howclose.to
DNA DATA STORAGE

The problemsIntelligence & MachinesDNA Data Storage

Updated · July 2026

How close are we to practical DNA data storage?Can DNA become a real computer archive?

Can DNA become a real computer archive?How close are we to practical DNA data storage?

Synthetic DNA has recovered a 200 MB random-access archive without bit errors and automated a complete write-to-read cycle for five bytes, but no public benchmark joins useful payload, throughput, retrieval and cost.DNA can hold computer files in an extraordinarily small sample. The missing step is a useful machine that writes a large archive quickly, finds one file and costs less than today's cold storage.

Current stageEngineeringEngineering a system that works outside the labBuilding a system that works outside the lab
Read the evidence Where we stand
The path so far

Recorded milestones

  1. 739 kB recovered exactly

  2. The integrated benchmark is still missing

Open tests · conceptual branches

TODAY
  1. OPEN TEST03 ↗
    Write it fast enough and cheaply enoughArchive-class cost and throughputThe most famous density experiment spent about $7,000 to synthesize 2.1 MB and another $2,000 to read it.Publish independently reproducible cost per terabyte and sustained write throughput, including redundancy, reagents, containers and instruments.
  2. OPEN TEST04 ↗
    Run a real petabyte archiveOperational petabyte archiveThe site's one-petabyte marker is an editorial scale test, not a company promise or a measured technical limit.Operate a searchable one-petabyte DNA archive with verified integrity, retention conditions and recovery procedures.
TODAY
OPEN TEST03 ↗
Write it fast enough and cheaply enoughArchive-class cost and throughputThe most famous density experiment spent about $7,000 to synthesize 2.1 MB and another $2,000 to read it.Publish independently reproducible cost per terabyte and sustained write throughput, including redundancy, reagents, containers and instruments.
OPEN TEST04 ↗
Run a real petabyte archiveOperational petabyte archiveThe site's one-petabyte marker is an editorial scale test, not a company promise or a measured technical limit.Operate a searchable one-petabyte DNA archive with verified integrity, retention conditions and recovery procedures.
Milestones are in recorded order. Open tests are separate questions, not a forecast.Milestones are in date order. Open tests are separate tests, not a promise.
DNA Data StorageWhere we stand

The path shows The integrated benchmark is still missing. Open thresholds are separate tests in a conceptual map, not a forecast. No single percentage represents progress.

01 · Where we standWhere we stand

Where we standWhere we stand

The current stage, the evidence recorded, and the single threshold that gates the next stage.The stage reached, the evidence recorded, and the one test still blocking the next step.

What the evidence has shownWhat the evidence has shown

DNA can hold computer files in an extraordinarily small sample. The missing step is a useful machine that writes a large archive quickly, finds one file and costs less than today's cold storage.Synthetic DNA has recovered a 200 MB random-access archive without bit errors and automated a complete write-to-read cycle for five bytes, but no public benchmark joins useful payload, throughput, retrieval and cost.

2 of 4 tests recorded2 of 4 thresholds achieved open
✓Get a large set of files backLossless multi-file archiveAchieved
✓Make the whole process automaticAutomated end-to-end cycleAchieved
Write it fast enough and cheaply enoughArchive-class cost and throughputOpen test
Run a real petabyte archiveOperational petabyte archiveOpen test

The next testBlocking threshold

Write it fast enough and cheaply enoughArchive-class cost and throughput

An appliance vendor must publish an audited end-to-end dollars-per-TB and bytes-per-second benchmark.An appliance vendor must publish an audited end-to-end dollars-per-TB and bytes-per-second benchmark.

02 · Thresholds & evidenceThresholds & evidence

What would count as proof?What would count as proof?

A clear test, the evidence behind it, and the limits of the claim.A clear test, the evidence behind it, and what it still leaves open.

✓Achieved · Mar 2018

Get a large set of files backLossless multi-file archive

A 200 MB pool containing 35 files was selectively retrieved and fully recovered without bit errors.Encode, synthesize, selectively retrieve and reconstruct a nontrivial multi-file archive with no bit errors.

Recorded result Microsoft and University of Washington · 200 MB

Next test Repeat at terabyte scale with public redundancy, error-rate and recovery-latency data.

SOURCE RECORD

2018 200 MB random-access DNA archive ↗

Still to demonstrateWhat this does not establishA separate system can still have different conditions and results.The named measurement has no wider implication until the supporting system test is also shown.

View every evidence record

Achieved

Get a large set of files backLossless multi-file archive

A 200 MB pool containing 35 files was selectively retrieved and fully recovered without bit errors.Encode, synthesize, selectively retrieve and reconstruct a nontrivial multi-file archive with no bit errors.

