IBM invests $3 billion into post silicon computer technology and 7nm chip research

IBM has proclaimed that it’s tilling $3 billion into 2 R&D programs that may hopefully create it the authority on 7-nanometer-and-beyond chip technologies. One R&D project can consider pushing standard Si chips as way as they're about to go around 7nm, and are searching for ways to take them even further in quantum computing, carbon nanotubes, grapheme and III-V. IBM additionally took the chance to inform everybody that it’s already the largest player in 7nm-and-beyond technology, with over five hundred applicable patents (more than double the closest competitor).whereas most major chip manufacturers (Intel, TSMC, GloFo, IBM) appear assured that they will take commonplace Si CMOS chips right down to 10nm, they're somewhat bit nervous concerning the prospect of 7nm and on the far side. At around 7nm, the present building blocks of Si transistors simply won’t behave within the same way and once the gate is just a couple of atoms across then classical physics goes out of the door and physics takes over. However, totally different semiconductor unit styles (such as 3D) enable U.S. to require Si somewhat additional, the laws of physics can eventually catch up.
The IBM analysis 3D chip stack concept’s alternate approaches like this, instead of brute-forcing smaller transistors, that square measure the possible way forward for pc.
To reach 7nm and on the far side, IBM analysis is taking a pincer approach. First, it'll take serious effort to truly get Si right down to 7nm and to significantly, develop processes that may create 14nm, 10nm, and 7nm chips economically.  

However, the second approach is without a doubt quite exciting, instead of pushing Si indefinitely, this second research can consider different materials and techniques which may simply take us to 7nm and on the far side. Materials like III-V semiconductors notably metallic element chemical compound, GaAs have around ten times the lepton quality of Si, providing smaller transistors with a lot of higher performance and lower power consumption. Likewise, IBM is trying into graphene and carbon nanotubes, each of that have unbelievably high lepton quality and may (theoretically) be designed into terribly little structures. IBM has already created a fullerene semiconductor unit with a 10nm channel that showed no sign of performance degradation thanks to its diminutive size (CNTs square measure simply single layers of carbon/graphene rolled up into a tube, and are thus extremely small).

But on the far side higher materials and therefore the ever-shrinking semiconductor unit, IBM is additionally trying into totally different ways of computation entirely, like neuromorphic computing (brain-like chips), quantum computing, and Si photonics/optoelectronics (optical school designed into electronic chips). These ways won’t essentially offer additional gigacycle per second or consume less power, however they might supply wildly additional capable pcs that may method way more knowledge than a traditional computer within the same quantity of your time. Finally, it’s necessary to notice that IBM is already acting on all of those technologies, and in several cases has been for years. The announcement of an extra $3 billion in expenditure is exciting, however IBM already spends roughly $6 billion p.a. on R&D, and corporations like Intel, Samsung, HP, Microsoft, and Google  all around that mark.