By Sophie Lindqvist, Chip Dispatch

A memory device built from synthetic DNA and perovskite consumes 100 times less power than comparable storage, according to Penn State researchers. The team published its work in Advanced Functional Materials and has filed a patent application. One gram of DNA can hold roughly 215 million gigabytes. That density, paired with perovskite electronics, forms the basis of the claim.

The device is a memristor. It retains a record of previous current flow after power is removed. That property lets it store and process information in the same place, an arrangement the researchers compare to neural behavior. The caveat is scale. The paper describes a laboratory device, not a manufacturable product, and the authors themselves note that practical commercial systems would still require enough capacity and power to become costly without DNA's spatial efficiency.

How did the team make DNA conduct electricity?

The researchers doped short, customized synthetic DNA sequences with silver nanoparticles. The doping made the DNA conductive and aligned its molecular units in a more orderly arrangement. They then integrated the DNA with thin films of crystalline perovskite, a semiconductor already used in solar cells, lasers and data storage.

Synthetic DNA mattered here. Natural DNA forms long, entangled strands. Short, rigid synthetic pieces can be placed precisely at very small scales.

"We can computationally determine exactly which sequences we need and how long they should be, and then we can rationally design them with synthetic DNA," said co-author Neela H. Yennawar, research professor and director of the Biomolecular Interactions Core Facility at Penn State's Huck Institutes. She said the structures can be doped with silver and other ions and engineered to interface with perovskites, turning DNA into a programmable nanomaterials platform.

The researchers did not report wafer-scale integration, cycling endurance data, or fabrication costs. Those gaps will determine whether this leaves the laboratory.

What performance numbers did the device achieve?

Electrons moved reliably at applied voltages under 0.1 volt. Standard U.S. outlets supply 120 volts, for comparison. The device responded predictably when current direction changed.

Stability figures were specific:

• Consistent operation at temperatures approaching 250 degrees Fahrenheit
• Room-temperature functionality for more than six weeks
• Equal memory function to comparable technologies at one-tenth the power

The team stated this performance substantially exceeds existing perovskite-based memory storage devices. No third-party verification exists yet.

"Usually, it takes more power to store more information," said co-corresponding author Bed Poudel, research professor of materials science and engineering at Penn State. "Our device, however, consumes 100 times less power and the storage capacity is higher than traditional storage devices, like flash drives."

That claim lacks a head-to-head flash comparison in the released material. Readers should treat it as the authors' assertion, not an independent benchmark.

Why does this matter for AI hardware?

Poudel tied the work to neuromorphic computing, which evaluates multiple inputs at once and weighs previous states. "As the demand for artificial intelligence grows, we need a new strategy for low-power, high-storage devices," he said.

Co-corresponding author Kavya S. Keremane, a postdoctoral researcher, framed the result as a materials advance rather than a product. "Biology and electronics are different domains," she said. "Bridging these two fields required developing an entirely new materials platform."

She also noted the result depended on the combination. DNA alone or perovskite alone did not produce a robust device.

The U.S. National Science Foundation, the National Institutes of Health, Penn State and the University of Minnesota funded the research. Co-corresponding author Rashmi Jha of the University of Minnesota contributed, alongside Penn State co-authors Luyao Zheng, Haodong Wu, Jiamao Zheng, Shashank Priya and Chiranth C. Ravi.

The team now plans to improve the technology and explore further bio-inspired applications. Whether silver-doped DNA can survive semiconductor manufacturing conditions, and at what cost, remains untested.