Recorded result: Mar 2018 · Microsoft and University of Washington · 200 MB

Next test: Repeat at terabyte scale with public redundancy, error-rate and recovery-latency data.

2018 200 MB random-access DNA archive ↗

Achieved

Make the whole process automaticAutomated end-to-end cycle

The first automated machine completed the full cycle, but its payload was only the five-byte word HELLO.Run encoding, DNA writing, storage preparation, sequencing and decoding without manual transfers.

Recorded result: Mar 2019 · Five bytes recovered after an approximately 21-hour write-to-read cycle

Next test: Automate at least a gigabyte while publishing write speed, read latency and error-correction overhead.

2019 automated end-to-end device ↗

Open test

Write it fast enough and cheaply enoughArchive-class cost and throughput

The most famous density experiment spent about $7,000 to synthesize 2.1 MB and another $2,000 to read it.Publish independently reproducible cost per terabyte and sustained write throughput, including redundancy, reagents, containers and instruments.

Next test: An appliance vendor must publish an audited end-to-end dollars-per-TB and bytes-per-second benchmark.

Open test

Run a real petabyte archiveOperational petabyte archive

The site's one-petabyte marker is an editorial scale test, not a company promise or a measured technical limit.Operate a searchable one-petabyte DNA archive with verified integrity, retention conditions and recovery procedures.

Next test: A deployment must publish usable payload, write and retrieval performance, integrity checks and operating cost together.

03 · The recordThe record

The record behind the verdictThe record behind the verdict

Major milestones, smaller context events and setbacks share one source-linked record.Milestones, smaller events and setbacks share one sourced record.

Event key
TheoryExperimentFundingSetbackPolicyDeployment Future / scheduled
1964-20111 event1 shown

Molecular memory is imagined

Recorded event: “Nucleic-acid memory is proposed”.

1964
Nucleic-acid memory is proposedTheory
Mikhail Neiman discussed the possibility of using nucleic-acid macromolecules as non-biological information memory.
Norbert Wiener and Mikhail Neiman
2012-20162 events2 shown

Digital files enter DNA

This record spans “A 5.27-megabit book enters DNA” to “Five files return with 100% accuracy”.

2012
A 5.27-megabit book enters DNAExperiment
A book containing more than 53,000 words, 11 images and a program was encoded in about 55,000 synthetic DNA strands. This was an encoded-payload demonstration, not the later no-bit-error archive record.
George Church, Yuan Gao and Sriram Kosuri
2017-20181 event0 shown

Density and random access

Recorded event: “DNA Fountain demonstrates extreme density at small scale”.

2017
DNA Fountain demonstrates extreme density at small scale
A 2.1 MB experiment reported 215 PB/g physical density. Synthesis cost about $7,000 and sequencing about $2,000, so density did not imply an affordable storage system.
Yaniv Erlich and Dina Zielinski
2019-20242 events1 shown

Automation and molecular search

This record spans “Automation completes a five-byte cycle” to “A molecular index searches by image similarity”.

2019
Automation completes a five-byte cycleExperiment
The first automated end-to-end device wrote, stored and recovered HELLO. The five-byte cycle took about 21 hours and only one of 30 extractable payload reads decoded perfectly in the reported run.
Microsoft and University of Washington
2021
A molecular index searches by image similarity
A DNA-based index performed similarity search over representations of 1.6 million images, demonstrating computation near the stored molecules rather than a larger archival payload.
University of Washington, Microsoft and collaborators
2025-20262 events1 shown

The integrated-system test

This record spans “Cas9 adds addressed and semantic retrieval” to “Writing and system integration remain the bottlenecks”.

2025
Cas9 adds addressed and semantic retrieval
Cas9-guided methods selectively enriched files and tested semantic image lookup. The semantic database represented 1.74 million images with 457 synthesized address sequences; it was not a 1.74-million-file archive.
University of Washington and Microsoft
2025
Writing and system integration remain the bottlenecksSetbackWe are here
A field assessment still identified cost-effective high-throughput writing, practical storage handling and integrated workflows as barriers to deployment.
DNA computing and data-storage researchers
2018-20181 event1 shown

Additional records

Recorded event: “Random-access archive reaches 200 MB”.

2018
Random-access archive reaches 200 MBExperiment
Thirty-five files totaling 200 MB were encoded in more than 13 million oligonucleotides, selectively accessed and fully recovered with no bit errors.
Microsoft and University of Washington
— end of record · 9 events shown —
04 · The dataThe data

The data behind the verdictThe data behind the verdict

Read each measure with its units, conditions, uncertainty and source; targets remain distinct from observations.Read each chart with its conditions, limits and source. Future targets remain distinct from recorded results.

