Deepak Sekar, Chief Scientist of MonolithIC 3D, has just published a really interesting blog post with a somewhat cryptic title, in my opinion. The title of the post is “Can 450mm Decommoditize the Semiconductor Industry?” The post discusses how the cost of jumping from 300mm to 450mm silicon wafers might indeed separate the big semiconductor fab players from the not-so-big. Sekar makes a compelling argument that the cost of upgrading an entire fab line from 300mm to 450mm is significantly more than the roughly $2 billion he estimates is needed to jump from one IC process node to the next node. The benefit realized from making this jump is a 20% to 25% savings in per-die manufacturing cost, so there’s a real competitive advantage to making such a move, if a company can afford it.
And that is Sekar’s premise: companies that can afford to make this jump will ultimately win.
Sekar’s subsequent analysis is based on a “rule of thumb” that says a company can only afford a fab upgrade if it has annual revenues that are twice the cost of the upgrade, which he pegs at $7 billion for a 450mm fab. He then compares DRAM vendors against DRAM vendors, NAND Flash vendors against NAND Flash vendors, and logic foundries against logic foundries to determine which might be the winners in such a conversion.
Very, very interesting reading. Click here for Sekar’s full post.