There is only so much you can cram inside a smartphone without sacrificing other things like phone size, internal space, thermal emission, and battery life. One solution is to shrink those components while maintaining or even improving their performance and efficiency. That is often the case with semiconductors, particularly the processors that power phones, laptops, and computers, and IBM’s first 2nm chip, also the world’s first 2nm chip, promises to do exactly that and then some.
Most of the processors that power high-end smartphones and devices today utilize 5 or 7 nm FinFET processes. To put it into perspective, a 5nm chip crams less than a million transistors per square millimeter while this proof-of-concept 2nm chip can hold more than 300 million. According to IBM, that results in a 45% improvement in performance.
With more transistors, of course, comes more processing power but also more power consumption. That isn’t the case here, however, as IBM advertises an even lower 75% energy use compared to current chips in use. In practice, that could make it possible to have smartphones that will take days, not hours, before needing to be recharged.
It’s not just about smartphone battery life, of course. IBM tries to envision a world where data centers will produce less carbon footprint and self-driving cars can detect objects faster and, consequently, avoid crashing into them. The smaller chips will also be a boon to the IoT industry that has always been constrained when it comes to available space but also need sufficient processing power to offer smart features.
That said, it’s too early to get excited over IBM’s 2nm processors, enticing as they may sound. It will probably take years before actual 2nm chips become commercially available, especially considering the global silicon shortage we’re facing now.