There’s something almost absurd about the idea. Every email you’ve ever written, every grainy phone video, every spreadsheet sitting forgotten on a corporate server somewhere in Virginia — all of it, in theory, fitting inside a coffee cup. Not a metaphorical coffee cup. A real one. The kind sitting next to your keyboard right now.
That’s the promise scientists have been chasing for years, and last month a team at the University of Missouri pushed it a little closer to reality. Their work, published in PNAS Nexus, takes a stab at one of the field’s most stubborn problems: once you write information into DNA, you can’t change it. It’s permanent. Useful for archives, maybe, but useless for almost everything else we expect modern storage to do.
| Topic Profile: DNA Data Storage | Details |
|---|---|
| Field of Research | Synthetic biology, molecular computing, biomedical engineering |
| Recent Breakthrough Institution | University of Missouri, College of Engineering |
| Lead Researcher | Li-Qun “Andrew” Gu, Professor of Chemical and Biomedical Engineering |
| Published In | PNAS Nexus (2025) |
| Key Industry Voice | Mark Bathe, PhD, MIT Professor of Biological Engineering |
| Storage Density | All of humanity’s data, theoretically, in something close to a coffee cup |
| Estimated Global Data by End of 2025 | 33 zettabytes (3.3 × 10²²) |
| Daily Data Added Worldwide | 2.5 million gigabytes |
| Current Synthesis Cost | Roughly $1 trillion per petabyte |
| Longevity Under Cool, Dry Storage | Several thousand years |
| Notable Milestones | 1999 (23-character message), 2013 (739 KB), 2018 (200+ MB), 2025 Library of Congress grant for 1.5 GB encoding |
| Energy Reduction Potential | Up to three orders of magnitude vs. tape storage |
| Building Blocks | Four bases — A, T, G, C |
Li-Qun Gu, the chemical and biomedical engineering professor leading the Mizzou research, has been working on this puzzle for a while. He talks about DNA the way some people talk about old vinyl records — with a kind of quiet awe at how something so small could carry so much. His team figured out a way to make the molecule rewritable at a faster, simpler, more efficient pace than what came before. Whether that scales is another question. It usually is.
The reason any of this matters comes down to numbers that don’t quite feel real. By the end of this year, humanity will have generated roughly 33 zettabytes of data — a figure with twenty-two zeros after it. Every single day adds another 2.5 million gigabytes to the pile. Data centers are running out of room. They’re running out of power, too. The Department of Energy expects their electricity demand to triple by 2028, eating up something like twelve percent of all U.S. power. That’s not sustainable. Most people in the industry know it isn’t.
DNA, as a storage medium, is almost embarrassingly good at this job. It’s dense beyond comparison. It’s stable for thousands of years if you keep it cool and dry. Once you’ve made the polymer, it sits there, costing nothing, drawing no current. Mark Bathe at MIT has been blunt about it for years. There isn’t another contender, he says. Nature picked the molecule for a reason. The molecule works.
And yet. The cost still hovers at something like a trillion dollars per petabyte, which is the kind of number that makes investors politely change the subject. A handful of start-ups — Atlas Data Storage among them — are betting they can drag that price down through better synthesis, better sequencing, smarter chemistry. Varun Mehta, who runs Atlas, compares traditional storage to painting the Golden Gate Bridge. By the time you finish one end, the other end is already rusting. You start over. Forever.
Watching the field evolve, it’s hard not to notice how quietly it’s all happening. The Library of Congress quietly funded a 1.5-gigabyte DNA encoding project earlier this year. Microsoft has been pouring money into AI infrastructure that will, eventually, demand storage solutions current technology can’t deliver. Shakespeare’s complete sonnets, eight of his tragedies, the entire English Wikipedia — researchers have already squeezed all of that into containers smaller than a test tube. Not compressed in the conventional sense. Translated, base by base, into the same alphabet that built every living thing on Earth.
There’s a feeling, watching this unfold, that we’re stumbling toward something genuinely strange. The future of the internet might not live in a Nevada data center after all. It could reside in a refrigerator. or a cup of tea. Or even smaller, if Gu and his associates don’t give up.
