Taking inspiration from living organisms to store data: DNA, a "new" medium

How and why store data in DNA?

Recent news has highlighted digital storage on DNA, notably mentioning the French startup Biomemory, which offers this new type of storage as an alternative to conventional hard drives. The DNA referred to in this context is synthetic DNA, not living DNA. In living organisms, DNA has a “purpose”: its structure and sequences produce biological effects, evolutionary changes, and so on. Here, however, DNA sequences are artificially constructed in laboratories with no intention of producing such effects (there is no biological activity involved), but solely to store information. Nucleotides – the molecules that make up the basic building blocks of DNA – are simply used as “bricks” assembled together to “write” code. This new technique falls squarely within the field of biomimicry, which involves drawing inspiration from living systems in innovation and engineering (while adapting them to technical constraints – for example, airplanes do not flap their wings).

In practical terms, data are encoded using a base-four system, since DNA is composed of four nucleotides: A, C, G, and T (adenine, cytosine, guanine, and thymine). Digital technology, by contrast, relies on a binary system (the well-known bits: 0 or 1). Information must therefore be translated into this new encoding scheme. It is important to note that this is encoding, not encryption: it is merely a different way of representing information. The security challenges remain essentially the same as with other storage systems. This technology also requires new DNA reading devices, which have become increasingly accessible in laboratories thanks to technological advances such as the famous CRISPR/Cas9 “genetic scissors”, which allow DNA manipulation.

The idea of storing information in biological material dates back to the late 1950s, alongside developments in genetics and cybernetics, particularly through comparisons between humans and machines (for example, in the work of Norbert Wiener). This concept would later inspire significant aspects of transhumanist ideology (see Nicolas Le Dévédec’s work on the topic). The first practical demonstrations of DNA data storage appeared in the late 1980s, with rapid progress since the mid-2010's:

• In 2011, a research team successfully encoded a 659-kilobit book.
• In 2012, another study encoded an HTML document containing a 53,000-word book, seven JPEG images, and a JavaScript program.
• In 2019, 16 gigabits of data from Wikipedia were encoded into synthetic DNA.

As a result, numerous research projects have been launched since the 2020s, often involving consortia of research institutes, companies, and specialized startups, such as the DNA Data Storage Alliance. Other initiatives have been funded by the Intelligence Advanced Research Projects Activity (IARPA), the U.S. intelligence agency responsible for directing research toward specific strategic challenges.

A revolution… still to come, and primarily for static storage

The technology currently faces several limitations:

• First, it is not immune to errors during encoding and decoding, particularly due to DNA-specific constraints (instability, structural limitations, etc.). However, these errors appear to be decreasing as research progresses and quality-control processes improve.
• Second, there is a significant time cost involved: reading and writing data are not yet fast processes. This makes DNA storage well suited for archiving but less appropriate for active databases.
• Third, the financial cost remains substantial. For example, in 2017, synthesizing "2 megabytes of data reportedly cost around $7,000, while reading it cost about $2,000". A 2024 PhonAndroid article referred to pricing in the range of €1,000 per kilobyte.

DNA can store enormous amounts of information in extremely small spaces while consuming very little energy (“This ultra-dense molecule –1 gram of DNA can hold 450 million terabytes – is both stable and long-lasting”). This is precisely why it appears so promising at a time when data centers are expanding rapidly and are expected to continue doing so. Another major advantage is longevity: whereas digital technologies often become obsolete relatively quickly, requiring regular replacement of storage devices, DNA storage could offer remarkable durability and resilience to environmental changes, such as temperature fluctuations. After all, DNA has transmitted information about the distant past through fossils.

DNA storage offers additional benefits. Once encoded, DNA can be freeze-dried, dramatically reducing the physical space needed for archiving vast amounts of data. As one estimate suggests, “the data currently stored in the world’s ten million data centers could fit into just 200 grams of DNA” (see the CNRS article on the topic). It is therefore important to distinguish between the liquid or “wet” form (sometimes referred to as wetware, by analogy with hardware and software), which allows information to be written and read, and the “dry” form used for long-term storage. Physically, such storage may take the form of cards resembling credit cards and based on silicon substrates, or metal beads and capsules that can then be stored in large cabinets.

It should be emphasized that DNA storage is primarily intended for archival purposes. The data are essentially “frozen,” and easy, frequent access does not appear to be the primary objective, since retrieving information requires converting the DNA from its dry state back into a wet one. This may evolve as the technology matures. Consequently, the challenges associated with these new storage media are comparable to those of more traditional archival formats, such as microfilm, which was long used as an analog archiving medium. Comparisons with data centers may therefore be somewhat misleading, since data centers serve functions far beyond archiving, including data access and real-time computation. Other approaches to long-term or “deep-time” archiving also exist, particularly in efforts to preserve human knowledge over extended timescales (see the CNIL's ethical event and the report associated on this topic).

What about data protection issues?

At first glance, it is difficult to identify entirely new data-protection challenges specific to this technology. First, because the storage medium is synthetic DNA rather than DNA from living organisms. Second, because DNA storage is essentially a new encoding method, replacing binary encoding with a base-four system built on DNA nucleotides. In this sense, DNA could simply become the next storage medium after film, USB drives, hard disks, and SSDs.

From a security perspective, DNA storage does not fundamentally change the landscape. Since it is merely a different encoding format, the same requirements remain: data encryption, access management, and related safeguards (see CNIL’s Guide on personal data security). That said, the exceptional resilience of DNA-based media – through freeze-drying, natural tolerance to temperature changes, and other characteristics – may help preserve their physical integrity, and therefore the integrity of the data they contain.

Like any storage medium, DNA can hold personal data and therefore raises all the usual issues associated with data protection. However, the exercise of individual rights may pose practical challenges. How easily can archived data be accessed? How can data be corrected once encoded into a DNA structure? Would modifications require restructuring and re-encoding the entire dataset, much like certain operations on blockchain systems?