PCR Tech and IT retail, distribution and vendor news
For years, the storage industry has relied on familiar assumptions with falling prices, rising density, and steady gains under Moore’s Law. But Dr Steffen Klewitz, Founder and CEO of Ewigbyte, has warned that these foundations are eroding, and urged the channel to prepare for a fundamental shift in how data is stored, priced, and protected.
“We have a gap,” Klewitz said. “We are using technology which is largely over 70 years old, and if you look at LTO hard drives, for example, they will probably not be able to cover it.” He was, of course, referring to the mismatch between projected data growth and the production capacity of existing storage media.
Even in optimistic scenarios, he noted, manufacturing growth rates for HDD, SSD, and tape hover around 27 per cent. Data growth, by contrast, is expected to exceed 40 per cent annually, even before accounting for the additional burden of AI workloads. “AI will definitely add more,” he added. “We don’t know how much, but the amount of data you need to store for documenting your learning models is much more than the data AI itself generates.”
This imbalance, Klewitz argued, will force the industry to confront the end of a long-held belief: that storage prices inevitably decline. “This is Moore’s Law,” he said, “and it has worked for over 30 years.” But Moore’s Law itself, he noted, has already faltered in semiconductors. “Once you reach the atomic scale, quantum effects take over, and the electronics you know in the analogue space will not work any longer – so Moore’s Law has come to an end because of the physical size of the chip. The implication, then, is that the same applies in the storage space, where you need to think about whether we can actually store more data in the same volume.”
The consequences are already visible. Klewitz pointed to analyses showing consumer prices for 10-terabyte hard drives rising, with enterprise-grade drives having 52-week lead times. “We also need to exchange the media in data centres every five to seven years,” he claimed. “We are creating, thereby, over 350,000 tonnes, minimum conservative estimate, of electronic waste. This can’t be efficient.”
Beyond economics, Klewitz highlighted new risks to data integrity. He cited threats ranging from AI-driven data alteration to geopolitical manipulation, as well as natural and man-made hazards such as solar flares and electromagnetic pulses. “If you ignite a nuclear warhead in orbit, you create an electromagnetic pulse, and within the cone of this pulse, all electronics are compromised – especially magnetic because you cannot shield magnetic fields,” he explained. “So, all the information is at least partly damaged, if not completely deleted.”
Against this backdrop, Ewigbyte is developing a radically different approach: storing data on glass surfaces using ultra-short‑pulse lasers. “We store data on the glass surface by ablating the information into the glass,” Klewitz said. “It’s similar to a punch card, but confined to the surface, otherwise you pay in speed.”
The process involves using a laser beam to inscribe QR codes directly into the glass. Each laser shot can embed several kilobytes of data, and with repetition rates of up to 60,000 pulses per second, Klewitz projected write speeds of around 500 megabytes per second per head. With multiple heads in a rack, throughput could reach four gigabytes per second. Reading is performed optically, using microscopes and cameras to capture the QR codes in parallel. “We etch the glass on the surface, and the QR code or the DPD code is, of course, generic,” Klewitz explained. “We will not try to lock in a customer with our proprietary code.”
The durability of the glass is central to the proposition. Unlike magnetic or solid-state media, which degrade over years, glass promises lifespans measured in millennia. Klewitz described ongoing research into identifying the “ideal glass” among tens of thousands of variants, with ageing simulations designed to ensure stability for 10,000 years. “There are about 50,000 different sorts of glass, and it is crucial for us to find the ideal glass,” he said. “This is a typical industrial development process – you start with the available glass, you find that it does not have the optimal parameters, and then you do research, and over time, you will develop better glass.”
Security, too, is embedded at the physical layer. Once a bit is written into the glass, it cannot be erased or altered without leaving visible evidence. “We bring the whole security thing down to the physical layer,” Klewitz said. “Once this is written, you have these holes where you have your bits. It is impossible to close a hole on a glass surface without leaving evidence.” He described plans to incorporate holograms into the glass as proof of authenticity, with serial numbers tied into checksum algorithms. “Once this hologram is in, it’s like money,” he said. “Suddenly we have something, and it’s an absolute original, physical, valuable piece of glass.”
Ewigbyte intends to operate its own facilities and offer storage as a service. “We don’t want to sell hardware,” Klewitz said. “We want to sell data storage as a service. We want to develop our own proprietary hardware and run this in our own way.” By housing racks in dedicated warehouses rather than conventional data centres, the company aims to eliminate cooling and humidity control costs. “When you have plenty of space that you don’t have to cool, you only have to control against physical attacks.”
Redundancy will be achieved using familiar techniques, but adapted to the new medium. Klewitz described block-level architectures similar to RAID, with physical glass carriers serving as replication units. “The logic is the same,” he explained. “You can implement RAID by writing two identical carriers, then add a third with XOR parity to create RAID 3. Even if one carrier fails, the data remains recoverable. This can be done at the block level or at the carrier level. The principles are unchanged – the difference is that redundancy is now applied physically, whereas in hard drives it was logical.”
Klewitz positioned Ewigbyte’s solution as a complement to existing object storage infrastructures, designed to address the challenge of cold data. “Organisations using today’s object storage face this issue,” he explained. “They still need hardware to manage storage. We have very interested people saying that if we can deliver the service, they could integrate us through an S3-style interface within their infrastructure, routing hot data to the specialist providers who specialise in it, and directing cold data to us.”
The vision extends to large-scale automated warehouses, where cartridges of glass media could be stored and retrieved by robotics. Klewitz has visited modern logistics facilities capable of moving hundreds of items per second, and he suggested similar automation could underpin petabyte-scale glass archives. “Ultimately, we will have a data density of six petabytes per cubic metre, which is possible,” he said. “We will have our own specialised warehouse, and in the core of this warehouse is our writing and reading facility. And then we just need to manage the logistics process.”
For the global market, Klewitz predicted that optical storage will squeeze out tape and relegate HDDs to niche roles. “My prediction is that the 70-year-old technology of tape will be squeezed out first,” he said. “And the second thing is that HDD is becoming a niche market.” Ultimately, he argued, the decisive metric will remain cost per terabyte. “If you can deliver something better than tape for a dollar per terabyte, you will win.”
Klewitz acknowledged that costs are currently uncompetitive for a startup, but he projected break-even within five to ten years, depending on advances in glass and learning curves. He also noted that specific market sectors, such as government agencies safeguarding irreplaceable data, may pay premiums for the assurance of immutability and resilience against electromagnetic pulses. “There are sectors that may decide they need to keep the data forever,” he said. “If you are able store data forever, as well as protect against physical attack, then it will happen.”
For the channel, long-held assumptions about storage economics and technology are under strain, and new paradigms are emerging. Glass-based storage may not replace existing media overnight, but its promise of permanence, security, and sustainability positions it as a contender in the evolving landscape of cold data. As Klewitz said: “This is a black swan event. This is the thing where we need to be different and try to develop a business model that relies more on data speed than on density.”