Open figure in archive: Selected payload demonstrations increased sharply, but they differ in error claim, file structure and retrieval method and are not a deployment curve.Open figure in archive: Automation was proved independently of archive scale: the complete unattended write-store-read demonstration handled five bytes.Open figure in archive: The measured molecular density was exceptional, but the payload was only 2.1 MB and the figure excludes containers, instruments, indexing and redundancy.Open figure in archive: Selective retrieval worked in a pool containing more than 13 million distinct designed DNA strands.Open figure in archive: The first fully automated payload was 40 million times smaller than the 200 MB manually integrated research pool.Open figure in archive: The DNA Fountain experiment reported about $7,000 for synthesis and $2,000 for sequencing its 2.1 MB payload.
Fully recovered research-archive payloadFully recovered research-archive payload
Scale

Five files · 100% accurate recovery · 0.739 MB user data · 2013 · verification unspecified

Fully recovered research-archive payload over time, with measured values, projected values, and a goal at 1e+9 MB user data.1101001,00010,000100,0001e+61e+7Fully recovered research-archive payload · MB user dataYear20132018GOAL 1e+9 · One-petabyte operational archive…Five files · 100% accurate…35 files · no bit errors~1e+9 MB user data to goal
NOTE - This archive-payload ladder is a proxy because no comparable public series combines payload, write cost, throughput, random access and automation. The one-petabyte endpoint is an editorial scale test, not a forecast. The automated 2019 system is excluded from the archive curve and shown separately because it handled only five bytes.
View data table
All metric observations
DateValueStatusCondition / labelVerificationCICensoringSource
20130.739 MB user dataobservedFive files · 100% accurate recoveryunspecified—none2013 practical high-capacity DNA archive
2018200 MB user dataobserved35 files · no bit errorsunspecified—none2018 200 MB random-access DNA archive
AS OF July 2026
View every data figure
Fully recovered research-archive payloadFully recovered research-archive payload
Scale

Five files · 100% accurate recovery · 0.739 MB user data · 2013 · verification unspecified

Fully recovered research-archive payload over time, with measured values, projected values, and a goal at 1e+9 MB user data.1101001,00010,000100,0001e+61e+7Fully recovered research-archive payload · MB user dataYear20132018GOAL 1e+9 · One-petabyte operational archive…Five files · 100% accurate…35 files · no bit errors~1e+9 MB user data to goal
NOTE - This archive-payload ladder is a proxy because no comparable public series combines payload, write cost, throughput, random access and automation. The one-petabyte endpoint is an editorial scale test, not a forecast. The automated 2019 system is excluded from the archive curve and shown separately because it handled only five bytes.
View data table
All metric observations
DateValueStatusCondition / labelVerificationCICensoringSource
20130.739 MB user dataobservedFive files · 100% accurate recoveryunspecified—none2013 practical high-capacity DNA archive
2018200 MB user dataobserved35 files · no bit errorsunspecified—none2018 200 MB random-access DNA archive
AS OF July 2026
Selected payload demonstrations increased sharply, but they differ in error claim, file structure and retrieval method and are not a deployment curve.Selected payload demonstrations increased sharply, but they differ in error claim, file structure and retrieval method and are not a deployment curve.

Independent records

Selected encoded payload demonstrations log scale · MB

log scale · MB
5.27-megabit book20120.6588 MB
Five files · exact recovery20130.739 MB
35-file random-access pool2018200 MB
observed self-reported target / pathway right-censored lower bound
View data table
All plotted observations
SeriesDateValueStatusCondition / labelVerificationCICensoringSource
Selected encoded payload demonstrations20120.65875 MBobserved5.27-megabit bookunspecified—none2012 next-generation digital information storage in DNA
Selected encoded payload demonstrations20130.739 MBobservedFive files · exact recoveryunspecified—none2013 practical high-capacity DNA archive
Selected encoded payload demonstrations2018200 MBobserved35-file random-access poolunspecified—none2018 200 MB random-access DNA archive
AS OF July 2026
Automation was proved independently of archive scale: the complete unattended write-store-read demonstration handled five bytes.Automation was proved independently of archive scale: the complete unattended write-store-read demonstration handled five bytes.

log scale · bytes by year

HELLO · ~21-hour cycle · 5 bytes · Jan 2019 · verification unspecified

Automation was proved independently of archive scale: the complete unattended write-store-read demonstration handled five bytes.5 bytes · Jan 2019HELLO · ~21-hour cycle: 5 bytes · Jan 2019Year
Log-scale line joins each new longest record; all observations remain individual dots.

