to outsource the whole enchilada to TSMC and others, rather than get in bed with TSMC in a JV. As you point out, Peter, owning a fab, or owning it in part, is not simply a matter of opening up a magic chip-building facility. There is a lot of work and know how that has to be done in terms of process development. All of that would have to come from TSMC or another foundry. In the end, it wouldn't be that much different from the situation that Qualcomm and others enjoy today--they pay money to foundries, the foundries provide the knowledge, capability and capacity to build their chips. Entering a JV would only mean that a fabless company was on the hook for the depreciation of the facility and equipment, which would be a liability as soon as oversupply conditions took hold. And oversupply is never that far off.
Not only that, but Qualcomm and its fabless brethen today enjoy the right to work with different foundries, play them off each other, and negotiate the best deal. Once they are a partner in the fab, that flexibility goes out the window.
Of course the advantage of such an arrangement, on the surface, would appear to be certainty of supply. Right about now, that sounds great. But as Qualcomm executives acknowledged the other day, at least part of the problem with the 28-nm capacity crunch is that demand for Qualcomm's 28-nm chips is simply greater than the company anticipated. If Qualcomm had been able to accurately predict what the demand would be, TSMC would have had the capacity available, I think.
I think that now is a good time to be an IDM. The fabless chip companies can't have it both ways. I agree with others that it is too late for them to catch Intel by the 14 nm node anyway. This problem of constrained capacity at the leading edge could go away in a couple of quarters between improving yields and more capacity coming on line. I don't think that now is the time to go back to a failed model - like committing to a JV if you are Qualcomm. I think there is more downside risk than upside potential there.
I wonder if circumstances are really as different today from what they were in the last joint venture backed fab episode?
Companies have been going fabless to begin with because *building* fabs is so enormously expensive that very few companies can afford to do it on their own, and it's likely a few that could simply see better returns on investment elsewhere. Joint ventures are fraught with inherent difficulties because of the issues of getting agreement from the partners on what is being done, what each is expected to put into it, and what each will get out of it, and *maintaining* that agreement as changes in the underlying economy require adjustments to the plans.
If I'm Qualcomm, I'm leaving money on the table because I can't get all the chips I want, but how *much* money is that? How much will it cost to build this joint venture fab? How long will it take me to recoup my investment? And what happens when the wheel turns as it always does and we move to a condition of oversupply again?
I may look at the numbers, and decide I'm better off *leaving* money on the table, because the amount of potential revenue I'm forgoing won't approach what it would cost me to build a fab, even with partners, to address capacity issues that are ultimately transitory.
I don't think that even Qualcomm is big enough to own JV fabs. It does not make any sense to own JV financially. Each generation of process development is multi-billion dollar effort PLUS it will cost JV $6-$7B to build just one 28nm fab. Qualcomm is running 28nm poly-Sion process not HKMG. So using HKMG is another story. For HKMG, TSMC and Common platform uses different gate stacks - gate last vs. gate first. This means Qualcomm has to port designs from one technology to the other along with all their IPs/design enablements. I don't believe they even have enough resources and money to maintain one generation of process and products with multiple technologies/foundries. How about 20nm? and 14-finfet? So this whole story simple does not make any sense.
There were a number of JV fabs built in 90's. UMC has USC, UICC, et al. Chartered had 3 fabs that were JV's with IDM/fabless companies. In my opinion, you picked the wrong one to use as an example. The Camus fab was built far from the other TSMC fabs and suffered some unique problems associated with that distance and culture. The UMC fabs were all built in Hsinchu and the Chartered fabs were all built in Singapore. There were pros and cons to the concept, but you are better served by looking at ones other than Camus for judging a potential return to the concept.
Interesting article. Intel does seem to hold quite a few important cards in their hand. I think one of the reasons that AMDs Athlon & Opteron products didn't do better is because Intel is a lot more than just design & archtecture. When their designs weren't as good as their competition, they were still to out-muscle them with faster transistors and lower die costs due to their powerful manufacturing division. Intel held about a process node advantage over AMD during all of that time and even when AMD was architecture-competitive, they were still beaten by process technology & Intel's lower costs. And now it's history. AMD is w/o a fab.
I expect Intel to do a repeat of this bit of history with the ultra mobiles. They have even more of a commanding lead in process technology than a few years ago and now have a competitive smartphone. They'll quickly move it to 22nm and to 14nm within a year after that. They just forecasted that ~25% of their mainstream CPU product shipments in Q2 with be 22nm. This is a huge volume.
It would seem that their 22nm process is pretty healthy. Die supply for the very much smaller Medfield chips and its successor will never be an issue. Intel has been hitting on all cyclinders for several years & I expect them to start pulling ahead by the end of 2012 and extending that lead in 2013 with 14nm.
Well one of the points that i tried to tease out in this analysis/opinion piece is that Qualcomm and others may need to enter into a Crolles-style joint venture that is as much about process development as it is about manufacturing capacity. OR watch an IDM continue to eat their lunch even when there is a foundry oversupply situation.
Qualcomm is much more likely to use their money to help the other foundries catch up to TSMC than to build their own fab. There mistake was waiting so long to start. They appear to be spreading it around now, but it will take awhile. 28nm may be old news by then. It might make more sense to help one of more fabs pull in the jump to 14nm with FinFets than to pour the money into 28nm capacity.
This article shows how important it is to have your own manufacturing. Besides the points raised in this article, there is the additional risk of foundries creating their own products directly or indirectly (Samsung anyone) in which case the fabless companies would have funded their competition.
I agree with Peter here. What Qualcomm and Co. need is a long term view. Intel had a process technology advantage because of the economies of scale of its x86 processor business. That is no longer the case (the PC market is no longer the cash cow it used to be and x86 in mobile does not make much sense IMO). So if Qualcomm and Co. invest heavily in R&D in the form of a JV fab, ultimately they will bridge the gap and surpass Intel in leading process technology.