This is the last time a graph like this will appear for a while. Because Chartered Semiconductor Manufacturing is now part of Globalfoundries there won’t be an opportunity to get information on the processes the company is running from financial reports. AMD will only report profit or loss in its figures now that the company has switched to equity accounting even though AMD holds the lion’s share of the key stock class that determines overall ownership.
Now that the results are in for 2009, it’s possible to see what effect the semiconductor industry’s bungee recession (thanks to Future Horizons’ Malcolm Penn for the inspiration for the phrase) has had on the shift towards more advanced processes. What’s interesting about the first set of charts is that you’d hardly know the foundries practically turned their machines off for a couple of quarters.
Utilisation plunged to 30 per cent in the dark days of early 2009 before bouncing back to near capacity by the middle of the year. Had it not been for TSMC’s 40nm yield problems, the transition towards 40/45nm processes might have been a bit quicker. But the severity of the recession arguably gave the number-one foundry a bit of breathing space, arguably helped by the better yield on more expensive flip-chip packages that the 40G-process chips would typically go into.
Chartered’s small spike in 45nm orders at the start of 2009 is not just a fluke of the graph, coming at a time when revenues had plummeted. Perhaps helped by delays at TSMC, the Singapore foundry saw about $8m worth of 45nm orders go out the door in the first quarter. These then fell back towards the autumn.
It’s not clear what happened in the last quarter — Chartered did not report 4Q09 figures because it had come under the wing of Globalfoundries by the end of December — but the trend for 45nm was in the wrong direction ahead of that. However, 65nm production was still growing reasonably well and still exceeded that of larger company UMC as recently as 3Q09.
Although UMC has been able to ramp 65nm production quickly in the past few quarters, it’s possible that the number-two foundry did not catch Chartered by the end of the year. But it’s now hard to tell and the focus at Globalfoundries will be on developing post-65nm processes.
In the second set of graphs, you can see how foundries have been affected revenue-wise by the late-2008/early-2009 recession. However, despite the economic and the yield problems at TSMC, it seems that 45nm-class processes are more or less tracking the ramp of 65nm two years earlier. At this scale, to match the other graphs, it’s more or less impossible to see the Chartered contribution to 45nm. It is in there but hard to see without a little magnification.
The question will be how big the foundries’ slice of the 45nm pie will grow. So far, each process has topped out in the $1.2bn range even if you aggregate the results from the big four.
In the past year, foundry sales were knocked badly by the decision by integrated device manufacturers (IDMs) to pull back from using external fabs. Their share is rising again but this is a volatile set of customers for the foundries as they have a greater range of choices fabless companies do not. As more IDMs go fab-lite, clearly they too run out of choice. But this is a market that often rewards the biggest players. To even play in some markets, chipmakers have to be able to sell hundreds of millions of dollars’ worth of product to be considered contenders. And if you are that big, owning your own fab is still an option.
It seems reasonable to assume that the peak revenue from 40nm will get much closer to $1.5bn, and possibly surpass it. But with the market on a two-year process-upgrade treadmill and high-k, metal-gate versions of 32nm offering some performance advantages — at a cost — there is still the risk that the companies who bet on 65nm will wait for 32nm and 28nm to roll along before making the jump. If companies play this waiting game, it will slightly favour Globalfoundries — all other things being equal — as it tries to build up capacity to service a wider range of customers. But for those who do pick 45nm-generation processes for the next year or two, TSMC has built up a similar early volume lead to the one it had with 65nm.