Thanks to this technique, AI could consume 2500 times less energy

Researchers at the University of Minnesota have developed an innovative device that could reduce the energy consumption of artificial intelligence (AI) by at least a factor of 1,000. This major advancement addresses growing concerns about the environmental impact of AI technologies, as their global energy demand is expected to double by 2026, according to the International Energy Agency.


Current AI systems require significant data transfers between memory and processors, a highly energy-intensive process. The new device, called Computational Random Access Memory (CRAM), allows data processing directly within the memory, eliminating the need for transfers. This method, experienced for the first time by the team at the University of Minnesota, could drastically reduce the energy consumption of AI applications.

Yang Lv, a postdoctoral researcher at the University of Minnesota and the lead author of the study, explains that CRAM enables data processing directly in the memory network, thereby eliminating energy-intensive transfer stages. This technology, which relies on the use of magnetic tunnel junctions (MTJ), allows for much more efficient storage and processing of information compared to current methods based on transistors.

Jian-Ping Wang, professor at the University of Minnesota and co-author of the study, emphasizes that this technology is the result of over 20 years of research and interdisciplinary collaboration. Initially considered a wild idea, CRAM has now proven its effectiveness and is ready to be integrated into existing technologies. According to Jian-Ping Wang, initial results show a reduction in energy consumption by a factor of 2,500 compared to traditional systems.

CRAM could represent a sustainable solution for AI development, offering unprecedented energy efficiency. This technological breakthrough opens up promising prospects for reducing the environmental footprint of AI systems while maintaining high performance. The researchers are now looking to collaborate with leaders in the semiconductor industry to develop this technology on a large scale.