How far are we from storing the world's archives in DNA?

The road so far

4 of 7 steps done, 1 in progress.

1. Done · 4 Feb 2015Encapsulated DNA survives an accelerated ageing test

   Grass and colleagues stored 83 kB in DNA encapsulated in silica, encoded in 4,991 DNA segments, and recovered it after one week at 70 degrees C. The authors propose that this test is equivalent to about 2,000 years of storage in central Europe. That equivalence is their proposal from one accelerated test, not observed ageing. [2]

2. Done · 1 Mar 2017About 2 MB of files stored and fully recovered

   The DNA Fountain scheme stored about 2.14 million bytes, including an operating system and a film, and recovered all of it from sequencing coverage equal to one Illumina tile. The paper reports a density of 215 petabytes per gram of DNA. That is a density figure, not an amount stored. [3]

3. Done · 19 Feb 2018More than 200 MB stored, with single files retrieved

   Organick and colleagues stored 35 files totalling over 200 MB in more than 13 million oligonucleotides. Each file was retrieved individually without errors, using primer-based random access. This is the largest stored-and-retrieved amount in the peer-reviewed papers opened for this page. [4]

4. Done · 21 Mar 2019Fully automated write, store and read cycle

   Takahashi and colleagues demonstrated an automated device that writes a 5-byte message with a custom synthesiser, stores it, and reads it back with nanopore sequencing. No automated cycle at archive scale appears in the sources opened. [5]

5. In progress · 23 Oct 2024Writing without making each strand from scratch

   Zhang and colleagues wrote about 275,000 bits (roughly 34 kilobytes) by adding epigenetic marks to premade DNA template carriers, which were chemically synthesised in advance, and read them with nanopore sequencing. This is a laboratory result at very small scale. The paper reports only short heat tests of the marks and no long-term stability data. [6]

6. Not started · 18 Oct 2022Writing at archive speed and cost

   A 2022 review puts current DNA archive writing speeds in the order of kilobytes per second, and estimates cost at about 800 million US dollars per terabyte. The same review gives tape at roughly 15 to 16 US dollars per terabyte. These are reported estimates, and the review notes they cite earlier work. For comparison, a 2026 Nature paper on laser-written glass reports a peak write rate of 25.6 Mbit/s per beam (about 3.2 MB/s), under different conditions and at a different scale. [1][7]

7. Not started · 18 Oct 2022Reading back at practical speed

   The 2022 review says reading is much slower than standard silicon-based storage. It notes that, without random access, retrieving one file would mean reading the whole database. Reading needs sequencing plus heavy computational decoding. [1]

Why it matters

Libraries, museums and governments hold records that must survive for decades or longer. DNA is very dense, so a small volume could in principle hold a large archive. Knowing where the gaps lie shows whether DNA is close to practical archiving or still at the bench, and how it compares with tape and glass.

What stands in the way

• Writing is the main bottleneck. A 2022 review puts current archive writing speeds in the kilobytes-per-second range.
• The same review estimates about 800 million US dollars per terabyte to write DNA, or about 800 US dollars per megabyte (our conversion, in decimal units). Tape is given at roughly 15 to 16 US dollars per terabyte in the same review.
• Reading needs sequencing and heavy decoding. The review says reading is much slower than silicon-based storage, and without random access retrieving one file would mean reading the whole database.
• Long-term life is inferred from accelerated tests. The 2,000-year equivalence is the authors' proposal from one week at 70 degrees C, not measured ageing.
• The sources opened describe no commercial archival product, standard or integration with existing archive systems.

Signals to watch

• A write speed or cost per megabyte measured at archive scale and reported in a peer-reviewed paper, replacing estimates.
• Random access or a fully automated write, store and read cycle run on hundreds of megabytes or more.
• Non-synthesis writing methods, such as epigenetic printing, shown at larger scale with read-out and stability data.
• A new review that updates the 2022 cost and speed figures.
• Laser-written glass moving from research to a product or release date, as a comparison point for the archive gap.

Sources

1. review article
   Emerging Approaches to DNA Data Storage: Challenges and Prospects
   Doricchi A et al.. ACS Nano, 18 Oct 2022.
2. peer reviewed paper
   Robust chemical preservation of digital information on DNA in silica with error-correcting codes
   Grass RN, Heckel R, Puddu M et al.. Angewandte Chemie International Edition, 4 Feb 2015.
3. peer reviewed paper
   DNA Fountain enables a robust and efficient storage architecture
   Erlich Y, Zielinski D. Science, 1 Mar 2017.
4. peer reviewed paper
   Random access in large-scale DNA data storage
   Organick L et al.. Nature Biotechnology, 19 Feb 2018.
5. peer reviewed paper
   Demonstration of End-to-End Automation of DNA Data Storage
   Takahashi CN, Nguyen BH, Strauss K et al.. Scientific Reports, 21 Mar 2019.
6. peer reviewed paper
   Parallel molecular data storage by printing epigenetic bits on DNA
   Zhang C, Wu R, Sun F et al.. Nature, 23 Oct 2024.
7. peer reviewed paper
   Laser writing in glass for dense, fast and efficient archival data storage
   Microsoft Research Project Silica Team. Nature, 18 Feb 2026.

Last reviewed 21 Oct 2026.