DEEPAK Sekar, the Chief Scientist of MonolithIC 3D, has an insightful blog post about the move from 300mm to 450mm wafers.
IBM, Intel, TSMC, Samsung and Globalfoundries are jointly contributing to a 450mm wafer R&D in New York. According to Sekar, the transition to 450mm is not just about the 20 to 25 percent cost savings on 300mm wafer technology.
Sekar says upgrading feature size (moving from 45nm to 32nm to 22nm, for example) only requires the replacement of a small percentage of tools, thus reducing the amount that needs to be spent in those upgrades. The average is US$2 billion.
This diminished cost for scaling situation applies for both logic fabrication facilities and NAND flash memory manufacturers (who move from 3x nm to 2x nm then to 1x nm).
However, Sekar points out that moving to a different wafer size (such as what the coalition above are doing) requires all equipment in the fab to be upgraded, with the average cost being US$7 billion.
“A rule of thumb people use is that your revenue must be at least 2x greater than your fab cost to be able to afford it. To pay for a $7B 450mm fab, your revenue must therefore be at least $14B,” Sekar wrote.
The sheer cost needed to upgrade facilities to a new wafer size (in order to catch up with other players) effectively creates an entry barrier that eliminates smaller competitors.
If the smaller companies cannot make the move, they will continue to operate, but at a 20 to 25 percent cost disadvantage to the big players who have upgraded. They will also see decreased support from equipment makers. Thus, smaller manufacturers may ask for government funding to upgrade, or form alliances for the upgrade.
Sekar found that in the DRAM market, only Samsung can afford the 450mm transition. In the NANDEEPAK Sekar, the Chief Scientist of MonolithIC 3D, has an insightful blog post about the move from 300mm to 450mm wafers.
IBM, Intel, TSMC, Samsung and Globalfoundries are jointly contributing to a 450mm wafer R&D in New York. According to Sekar, the transition to 450mm is not just about the 20 to 25 percent cost savings on 300mm wafer technology.
Sekar says upgrading feature size (moving from 45nm to 32nm to 22nm, for example) only requires the replacement of a small percentage of tools, thus reducing the amount that needs to be spent in those upgrades. The average is US$2 billion.
This diminished cost for scaling situation applies for both logic fabrication facilities and NAND flash memory manufacturers (who move from 3x nm to 2x nm then to 1x nm).
However, Sekar points out that moving to a different wafer size (such as what the coalition above are doing) requires all equipment in the fab to be upgraded, with the average cost being US$7 billion.
“A rule of thumb people use is that your revenue must be at least 2x greater than your fab cost to be able to afford it. To pay for a $7B 450mm fab, your revenue must therefore be at least $14B,” Sekar wrote.
The sheer cost needed to upgrade facilities to a new wafer size (in order to catch up with other players) effectively creates an entry barrier that eliminates smaller competitors.
If the smaller companies cannot make the move, they will continue to operate, but at a 20 to 25 percent cost disadvantage to the big players who have upgraded. They will also see decreased support from equipment makers. Thus, smaller manufacturers may ask for government funding to upgrade, or form alliances for the upgrade.
Sekar found that in the DRAM market, only Samsung can afford the 450mm transition. In the NAND flash memory market, Samsung and Toshiba/SanDisk can viably upgrade wafer sizes.
“In the foundry business, only TSMC and possibly Samsung can afford a 450mm fab... Globalfoundries, as we know, is lucky enough to be sponsored by the Government of Abu Dhabi, who could potentially get them a 450mm fab,” Sekar wrote.
“The other foundry players, who make up 36% of today's foundry market, may be unable to afford a 450mm fab.”
Sekar says in the new market, “semiconductor products would no longer be commodities,” with its relevant implications for the rest of the electronics industry.
For the full post, click here http://www.monolithic3d.com/2/post/2011/10/can-450mm-decommoditize-the-semiconductor-industry.htmlD flash memory market, Samsung and Toshiba/SanDisk can viably upgrade wafer sizes.
“In the foundry business, only TSMC and possibly Samsung can afford a 450mm fab... Globalfoundries, as we know, is lucky enough to be sponsored by the Government of Abu Dhabi, who could potentially get them a 450mm fab,” Sekar wrote.
“The other foundry players, who make up 36% of today's foundry market, may be unable to afford a 450mm fab.”
Sekar says in the new market, “semiconductor products would no longer be commodities,” with its relevant implications for the rest of the electronics industry.
For the full post, click here.