iSuppli predicts 20nm is where we get off Moore's curve

Michael Feldman at HPCwire posted a blog item yesterday following an announcement by consulting firm iSuppli that reckons that below 20nm chip foundries will become too expensive to ever recoup the capital investment needed to build them. Ever.

The price tag on a new 45nm fab is over a billion dollars today. AMD’s new foundry partner, Globalfoundries, is constructing a 32nm fab in New York with a budget of $4.2 billion, and Intel has already committed $7 billion to upgrade its fabs to produce 32nm chips. You have to sell a lot of chips to recoup those kinds of costs. And those are just capital expenditures.
In the iSuppli announcement, Len Jelinek, the firm’s director and chief analyst for semiconductor manufacturing, explained it thusly:

“The usable limit for semiconductor process technology will be reached when chip process geometries shrink to be smaller than 20 nanometers (nm), to 18nm nodes. At those nodes, the industry will start getting to the point where semiconductor manufacturing tools are too expensive to depreciate with volume production, i.e., their costs will be so high, that the value of their lifetime productivity can never justify it.”

Michael points out that, if it happens, this could shake up our chip monoculture, giving rise to new architectures at both the chip and system level that can deliver the competitive advantage that currently arises from shrinking processes and the relentless tick tock of Intel’s processor cadence.

Of course, this isn’t the first time Moore’s law has been declared in poor health

Of course, none of this may come to pass. Moore’s Law is periodically declared dead and has thus far defied its doomsayers. Additional transistor density may be achieved in other ways, such as 3D semiconductor structures. And there’s no shortage of more exotic approaches like carbon nanotubes, silicon nanowires, molecular crossbars, and spintronics. In any case, whatever happens in 2014, we’re bound to be living in interesting times.

Comments

  1. Ignoring economic effects, there is a hard physical limit, which changes us from a “classical” or “semi-classical” physics regime, into a hard “quantum” regime. This is when a length scale of the semiconductor (width, depth, length, spacing) becomes comparable to the thermal de Broglie wavelength of a charge carrier (for an electron at room temperature, this is about 4.3nm or 43 Angstrom. Then our device physics will be significantly different than it is today. Different doesn’t mean worse, there may be some very nice new devices we can build from this (a quantum ALU anyone?) where we leverage and embrace the quantum nature of the system. There is quite a bit of work going on in quantum computing. It would be interesting to see if it can be made practical.

    Note that we are at 45nm devices now, about 10x the de Broglie wavelength. 20nm devices are about 5x, so you should see a serious onset of the quantum regime happen below this. The charge carrier waves tend to interact with an exponential tail that looks something like exp(-1 x size / constant) where the size is the structure size or spacing, and the constant is a multiplicative factor times that de Broglie wavelength. If you get a bunch of “square” wires close enough together, you will get basically a Kronig-Penny model, which is a homework problem for many an undergraduate physics student in their quantum mechanics class. We are getting nearer to this.

  2. Don’t worry, there will likely be a completely different kind of technology that beats the curve. When high technology was electrically actuated switches, they could never have imagined electronic components, let alone integrated circuits or microchips probably smaller than the heads of the bolts that held early computers together. Physicists pontificated the speed of sound was impossible to break, then Chuck Yeager broke it in the X-1 rocketship in 1947. Little did they know that only a few years later they would test and fly a true airplane, the SR-71 that cruised faster than a rifle bullet!

Resource Links: