Cerabyte: Shaping the Future of Sustainable Long-Term Data Storage

Cerabyte: Shaping the Future of Sustainable Long-Term Data Storage

Long before the recent surge in machine learning, humanity had already developed an insatiable appetite for data. Streaming, social networks, computer simulations – with humanity producing more and more data, data storage has become an important but often overlooked part of data processing.

The current use of hard disk drives for data storage leads to the consumption of a considerable amount of electricity, thereby contributing to carbon emissions and – due to the short service life – huge amounts of electronic waste. At the same time, most of the data is stored for long-term purposes and is ‘cold’ – meaning it’s rarely, if ever, retrieved. 

Cerabyte pioneers a new approach to long-term data storage, using laser pulses to write patterns into ceramic nano-coatings on a glass substrate and thereby imprinting data practically forever. This isn’t a scene from a sci-fi movie, but how their ceramic storage technology works under the hood. And it is recyclable, reduces energy consumption by two orders of magnitude, and massively cuts down the cost of data ownership. 

Founded by Christian Pflaum, Martin Kunze, and Alexander Pflaum in 2022, Cerabyte went through the Intel Ignite accelerator program and raised a seed round of $7M the same year, led by Daniel Aegerter and with Porsche’s Sajjad Khan becoming chairman of the board. 

Learn more about the future of sustainable long-term data storage from our interview with the co-founder and CFO, Alexander Pflaum

Why Did You Start Cerabyte?

Cerabyte came about when Christian and I met Martin, who was working on storing information forever on ceramics for an art project. However, when we looked at the market for long-term data storage, we learned about the scale of the problem that nobody seemed to be tackling – and that ceramic storage technology could be a once-in-a-lifetime chance to change how the world stores information. 

More than 70% of data is cold data. It is practically never retrieved but stored for archival purposes on current-day storage media like hard disks that must be replaced every five to ten years. This not only leads to considerable capital expenditure, but the electricity bill for operating hard disks also results in considerable operating expenses.

Ceramic data storage may last virtually forever, withstanding temperatures close to absolute zero to hundreds of degrees and being resistant to electromagnetic pulses and radiation. It is also cheaper than other long-term data storage solutions and avoids rare earths and complicated moving parts, as we just need ceramic nano-coating and glass, which we can produce in large quantities. Our technology is thus very scalable and works reliably – there’s no merit to having a great idea or technology if you can’t scale it. 

How Does Ceramic Data Storage Work?

Cuneiform tablets are like a 5000-year-old proof-of-concept for ceramic data storage: you can still read them today! But unlike the old Babylonians, we don’t need chisels but instead state-of-the-art laser technology.

We use femtosecond lasers to write information swiftly. The physics is simple: Take a glass substrate, coat its surface with nanolayers of a ceramic that absorbs a lot of energy, and then bombard it with ultrashort laser pulses. Once a laser pulse hits the ceramic coating, it leads to a Coulomb explosion and produces a hole in it. By having a hole or no hole, we can encode binary information. And we can even use a matrix of multiple laser beams at once and stamp data like QR codes in the ceramic coating. 

The data is read using ultra-fast LED illumination and by detecting whether a given pixel has a coating or not. Thanks to the large number of high-resolution image sensors produced for the smartphone industry, the readout can be achieved at a very low cost. And with every new generation, our technology gets faster, the pixel size decreases, and we can scale by leveraging existing semiconductor technology.

Of course, as with all optical media, you can’t make patterns arbitrarily small. If smaller than about half a wavelength, you can’t resolve distinct structures anymore, and they blur into each other. Yet, we have found some tricks to overcome this diffraction limit by using structured illumination and image processing to turn blurry images into ones from which we can read information.

Venture capitalists often ask founders the “why now?” question, which translates to: if you believe this is a smart idea, why has no one else done it before? As for ceramic data storage, we first had to develop the right kind of machinery to make encoding and decoding information on top of glass fast enough to be a match for existing computing infrastructure – and only since two years has this become technologically feasible.

The main goal when developing our technology is to make data storage really, really cost-efficient. We’ve been talking extensively to hyperscalers and learned that long-term data storage is primarily a cost game. Hyperscalers want to offer storage at the lowest cost as they make their profit on compute: you do compute where all of your data is, so it’s of strategic importance to bring all the data in one place. 

While our primary objective is to deliver cost savings, we also work on improving performance. We’re making writing and reading data much faster in the future – thanks to our modular approach, we can add reading devices depending on retrieval rates and how short the required reading time needs to be. And our approach comes naturally with resilience to harsh conditions – even when we tortured our data carriers with extreme temperatures or radiation, we didn’t lose a single bit.

How Did You Evaluate Your Startup Idea?

The market for long-term data storage is giant, so no one ever doubts that. The question is more about go-to-market and how to capture it. 

Our patent portfolio is key to capturing this opportunity, covering about 80% of global GDP. It puts us in a leading position to develop a global standard for long-term data storage – and once there is a global standard at an attractive price level, this will unlock crazy demand. 

Our technology has garnered high interest as we recently came out of stealth. And once the big players in cloud computing realize that ceramic data storage is the future, they will order massive amounts, and we’ll partner with third parties to unlock more production capacity to meet demand. It’s then all about orchestrating the value chain – and that’s why we built a business development team with lots of experience in producing large quantities of physical products.

Finally, we will also be able to support large scientific experiments or high-performance computing applications at supercomputer labs around the world. Running these scientific experiments costs a fortune, so you want to gather all the data in a short period of time and store it practically forever – to evaluate it even decades down the line if you need to test a new hypothesis. Cerabyte could come in as second-tier storage: at first, you use flash storage as fast, intermediate storage, buffering data until it’s written to Cerabyte storage, so you can then overwrite the primary flash storage. 

What Advice Would You Give Fellow Deep Tech Founders?

After founding two startups before, I learned that there are thousands of answers to this question. As a founder, you’re responsible for everything: one day, you need to talk to investors, and the next day, you have to get your financials in order or solve some technical issues. So, most importantly, it’s about learning each new day. 

Your biggest asset as a startup is that you can be fast, make rapid changes to your strategy, or pivot into an entirely different domain. You need the mental flexibility to cope with the challenges that pop up along the way and adjust accordingly to what the market demands. We’re very happy to have an amazing, diverse, and strong team to execute our vision for Cerabyte: Becoming the global standard for long-term data storage so that people need just one more data migration: their last one to Cerabyte.

Learn More About Cerabyte*

If you’re an enterprise, research institution, or institutional investor and interested in working with Cerabyte, feel free to reach out to them at office@cerabyte.com.

*This section is sponsored by Cerabyte – we greatly appreciate the support!