Foundries fall at cutting edge
26 January 2011
There are predictions that only three ‘leading-edge’ foundries will remain by the end of 2011. Is this healthy?
The prediction comes from HIS iSuppli, who believe that by the end of this year there may be only three remaining sources for high-volume, leading-edge semiconductor foundry production: Taiwan Semiconductor Manufacturing Co. (TSMC), GlobalFoundries and Samsung Electronics.
By ‘leading-edge’ iSuppli means CMOS processes at the 22 and 20 nm nodes, and another couple of foundries join the party at the 28 and 32 nm level. HIS iSuppli also believes that this is the year that Intel may decide to use some of its advanced manufacturing capacity and offer foundry services to design companies and fabless semiconductor suppliers that are incorporating the Atom microprocessor into their designs. Although such a move would represent a dramatic change of philosophy at Intel, say HIS iSuppli, it potentially could lead to significant revenue growth as well as more favourable asset utilisation.
“The enormous cost of advanced semiconductor process technology is whittling down the ranks of leading-edge foundries, with just three firms likely to remain at the end of the year,” said Len Jelinek, Director and Chief Analyst of semiconductor manufacturing at IHS. “Unless additional foundries join the party, semiconductor companies will face minimal competitive choices when it comes to advanced chip geometries.”
This last point is the interesting one. Given the vast fluctuations in market conditions over the past two years, putting this segment of the market into the hands of so few companies cannot be good for stability. Each of these companies will naturally have their own philosophies regarding overall capacity, customer profiles, scheduling and customer priorities. But will these philosophies suit every customer, every designer and every fabless company? Of course not.
The lead times we have seen through 2010 have in extreme cases been as long as the year itself. Having so little choice at the top end would only discourage both designers of these components, or those designing with them, of steering clear of anything that was not already established or anticipated to be in very large volumes.
Any market, whether it be silicon or sandwiches, has a habit of balancing the supply and demand requirements, and it would be nice to think that the door was open for some cutting-edge/medium volume players to enter the market. Unfortunately, with vast sums of money required to set-up a new foundry, or process within an existing foundry, it could be that the electronics industry is being left short.
Contact Details and Archive...
Most Viewed Articles...