IBM unveils world’s first sub-1nm transistor architecture

IBM unveils world’s first sub-1nm transistor architecture


IBM has announced a major semiconductor breakthrough, unveiling the world’s first sub-1nm chip technology. Nominated as the 0.7nm (7-angstrom) generation, the milestone comes as global technology firms aggressively compete to engineer silicon capable of processing heavy, next-generation AI workloads. Following the announcement, shares of the Armonk, New York-based giant jumped over 6% in pre-market trading, rebounding slightly from an 11% decline so far this year.

Read: Samsung Galaxy Buds 4 Review: Brilliant brains, flawed physics

The development signals a massive leap forward at a time when traditional chipmakers are hitting physical boundaries in their quest to maintain Moore’s Law, the decades-long trend of squeezing more computing power into increasingly confined spaces.

To breach the sub-1nm threshold, IBM’s research division moved away from traditional horizontal layouts to pioneer a proprietary three-dimensional transistor design called “Nanostack.”

Instead of laying transistors flat across the silicon substrate, the Nanostack architecture stacks them vertically in three dimensions. This structural shift allows IBM to fit exponentially more computing components into the exact same volume of space.

“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,” explained Jay Gambetta, Director of IBM Research.

IBM projects that the 0.7nm node will pack nearly 100 billion transistors onto a surface roughly the size of a fingernail. This represents twice the transistor density of IBM’s own 2nm test chip unveiled in 2021.

According to laboratory metrics, the node is capable of delivering two distinct operational advantages depending on how the silicon is optimized:

See also

  • Performance Mode: Offers up to a 50% increase in processing speeds over previous generation standards.
  • Efficiency Mode: Yields up to a 70% reduction in energy consumption, a critical metric for power-hungry modern data centres.

This breakthrough significantly bolsters IBM’s position as a foundational research house competing against foundry giants TSMC and Intel. The announcement follows closely on the heels of Intel moving its 1.8nm “18A” manufacturing process into risk production—the final testing phase required before commercial fabrication begins.

IBM anticipates that commercial production for its 0.7nm chips could begin within the next five years. While the company has historically licensed its high-end chip blueprints to global manufacturers like Samsung and Japan’s Rapidus to bring its designs to market, it has not yet named an official manufacturing partner for this specific 7-angstrom technology.