Technology - more powerful and energy-efficient chips

The American giant IBM has presented a new semiconductor technology that could form the basis of a generation of chips with double the performance and significantly reduced energy consumption compared to current processors. According to the company, large-scale production of the new chips could become possible in the next five years, informs[ AFP.

The new technology aims to create chips of "0.7 nanometers" (nm), a name used in the industry to indicate the level of miniaturization and density of transistors, without representing the exact physical size of the components. IBM had previously presented the 2 nm technology in 2021, but mass production did not begin until the end of 2025. The new manufacturing process promises a significant leap in performance. According to the company, the 0.7nm architecture allows the integration of nearly 100 billion transistors on a chip the size of a fingernail, doubling the density offered by 2nm technology.

• Higher performance, lower consumption

IBM representatives claim that the new chips could offer approximately 50% more processing power compared to the 2nm generation. At the same time, they could perform 1.7 times more computing operations with the same energy consumption, an essential aspect in the context of the accelerated development of artificial intelligence and the increasing demand for computing power in data centers.

• A new three-dimensional architecture

To achieve these performances, IBM researchers developed a three-dimensional architecture called "nanostack” (nano stacking). Unlike conventional chips, which use a single layer of transistors, the new technology allows multiple layers to be stacked, thus increasing the density and efficiency of electronic components. According to Huiming Bu, vice president of Semiconductor Research and Development, this architecture could support the further miniaturization process to the level of 0.1 nanometers (one angstrom) around 2040.

• Benefits for artificial intelligence

The new technology will not be limited to classic processors, but can also be used for GPUs, the essential components in the development and training of artificial intelligence systems. In addition, IBM says that the architecture will also bring significant improvements to SRAM memory, where performance could increase by about 40%. "This is an advance that we have not seen in decades," said Jay Gambetta.

• IBM does not mass-produce chips

Although it has carried out research and development of the technology, IBM does not manufacture processors on an industrial scale. Its business model is based on licensing innovations to large semiconductor manufacturers. Currently, 2nm technology is used by TSMC, the world's largest chipmaker, and Japanese company Rapidus plans to start mass production in the second half of 2027. Meanwhile, TSMC has already announced plans to switch to 1.4nm technology in 2028, which shows the extremely fast pace of global competition to develop next-generation semiconductors.

• A strategic technology race

IBM's announcement highlights the intensifying competition in the semiconductor industry, at a time when demand for high-performance chips is fueled by the expansion of artificial intelligence, data centers and connected devices. If 0.7nm technology can be transformed into a viable industrial process, it could represent one of the most important technological leaps of the next decade, paving the way for faster, more energy-efficient and more powerful computing systems than those available today.