IBM Creates The Worlds First Commercially Viable 7nm Chip

While Intel and AMD are battling to get 14nm chips out on the market, IBM has announced it’s created the first fully-functional 7nm chip. IBM had originally announced its intentions to create the first transistor equipped 7nm chip last year, and now it’s here. It’s all part of IBM’s megabucks five-year, $3 billion plan to shrink the fabrication process ever smaller, having initially feared 7nm wouldn’t be possible.
As it stands, the 7nm chip IBM has created is the first next-next-generation chip, and the world’s first commercially viable sub-10nm product. IBM created it in partnership with Samsung and GlobalFoundries, using a combination of advanced technologies to break through the 10nm barrier.
The results of shrinking process technologies are more efficient, higher-density chips. You can fit more transistors on to the same amount of space basically, while also dropping power consumption. At the moment we’re seeing Intel preparing to push out its 14nm Broadwell chips, while AMD is still trucking on with Bulldozer-based 32nm chips until next year’s Zen micro-architecture.
While this is all well and good, in order to achieve Moore’s Law, which suggests the number of transistors should double every 18 months - two years, traditional methods weren’t going to cut it. Without branching out to other methods, pursuing such small dimensions results in escalating costs and current leakages that negate the benefits of dropping down in the first place.
IBM worked around this restrictions through a number of methods. Firstly, it worked on stacking its FinFET transistors closer together. They’re

In terms of the benefits, you’re looking at a 50% reduction in surface and a 50% boost in performance in comparison to a like-for-like 10nm chip, and the differences are going to be even more pronounced when compared to the upcoming 14nm fabrication process used for Intel’s Broadwell.
We know from yesterday’s news that Intel is struggling with its own die shrink, opting for the stop-gap 14nm Kaby Lake family next year, rather than the anticipated 10nm Cannonlake.
IBM's got us future-proofed for a little while yet then. Now it's time for the folks at Intel and AMD to play catch-up.