DNA can store data for centuries, but at a price no one can pay – yet.
Every few centuries, humanity reinvents how it remembers itself. Clay tablets became paper, paper became magnetic tape and more recently silicon. Each leap made storage faster, cheaper, and denser, but none made it permanent. Every digital format we have invented so far is destined to decay.
Now, a molecule that has preserved life’s code for billions of years is being asked to do the same for our data.
DNA and its family of molecules have revolutionized modern life. The ability to read, understand, and manipulate DNA has transformed how we treat diseases, engineer organisms, and grow food. Yet despite these breakthroughs, DNA has not yet delivered on one of its most hyped promises: revolutionizing the way we store our ever-growing mountains of data.
Humanity generates more digital information each year than conventional media can economically hold. For more than a decade, scientists and the tech press have pointed to DNA as a potential solution, nature’s preferred medium for storing and transmitting information, prized for its incredible density and longevity.
Once written into DNA, data can be copied endlessly at comparatively low cost. Reading DNA has become progressively cheaper, driven by the genomics revolution and the race to decode the human genome. The potential data density of DNA is unmatched by any current storage medium, and under the right conditions, DNA molecules can remain stable for centuries. Unlike magnetic or optical formats, DNA as a storage format will never become obsolete.
On paper, it sounds like a no-brainer. So why aren’t we storing all our data in DNA already?
Storing one megabyte in DNA costs more than a million times as much as putting it on an SSD, and more than 2.5 million times the cost of magnetic tape, the current standard for archival storage. Latency is another hurdle: retrieving data can take hours or days, not milliseconds. Because DNA is stored in liquid, random access to specific files adds complexity and cost. The field also lacks the standardized infrastructure and automation needed to plug into modern data pipelines.
Over the years, several startups have tried and failed to turn DNA storage into a product. Many of their demonstrations, encoding songs or books into DNA and embedding them in artworks, generated headlines and venture capital, but few customers.
So does DNA really stand a chance of becoming a viable storage technology, and if so, when and at what cost?
Modern DNA data storage began with two landmark papers in 2012 and 2013, one from George Church’s lab at Harvard and another from Nick Goldman’s team at the European Bioinformatics Institute. They were the first to show that digital data could be reliably encoded, stored, and retrieved from synthetic DNA molecules.
Since then, academic attention has grown steadily, from around 70 papers in 2012 to more than 800 in 2024. Much of that focus, however, has been on the software side: developing codecs, optimizing storage architectures, and refining retrieval algorithms. Valuable work, but not addressing the real bottleneck of high writing cost.
Progress there has been slower, though not static. Some promising ideas are emerging: cassette-like DNA media with addressable partitions and protective coatings, and new synthesis methods that hint at lower costs. Incremental as they are, these efforts show how practical DNA storage might one day work outside the lab.
Venture capital activity tells a similar story. With fewer than 100 deals since 2012 and about $1.4 billion in total investment, DNA storage remains a rounding error compared to fields like quantum computing or fusion energy, which each attract several times that amount every year.
Nearly 80 percent of that funding has gone to just two companies, Twist Bioscience and DNA Script, both focused primarily on making DNA for biological research rather than data storage. The needs of those two domains are almost opposites: biology demands long, pure DNA strands, while storage can tolerate noise and short fragments, compensating with error correction and redundancy. Progress in one field rarely translates to the other.
Acknowledging this gap, Twist Bioscience spun off its storage division in May 2025 into a new company, Atlas Data Storage, with $155 million in seed funding, a sign that investor confidence in the concept may be bruised, but not broken.
Another encouraging signal comes from the Storage Networking Industry Association (SNIA), best known for standardizing hard-disk, SSD, and tape interfaces. SNIA now hosts the DNA Data Storage Alliance, an international consortium defining file encoding, metadata, and physical media standards for DNA, the stage where experimental technologies begin to edge toward industry.
While signals from the market and academia remain mixed, the question is no longer whether DNA storage works; it does. The real challenge is how to make it affordable.
With DNA storage’s cost and latency limitations, only a few applications make sense: cultural and scientific archives that need century-scale retention with minimal maintenance and critical knowledge preservation, such as legal frameworks, scientific records, or cultural artifacts meant to outlast current institutions.
In short, DNA storage is built for the deepest, coldest layer of the data hierarchy, information written once and read rarely.
Two puzzle pieces need to fall into place for the technology to break out of the lab: First, write costs must drop below $1 per MB, and second, investments must scale dramatically.
Even if DNA storage achieves a 1,000–10,000× cost reduction from today’s ~$100 per MB, it will remain 2-3x pricier than magnetic tape. But its advantage isn’t cost per byte, it’s cost over time.
At realistic packing densities, a cubic centimeter of DNA could hold hundreds of petabytes, the equivalent of an entire hyperscale data center condensed into the volume of a sugar cube. Once written, DNA requires no power, cooling, or maintenance. Tape libraries and hard drives, by contrast, consume constant power and need migration every 5–10 years. Over 25 years, those energy and maintenance costs often outweigh the media itself.
DNA’s zero-idle energy footprint flips that model, offering an inert, maintenance-free, and inherently green form of storage for organizations under pressure to reduce emissions. Over very long horizons, what looks expensive upfront can become the cheapest option of all.
The problem: markets and corporations aren’t wired for centuries. Investors think in quarters and governments, if lucky, in decades. That’s why public funding may determine whether DNA storage escapes the lab, rather than private investors.
There is precedent: the Human Genome Project, a global collaboration aimed at decoding the DNA that defines the human genome. Between the end of the Human Genome Project in the early 2000s and now, DNA sequencing costs fell by nearly a million-fold, largely due to the publicly funded research project. That effort took around 15 years, cost about $5 billion, and mobilized hundreds of scientists. The payoff was not just a sequenced genome, but an entire sequencing industry that made reading DNA cheap enough to be routine.
A comparable effort for DNA writing could do the same for storage, a “Human Data Project” aimed at making information preservation a public good.
The history of flash storage showed that early skepticism is not a bad sign. When flash memory debuted in the late 1980s, it was dismissed as too expensive, too small, and too fragile to matter. One megabyte of flash storage cost around $1,000, a luxury reserved for spacecraft and high-end industrial gear. But it offered something magnetic media never could: speed. By the mid-2000s, prices were down by a factor of 10,000 and performance soared. Flash memory crossed its tipping point and redefined computing from digital cameras to smartphones and AI.
DNA storage sits at a similar inflection point, just in reverse. Where flash climbed the hierarchy from slow to fast, DNA descends from long-term storage to forever storage. It will never compete with SSDs or hard disc drives on latency, but it doesn’t need to for archival storage cycles. It occupies a new foundational tier, deep-cold storage, built for data that must outlast any technology cycle.
DNA’s promise hasn’t died; it’s just waiting for the economics to catch up.
If DNA storage fulfills even a fraction of its promise, the world’s information could outlive every disk, cloud, and company that created it. What it needs is a catalyst like the Human Genome Project, and the kind of stubborn engineering persistence that turned flash storage from a 1980s curiosity into the backbone of modern computing. Along the way investors, public or private, may start to see its true potential.
Bootnote
Philipp Antkowiak is an academic researcher turned strategy consultant based in Zurich. He did his PhD thesis on DNA data storage and has closely followed academic and industry activities over the past decade.
