The new chips are among the first to be made with 7 nanometer transistors, an advancement made possible by using a silicon-germanium mixture instead of 100% silicon. The benefit here? As transistors get smaller, you can pack more of them onto a single chip, greatly improving chips’ performance. Today’s best chips have 14 nanometer transistors, while the next generation is expected to whittle that size down to 10nm.
Basically, smaller transistors = faster chips, and that’s a good thing.
IBM’s new chips were made specifically to prove the jump to 7nm is possible, using a complex manufacturing process that won’t be easy to scale up. But by establishing that 7nm is doable, it clears a mental roadblock to getting commercial-grade chips to that point. Now it’s up to chip manufactures like Intel to figure out if it’s possible to make 7nm chips in a commercially viable manner.
Computer chips have for decades been governed by something called Moore’s Law. Moore’s Law states the number of transistors per square inch on an integrated circuit should double every two years, but it basically means “stuff gets faster.” Because the boost in chip performance has been fairly predictable, computer manufacturers and other innovators are able to plan in advance for the arrival of faster and more efficient machines.
processor design is starting to bump up against some pretty firm rules of atomic physics. What IBM’s new chip proves is that Moore’s Law should hold true for at least a little while longer, adding a little more certainly to computing’s future.