The Tsunami of Data

Humanity Has Already Hit Its Storage Limit Twice. Next Time Will Be Different. In the span of just a decade, humanity’s data production has surged from a relatively humble six zettabytes in 2015 to over 200 zettabytes in 2025. If you're unfamiliar with the unit, one zettabyte is a billion terabytes—and one terabyte holds about a million books. That means in just ten years, we've effectively generated data equivalent to 200 quintillion books.

Visualizing this is difficult. If we stored all that data on 3.5-inch 1TB hard drives and stacked them, the tower would stretch3.16 billion miles—13,000 times the distance from Earth to the Moon.

And the numbers are only growing. According to projections made in 2018, we are headed toward a jaw-dropping 19,267 zettabytes by 2035. That’s nearly 100 times more than today.

Image Credit: Unsplash

What's in All This Data?

The vast majority of this data is video. Although the exact figure is elusive, estimates suggest that 70–80% of all internet traffic is video—from TikTok clips to YouTube documentaries. The remainder is a mix of photos, emails, documents, enterprise and scientific data. Interestingly, scientific data—despite including outputs from giants like the Large Hadron Collider—comprises only about 1% of global data.

What’s more startling is that 90% of all data in existence was created within the past two years alone. And we’re not slowing down. The rise of generative AI has turbocharged the trend, spitting out terabytes of text, images, audio, and video with every interaction.

Our Growing Forest of Data Centers

Image Credit: Unsplash

To store this ever-expanding ocean of information, we've built over 7,000 data centers worldwide, including 1,000+ hyperscale data centers—giant facilities holding between 5,000 and 100,000 servers. These are run by the likes of Amazon, Google, and Microsoft. And we’re building over 100 more every year.

Each of these hyperscale projects costs hundreds of millions—or even billions—of dollars. But even this massive global infrastructure can’t keep up with exponential growth forever. Are we on the verge of a storage crisis?

The Real Limits to Storage

Image Credit: Unsplash

We’ve faced data storage crises before. Twice, humanity has bumped up against so-called “soft caps”—technological limitations that temporarily halted progress. The first came when magnetic hard drives reached the superparamagnetic limit. Bits became so small that thermal fluctuations began randomly flipping their polarity, corrupting data. The solution came in 2005 via perpendicular magnetic recording.

Then flash memory hit its own wall. When transistors shrank to around 20 nanometers, electrical interference again caused data loss. This was solved through a new structure called 3D NAND flash—where transistors are stacked vertically, greatly boosting capacity.

Yet with every push into smaller and denser territory, we hit new soft caps. And there's no guarantee we'll always overcome them.

The Threat of Data Decay

Even if we solve density issues, data permanence is another matter. Solid-state drives typically last 10 years before data fades. Traditional hard drives and tapes fare better, lasting up to 30 years. Some optical discs promise 100 years. But none are truly permanent. Every storage medium eventually succumbs to entropy.

That brings us to an intriguing alternative: DNA.

DNA: Nature's Ultimate Storage Medium

DNA isn’t just the blueprint for life—it's also the most efficient, compact, and durable data storage system ever discovered. In 2003, after 13 years of effort, scientists sequenced almost the full human genome, revealing that all the information needed to build a human being fits into just 3 gigabytes.

And you don’t have just one copy of this blueprint—you have it in every one of your ~30 trillion cells. That means your body is walking around with 90 zettabytes of data—nearly half of all digital data currently in existence.

Imagine: theoretically, we could store all of humanity’s data in just two human bodies, assuming, of course, we format them to remove all the biological information.

How DNA Storage Works

Image Credit: Unsplash

DNA uses four nucleotide bases—adenine (A), cytosine (C), guanine (G), and thymine (T)—as its "letters." Unlike binary storage, which uses just 0s and 1s, DNA uses a quaternary (base-4) system. It’s surprisingly easy to map binary data into DNA sequences: 00 = A, 01 = C, 10 = G, 11 = T.

To store data in DNA, scientists use DNA printers—real machines that assemble synthetic DNA by printing nucleotides, one at a time. It’s not unlike how an inkjet printer operates, except it prints in sugar, phosphate, and nitrogen rather than ink.

DNA is remarkably stable. Stored with salt in a dry ampule, DNA can last for thousands of years without electricity or environmental control.

In 2012,Harvard researchers stored a full-length book in DNA, converting all text and images into DNA code, then successfully re-extracted it with no loss. In 2019, a Boston-based startup called Catalog encoded the entire English Wikipedia, 16 GB, into a single DNA vial.

So Why Aren’t We All Using DNA Storage?

There’s just one problem: cost.

As of today, writing 1 megabyte of data to DNA costs between $1,000 and $4,000. A typical YouTube video (raw) might be 100 GB—meaning you’d need $100 million to store a single video in DNA. Needless to say, most of us are sticking with traditional hard drives for now.

But that might not always be the case.

DNA synthesis is used not just for data storage but also in pharmaceuticals, gene research, and biosecurity. That means there's major investment in making the process cheaper and faster. Costs are falling, and the pace of innovation is accelerating.

Just last year, a Chinese team invented a method to increase writing speed into DNA by 350x. Instead of writing nucleotide-by-nucleotide, they parallelized the process using pre-made templates, allowing hundreds of nucleotides to be printed simultaneously. In a proof-of-concept, they encoded and retrieved a digital image of a tiger using this technique.

The Future of Data Is Biological

It’s hard to believe that something as ancient and organic as DNA could be the future of data storage. But history often favors elegance over brute force. DNA offers unmatched density, energy efficiency, and longevity.

With the cost of synthesis dropping and scalability on the horizon, we may one day abandon server farms in favor of genome vials. No more humming fans or massive data center cooling bills. Just vials of data, quietly sitting on a shelf—for thousands of years.

The idea sounds like science fiction. But it's already real. And if our current trajectory of data creation continues, we may not have a choice but to embrace it.