BM has demonstrated a nine nanometer (9nm) carbon nanotube transistor (CNT) — the smallest CNT ever made, and significantly smaller than any commercial silicon transistor. At 9nm, IBM’s transistor is also smaller than the physical limit of silicon transistors, which is around 11nm.

Beyond its diminutive size, the 9nm CNT is capable of switching at very low voltages (0.5V), thus consuming less power than its silicon counterparts — but it can also carry four times as much current, meaning a better signal quality and a wider range of applications.

Carbon nanotubes, much like graphene, have long been heralded as the eventual replacement for silicon transistors due to their improved electrical qualities. There are (obviously) problems for the adoption of CNTs, though: They’re hard to mass produce, and they also have to reach a maturity level that can unseat a semiconductor technology that has ruled supreme for more than 40 years. It’s not that Intel & Co. don’t want to use carbon nanotubes, but when you’re churning out billions of dollars worth of silicon chips there’s an awful lot of inertia preventing a sideways leap to a new technology.

It is this inertia that resulted in Intel’s 3D FinFET chips — a last gasp effort to squeeze a few more years out of silicon semiconductors. The question is, does Intel also have a working CNT, or does IBM now have the upper hand? Intel’s projected roadmap has an 11nm node in 2015 — but what about after that? It’s important to remember that IBM has one of the most advanced semiconductor processes in the world, along the same lines as TSMC and GloFo. If IBM is the first to market with carbon nanotube transistors, Intel might finally have a challenger.