Moore’s Law, coined in 1965 by Intel’s co-founder Gordon Moore, predicts that the number of transistors on a microchip will double about every two years, while the cost continues to go down. That trend was true for decades, but lately semiconductor industry insiders believed Moore’s Law was all but dead. IBM’s latest invention, the 2-nanometer chip, suggests the principle is still alive and well.
Semiconductor chips hold the data needed for us to use our computers, cellphones, appliances, cameras, and cars. As the pandemic spurred a massive transition to remote work, increasing our reliance on computers, it fueled even more of a necessity for chips—and also helped precipitate a global chip shortage. “People take it for granted,” says Mukesh Khare, vice president of IBM Research. “But everything runs on semiconductors. Chips are the backbone of everything that we’re doing in modern technology.”
Transistors have shrunk over the decades, from an initial 10,000 nanometers (nm) in 1971 to 5nm in 2020. The 2nm transistors on IBM’s new chip, which is essentially a circuit of connected transistors, are much smaller than the eye can detect. The width of a human hair is 100,000nm; a red blood cell about 7,000nm; a strand of DNA about 2.5nm. Size matters: The smaller the transistors, the more fit on a chip, improving efficiency. “Every time you make things smaller, you can do more,” Khare says. IBM’s 2-nanometer chip is the winner of the enduring impact (15+ years in business) category of Fast Company’s 2022 World Changing Ideas Awards.
The 2nm chip will likely raise performance across all the gadgets that use them—like our cellphones and Microsoft Xboxes—while also making them smaller and allowing for impressive new features. IBM forecasts an average increase of 45% in product performance, which could allow us to charge cellphones once every four days rather than daily, and autonomous cars to make smarter and quicker decisions. All the while, these electronics should be cheaper, because smaller means less expensive to produce.
Getting the transistors so small is not easy; it’s been the work of a decade and counting. “People say, ‘it’s not rocket science,'” Khare says. “It’s actually much more difficult than rocket science.” On a single silicon chip the size of a fingernail, there are 50 billion transistors, each the size of about five atoms. They’re stacked on top of each other, in the same way that building skyscrapers makes the most of limited ground area. “You start to stack things on top of each other, so that you can get more efficiency in the same real estate,” Khare says. Those transistors then work together in a circuit to make our favorite devices tick.
The tech industry is responsible for almost 2% of global carbon emissions, nearly as much as aviation—a share that could skyrocket to 20% in the next 12-or-so years. Another benefit of the 2-nanometer chip is that this invention could potentially create a 75% reduction in energy consumption, according to IBM, because powering a smaller device requires less energy. IBM predicts that, if every data center in the world switched to the 2nm chips, while maintaining the same capacities, the energy savings could power more than 43 million homes for a year.
In the context of the semiconductor shortage, the invention is welcome news for the industry. Before the pandemic, with the stamina of Moore’s Law questioned, investment was dwindling, and manufacturing jobs were moving overseas. To bring its blueprint into fruition, IBM is now working with manufacturing partners, both in the U.S. and abroad, to begin mass production of the semiconductor in 2024. Khare says this secures a clear roadmap for scaling semiconductor tech, while potentially lowering carbon footprints, for at least the next 10 years.
After that, how much smaller can they get? Khare is optimistic, but says the main challenge will be sustainability. “Is there enough power in the world to be able to fuel all these chips?” he asks. “If we can keep power in check, I think we can keep going for a long time.”