IBM unveils a revolution: transistors smaller than 1 nanometer — a new frontier in microelectronics
IBM Corporation has achieved a breakthrough in the semiconductor industry by announcing a chip manufacturing technology with a transistor architecture of 0.7 nm, equivalent to 7 angstroms. This step marks a transition to a fundamentally new level of miniaturization, where traditional planar structures give way to multi-layered solutions.
The key innovation lies in the "nanostack" concept — transistors are placed not in a single plane but in multiple layers, allowing for a dramatic increase in the density of logic elements. In my assessment, this is not merely an evolution but a paradigm shift in chip design, where vertical scaling becomes the primary driver of performance.
According to data provided by the developers, the new approach will enable the placement of nearly 100 billion transistors on a chip the size of a fingernail. For comparison, this is several times higher than the density of current 3-nm and 5-nm solutions. As a result, the performance of such chips could increase by up to 50% compared to the 2-nm technology introduced by IBM in 2021, while energy efficiency could improve by an impressive 70%.
However, it is important to understand that the commercialization of the technology will take time. Experts predict that the first industrial samples will appear on the market within five years. This is due to the need to adapt production lines, develop new materials, and refine lithography methods. Nevertheless, the very achievement of the 0.7-nm barrier already sets the direction for the development of the entire semiconductor industry over the next decade.
My expert assessment: This development by IBM is not just a technological demonstrator but a clear signal to the market that Moore's Law continues to hold, albeit in a modified form. The transition to multi-layered transistors is inevitable, and the companies that master this approach first will gain a tremendous competitive advantage in the era of artificial intelligence and high-performance computing.