Take a look at this graph – it’s obvious, isn’t it? FD-SOI is significantly cheaper, outdoes planar bulk and matches bulk FinFET in the performance/power ratio, and keeps the industry on track with Moore’s Law.
This was part of a presentation by ST’s Joël Hartmann (EVP of Manufacturing and Process R&D, Embedded Processing Solutions) during Semi’s recent ISS Europe Symposium.
If you’re a designer, and you want faster-cheaper-cooler, you’ll go with FD-SOI, right? You can leverage existing designs, it’s an easy port of IP, and you get terrific power-performance benefits from back biasing.
Which leads to the question: is FD-SOI really what’s best for the industry, too? From the designers’ and fabless perspective, it’s a clear “yes!”
Of course, the pure-play foundries clearly need the volumes to make FD-SOI worth their while. We all remember last year when this GloFo slide made the rounds (it was shown at the SOI Consortium workshop in Kyoto, Japan last summer – see the full ASN article here):
When I spoke to Subi Kengeri, GloFo’s VP Advanced Technology Architecture at a developers event this past fall, he said that GloFo was enthusiastic about the prospects for FD-SOI. But the foundry giant was still considering its options, while comparing FD-SOI to (bulk) FinFETs. While FD-SOI comes out a winner when leakage is the primary issue, he said, GloFo potentially sees a bigger opportunity in FinFETs.
This was last fall, remember, and he said that some customers had expressed reservations about the difficulty of back-biasing in FD-SOI. Now surely with all we’ve heard about the amazing things you can do with back-biasing in FD-SOI (click here if you need a refresher), nobody really thinks it’s hard anymore, do they?
So why should the fabless world swallow a major and very costly veer off the trend line of Moore’s Law just to do FinFETs at this point?
Subi told me, “We are a foundry and are here to serve our fabless customers.” He said if there is serious high-volume customer interest in FD-SOI, they can make it happen.
ST’s now indicated that a big foundry announcement will be made shortly, which is very exciting. (Read about that here if you missed it before.)
So, the foundries can make it happen. And they will. Because maintaining cost-efficiencies is what makes this industry tick. Ultimately, the best interests of the industry will always prevail. Just ask!