Latest record

HELLO · ~21-hour cycle

5 bytes · Jan 2019

observed · Verification unspecified

View data table
All plotted observations
SeriesDateValueStatusCondition / labelVerificationCICensoringSource
Automated end-to-end payload20195 bytesobservedHELLO · ~21-hour cycleunspecified—none2019 automated end-to-end device
AS OF July 2026
The measured molecular density was exceptional, but the payload was only 2.1 MB and the figure excludes containers, instruments, indexing and redundancy.The measured molecular density was exceptional, but the payload was only 2.1 MB and the figure excludes containers, instruments, indexing and redundancy.

log scale · PB/g by year

2.1 MB experiment · 215 PB/g · Jan 2017 · verification unspecified

The measured molecular density was exceptional, but the payload was only 2.1 MB and the figure excludes containers, instruments, indexing and redundancy.215 PB/g · Jan 20172.1 MB experiment: 215 PB/g · Jan 2017Year
Log-scale line joins each new longest record; all observations remain individual dots.

Latest record

2.1 MB experiment

215 PB/g · Jan 2017

observed · Verification unspecified

View data table
All plotted observations
SeriesDateValueStatusCondition / labelVerificationCICensoringSource
DNA Fountain demonstrated physical density2017215 PB/gobserved2.1 MB experimentunspecified—none2017 DNA Fountain density and cost experiment
AS OF July 2026
Selective retrieval worked in a pool containing more than 13 million distinct designed DNA strands.Selective retrieval worked in a pool containing more than 13 million distinct designed DNA strands.

linear scale · million oligonucleotides by year

35 files · 200 MB · 13 million oligonucleotides · Jan 2018 · verification unspecified

Selective retrieval worked in a pool containing more than 13 million distinct designed DNA strands.13 million oligonucleotides · Jan 201835 files · 200 MB: 13 million oligonucleotides · Jan 2018Year
Log-scale line joins each new longest record; all observations remain individual dots.

Latest record

35 files · 200 MB

13 million oligonucleotides · Jan 2018

observed · Verification unspecified

View data table
All plotted observations
SeriesDateValueStatusCondition / labelVerificationCICensoringSource
Random-access library size201813 million oligonucleotidesobserved35 files · 200 MBunspecified—none2018 200 MB random-access DNA archive
AS OF July 2026
The first fully automated payload was 40 million times smaller than the 200 MB manually integrated research pool.The first fully automated payload was 40 million times smaller than the 200 MB manually integrated research pool.
Automated payload5 bytes
+2e+8 bytes · gain 40,000,000
Research pool2e+8 bytes
Automated payload5 bytes
+2e+8 bytes · gain 40,000,000
Research pool2e+8 bytes
View data table
All aligned comparisons
ClaimABSource / provenance
The first fully automated payload was 40 million times smaller than the 200 MB manually integrated research pool.5 bytes · Automated payload200000000 bytes · Research pool2019 automated end-to-end device
AS OF July 2026
The DNA Fountain experiment reported about $7,000 for synthesis and $2,000 for sequencing its 2.1 MB payload.The DNA Fountain experiment reported about $7,000 for synthesis and $2,000 for sequencing its 2.1 MB payload.
Synthesis7,000 USD
−5,000 USD · 0.29× of A
Sequencing2,000 USD

Each square = 200 USD

Synthesis7,000 USD
Sequencing2,000 USD
5,000 USD less from sequencing than from synthesis
View data table
All aligned comparisons
ClaimABSource / provenance
The DNA Fountain experiment reported about $7,000 for synthesis and $2,000 for sequencing its 2.1 MB payload.7000 USD · Synthesis2000 USD · Sequencing2017 DNA Fountain density and cost experiment
AS OF July 2026
05 · What it unlocksWhat it unlocks

If the remaining tests pass.If the remaining tests pass.

Downstream capabilities remain conditional on results that are not yet in.These are possible next capabilities, conditional on results that are not yet in.

