IBM has unveiled a new chip design that crams 100 billion transistors onto a surface the size of a fingernail, claiming it is the world's first known chip technology below 1 nanometre. The breakthrough, which the firm calls NanoStack, is the equivalent of around 0.7nm, compared with the current industry-standard size of about 2nm.
In tests, the prototype performed 50% better than IBM's own 2nm chip and was 70% more energy efficient. The company reported similar leaps in performance and energy efficiency when it debuted its 2nm chip technology in 2021.
“IBM claims world-first sub-1nm chip technology with 100 billion transistors on a fingernail-sized chip.”
Jay Gambetta, director of IBM Research and IBM Fellow, described the development as a "landmark moment" for the future of chips. "With our new NanoStack architecture, we're not just making smaller transistors, we're reinventing how chips are built to deliver dramatically more power and energy efficiency," he said.
Transistors are the building blocks of silicon chips, powering everything from smartphones to data centres that handle streaming, online banking and generative AI. The more transistors that can be squeezed onto a chip, the more powerful it becomes, and designers strive to make the chips themselves ever smaller.
For decades, the number of transistors per chip has doubled every two years – a phenomenon known as Moore's Law. But with billions already on some chips, sustaining this pace has grown harder, and experts broadly agree it cannot continue indefinitely. To try to extend it, chip designers have turned to 3D architectures that alter the shape of transistors to make them taller. IBM's approach is to layer sheets of transistors on top of each other.
Professor Alan Woodward, a computer scientist at Surrey University, compared the technique to building vertically in a city. "IBM's NanoStack is like proposing a 100-storey skyscraper," he said, adding that in his view, the firm's closest rivals are pursuing similar strategies.
Despite the advance, it will be several years before the chip technology is ready for production. The question remains whether the industry can keep shrinking chips fast enough to meet the demands of an AI-driven world.