Can DNA Become the Next Super‑High‑Density Storage Medium?

DNA storage uses the four nucleobases (A, T, G, C) to encode binary data, mapping 0‑3 to each base, theoretically allowing 1 g of DNA to hold 455 EB of data.

Reading is performed by DNA sequencing, which now can achieve up to 960 Gb per run at low cost, while writing remains a bottleneck: current synthesis can only write megabytes per day, making commercial deployment far off.

The architecture consists of a codec (storage controller) that converts binary to DNA sequences and handles error correction and indexing, a write device that synthesizes DNA strands, a storage device (e.g., cell nuclei or DNA “disk cabinets”), and a read device that uses sequencing (commonly Sanger sequencing).

Key technical challenges include:

Encoding and error correction : avoiding repeats and adding verification data; Microsoft uses a ternary coding scheme where one base signals the previous base.

Indexing : key‑value‑style DNA indexes that embed file headers and addresses.

Synthesis : large‑scale DNA synthesis is expensive and limited to specialized providers such as GeneArt or Twist Bioscience.

Copying : PCR‑based replication, a mature technique since 1983.

Despite these hurdles, DNA offers unparalleled density, longevity (hundreds of thousands of years under dry, cold conditions), and low energy consumption, making it attractive for archival scenarios. Major institutions (U.S. Library of Congress, Wikipedia, Google) and military projects are exploring DNA as a “cloud‑hard‑drive”.

Recent milestones include George Church’s 2012 650 KB write, EMBL’s 2013 20 MB write, and the 2016 Microsoft/University of Washington prototype that stored 200 MB and introduced new error‑correcting codes for random access.

Future progress depends on advances in DNA synthesis and sequencing technologies (e.g., PacBio, Illumina) and the development of DNA chips, synthesis platforms, and sequencing pipelines.