IBM’s $3 Billion Investment Into 7nm Chip Research
IBM have now announced that they are now investing over £3 billion into two Research & Development (R&D) that will push it to be the leading company in the world upon the 7 nanometer (nm) and smaller chip technology.
The first R&D project is scaling already developed chips as far as they will go attempting to reach the 7nm, pushing those to new extremes of processing power.
The second is more R&D investigating new materials and innovating chip functionality utilising these materials. They want to take chips to a new era of design of processing power, looking into such subjects as quantum computing, graphene development and carbon nano-tubes.
IBM also confidently remained the media that it is already one of the big players within this sector with 500+ applicable patents to this new technology. (That is more than double than its closest competitor)
However the other major chip designers (Intel,GloFo,TSMC) are confident at taking the chips down to 10nm but wary at the prospect of the 7nm CMOS chip.
There is reason to be concerned here as when you start going down to 7nm silicon transistors just do not behave in the same way, with the gate only being a few atoms across only quantum physics can be applied here as classical no longer applies. This can be a problem as quantum physics isn’t a strait forward subject as most physicists will know.
However there are also 3D chips which utilise 3D transistor technology out there filling the gap in chip development, this a great substitute until physics just catches up to the quantum silicon chip development.
Some believe that 3D chips may be the way forward as we already can develop these they suggest staying at a level we already know how to utilise and push the design.
IBM firmly believe that they can hit this subject hard and produce something altogether new to dominate the market of chip design and transistor size. However the problem here isn’t only getting the size down as this is possible, it’s getting the machines that can do this economically as they for even business as big as IBM they are still extremely expensive.
IBM wants to develop new machines and techniques that will help silicon scale down to 7nm, and potentially beyond.
However instead of just pushing silicon to new boundaries IBM are also investing in the research of new materials and techniques which could be used instead to get chip development down to the 7nm and beyond.
These could be III-V semiconductors (notably gallium arsenide, GaAs) which give about 10 times more mobility to the electrons than silicon, this could intern lead to transistors which have much higher performance but lower power consumption.
They are also looking into the architecture of materials too, such as graphene and carbon nanotubes which both could be designed with new architectures to increase electron mobility.
On-top of this IBM have already created the Carbon Nanotube Transistor (CNT) which are single layers of carbon or graphene rolled up into a tube. With transistors such as these of around 10nm channels that showed no problem in efficiency even at this scale.
Adding to this IBM are also looking into new methods of computation entirely too, with neuromorphic chips (emulating the function neurons within the human brain), optoelectronics (using paths of light inside of chips), silicon photonics and quantum computing.
IBM however already spends about $6 billion per year on R&D upon this technology with Intel, Microsoft, Samsung, HP and Google pushing about the same amount of money there too. So this announcement appears to be more of a message to customers and stakeholders that IBM still wants to have the leading edge in computing technologies.