IBM Research revealed a significant breakthrough for the chip industry last week (and, by association, the whole world). After four years of development, the company unveiled the world’s first 2-nanometer (2nm) chip.
According to IBM Research, the technology can make chips 45 per cent faster or 75 per cent more energy-efficient than today’s leading choices.
- To put it in perspective, Dario Gil, IBM’s SVP and director of IBM Research, told us that it’s the equivalent of using a cutting-edge iPhone for four days on a single charge.
- Manufacturing is expected to start in late 2024 or, more likely, early 2025.
According to Gil, “the essence of this advancement is not about producing a chip, but rather the technology that will make all chips.”
How the IBM 2nm Chip Works
“Behind the scenes of everything we do,” says Gil, there are zeros and ones. Any language, such as English or math, can be converted to a string of zeros and ones. So, according to Gil, you must translate, or “map the problem you want to solve,” into those zeros and ones.
Then you’ll need a way to process those zeros and ones quickly so you can finish your job. You can do this by creating a transistor, a microscopic physical unit.
How it works: You add voltage and an electric field to the transistor’s “gate,” which does as it says: either allows or prevents electric current from flowing. That’s why, according to Gil, it’s referred to as a switch: “When current is flowing, we can call it a one, and when current is not flowing, we call it a zero.”
The bottom line is that the more transistors on a chip, the more zeros and ones you can process—and the more calculations you can do.
Small Device which does Big
Engineers would fit 50 billion transistors into a chip the size of a fingernail using the two-nanometer node. A single strand of human DNA is thinner than a transistor.
“It’s this tech and science marvel,” Gil explains. “Billions of transistors…all interconnected, processing the zeros and ones in the background—doing arithmetic for you, translating languages for you, and so on. So that’s the beauty of it: it’s right in the middle of the digital revolution.”
Significant advancements in graphics rendering, video quality and loading speeds, and payment processing are only a few examples of real-world applications. AI-driven scientific advances, such as molecule simulation, drug discovery, and battery technology, as well as accelerated language processing, are also possible.
AI is “extremely power-hungry,” according to Gil, who believes that as a result of the technology, we will see even larger language models emerge. On the user side, this may mean improved translation, voice recognition, and automated customer support, according to him. However, as scholars and ethicists have recently said, bigger is not necessarily better.
For IBM News and Latest IT Industry News
Follow us on Google News and Twitter
For Latest IT Industry News, Follow us on Google News and Twitter