Scientists turn DNA into a memory device that uses 100x less power

Researchers have engineered a hybrid memory device that pairs synthetic DNA with semiconductor technology, achieving storage and data processing in a single system while consuming 100 times less power than conventional approaches. The breakthrough addresses a fundamental inefficiency in modern computing where data must shuttle between separate storage and processing units.

The team embedded DNA strands into a semiconductor substrate, enabling the material to function simultaneously as both memory and processor. This integration eliminates the energy-intensive data movement that plagues current computer architectures. DNA's natural capacity to store vast amounts of information in compact molecular structures makes it an ideal candidate for this application. The synthetic DNA used in the device offers greater stability and control compared to natural DNA.

The research builds on decades of work in DNA computing and bio-hybrid electronics. Previous efforts demonstrated DNA's theoretical potential for information storage, but practical implementation required solving engineering challenges around integration with silicon-based systems. This team successfully bridged that gap by developing a process to incorporate DNA into semiconductor fabrication.

The implications for artificial intelligence and computing power consumption are substantial. Data centers currently account for roughly 1-2 percent of global electricity use, with much of that spent moving information rather than processing it. If this technology scales beyond laboratory conditions, it could dramatically reduce energy demands for AI training and inference.

The researchers acknowledge limitations. The current prototype operates at speeds slower than conventional semiconductors, and the technology requires further development before commercial viability. Stability over extended periods and integration with existing chip manufacturing remains unproven at scale.

The work represents a proof-of-concept demonstration rather than a ready-made solution. The team plans continued research into optimizing both the molecular and electronic components to improve processing speed while maintaining the power efficiency gains. If successful, bio-hybrid memory could become a critical component in next-generation computing architectures designed for energy constraint environments, from edge devices to large-scale data centers.