Sealed DNA archive cartridge

Speculative concept
AI illustration · Illustrative scene, not a photograph of the specific device or experiment.

This is an idea, not a product. Nothing here has been built or tested unless a linked technology profile says so.

The problem

Some organisations need to keep a small, fixed set of records for many years. DNA is a dense medium that can be stored without power, but it remains a laboratory technology, so the open question for an archive is whether it could ever be practical.

The idea

A first step would be a sealed, lab-scale archive cartridge that holds one small, fixed set of files, not a computer that works on DNA. The files would be encoded with error-correcting redundancy, synthesised as short DNA strands, encapsulated in silica and kept with a separate written index. Recovery would use a sequencing service and a decoder, so the cartridge would be written once and read only when needed. The concept is limited to files of no more than 83 kB, the amount held in silica in the 2015 study.

Enabling technology

  • DNA data storage: DNA data storage encodes digital files in synthetic DNA sequences, then reads them with sequencing. DNA is extremely dense and can remain stable for long periods when stored well, but writing and reading remain slower and more expensive than electronic storage. In 2026, researchers reported advances in composite-letter coding and a codec tested in vitro and in living cells. These studies address parts of the system, not the cost and throughput of a complete archival service.

What stands in the way

  • Writing is the bottleneck. A 2022 review, citing earlier estimates, put current writing speeds in the kilobytes-per-second range and cost at about 800 million US dollars per terabyte. These are reported figures and have not been independently measured for this concept.
  • Reading needs sequencing plus heavy computational decoding. Fully automated write, store and read cycles have so far been shown only for a 5-byte message.
  • Synthesis and sequencing introduce errors and strand dropouts, so the cartridge needs substantial redundancy and error correction, which adds to storage cost.
  • Random access has been shown across more than 200 MB in research settings only, and there are no established standards or integration with existing archive systems.
  • Long-term life has not been measured directly. The 2015 study's equivalence to about 2,000 years of storage is the authors' proposal, based on one accelerated ageing test at 70 degrees C for one week.

A cheap way to test it

Bench test with no human subjects. Encode a test file of no more than 83 kB, using a sequence design screened for biosecurity concerns, and order the oligonucleotides from a commercial synthesis provider. Encapsulate the DNA in silica, then split it into at least two replicate tubes. Heat one replicate at 70 degrees C for one week, as in the 2015 study, and keep a second replicate unheated as a control. Sequence and decode both replicates, then compare the recovered file with the original by checksum. Pass: both replicates recover the full file exactly after error correction. Fail: any uncorrected error in either replicate. Also record the time and reagent cost of writing per kilobyte and compare these with the 2022 review's estimates.