TODAY
  1. PASSED · Mar 201801 ↗
    Get a large set of files backLossless multi-file archiveA 200 MB pool containing 35 files was selectively retrieved and fully recovered without bit errors.Encode, synthesize, selectively retrieve and reconstruct a nontrivial multi-file archive with no bit errors.
  2. PASSED · Mar 201902 ↗
    Make the whole process automaticAutomated end-to-end cycleThe first automated machine completed the full cycle, but its payload was only the five-byte word HELLO.Run encoding, DNA writing, storage preparation, sequencing and decoding without manual transfers.
  3. OPEN TEST03 ↗
    Write it fast enough and cheaply enoughArchive-class cost and throughputThe most famous density experiment spent about $7,000 to synthesize 2.1 MB and another $2,000 to read it.Publish independently reproducible cost per terabyte and sustained write throughput, including redundancy, reagents, containers and instruments.
  4. OPEN TEST04 ↗
    Run a real petabyte archiveOperational petabyte archiveThe site's one-petabyte marker is an editorial scale test, not a company promise or a measured technical limit.Operate a searchable one-petabyte DNA archive with verified integrity, retention conditions and recovery procedures.
PASSED · Mar 201801 ↗
Get a large set of files backLossless multi-file archiveA 200 MB pool containing 35 files was selectively retrieved and fully recovered without bit errors.Encode, synthesize, selectively retrieve and reconstruct a nontrivial multi-file archive with no bit errors.
PASSED · Mar 201902 ↗
Make the whole process automaticAutomated end-to-end cycleThe first automated machine completed the full cycle, but its payload was only the five-byte word HELLO.Run encoding, DNA writing, storage preparation, sequencing and decoding without manual transfers.
TODAY
OPEN TEST03 ↗
Write it fast enough and cheaply enoughArchive-class cost and throughputThe most famous density experiment spent about $7,000 to synthesize 2.1 MB and another $2,000 to read it.Publish independently reproducible cost per terabyte and sustained write throughput, including redundancy, reagents, containers and instruments.
OPEN TEST04 ↗
Run a real petabyte archiveOperational petabyte archiveThe site's one-petabyte marker is an editorial scale test, not a company promise or a measured technical limit.Operate a searchable one-petabyte DNA archive with verified integrity, retention conditions and recovery procedures.

Possible, if the open tests pass.

  1. Ultra-dense cold archives

    This is a possible benefit only if ultra-dense cold archives works safely, reliably and at useful scale.Potential benefit only if the stated use (Ultra-dense cold archives) is demonstrated with evidence for safety, reliability, affordability and scale.

  2. Low-standby-energy preservation

    This is a possible benefit only if low-standby-energy preservation works safely, reliably and at useful scale.Potential benefit only if the stated use (Low-standby-energy preservation) is demonstrated with evidence for safety, reliability, affordability and scale.

  3. Search inside molecular libraries

    This is a possible benefit only if search inside molecular libraries works safely, reliably and at useful scale.Potential benefit only if the stated use (Search inside molecular libraries) is demonstrated with evidence for safety, reliability, affordability and scale.

These are possible outcomes, if the open tests pass. They are conditional, not promises.
06 · SourcesSources

Follow the evidence.Follow the evidence.

9 sources — every figure on this page traces to one.9 sources — every figure on this page traces to one.

Source · what this page citesScope and limits
  1. 01

    Scope: Not recorded. Limit: Not recorded.

  2. 02

    Scope: Primary publication named by 2013 practical high-capacity DNA archive. Limit: The publication's measured scope does not establish downstream engineering or deployment claims.

  3. 03

    2017 DNA Fountain density and cost experiment

    Read source ↗science.org
    Used on this page
    All 8 uses across 6 destinations

    Scope: Primary publication named by 2017 DNA Fountain density and cost experiment. Limit: The publication's measured scope does not establish downstream engineering or deployment claims.

  4. 04

    2018 200 MB random-access DNA archive

    Read source ↗microsoft.com
    Used on this page
    All 10 uses across 7 destinations

    Scope: Not recorded. Limit: Not recorded.

  5. 05

    2019 automated end-to-end device

    Read source ↗nature.com
    Used on this page
    All 8 uses across 6 destinations

    Scope: Primary publication named by 2019 automated end-to-end device. Limit: The publication's measured scope does not establish downstream engineering or deployment claims.

  6. 06

    2021 molecular-level similarity search

    Read source ↗nature.com
    Used on this page
    All 2 uses across 2 destinations

    Scope: Primary publication named by 2021 molecular-level similarity search. Limit: The publication's measured scope does not establish downstream engineering or deployment claims.

  7. 07

    Scope: Primary publication named by 2025 Cas9 random access and semantic search. Limit: The publication's measured scope does not establish downstream engineering or deployment claims.

  8. 08

    Scope: Primary publication named by 2025 DNA computing and storage challenges. Limit: The publication's measured scope does not establish downstream engineering or deployment claims.

  9. 09

    DNA stability and systems review

    Read source ↗nature.com
    Used on this page
    All 2 uses across 1 destination

    Scope: Primary publication named by DNA stability and systems review. Limit: The publication's measured scope does not establish downstream engineering or deployment claims.

Sources describe evidence. Inclusion is not endorsement.