5D glass “memory crystals” provide up to 13.8 billion years of data storage life, which is also the age of the universe – packing 360 terabytes into a 5-inch glass disk using a femtosecond laser
SPhotonix says it has moved its so-called 5D Memory Crystal technology out of the lab and closer to real-world implementation, laying out plans to pilot glass-based refrigeration systems in data centers over the next two years, according to comments made during an interview with Register. The British start-up, which originated from research at the University of Southampton and was founded in 2024, made the announcement, providing details of its first round of external funding.
The company’s data carrier is a fused-silica glass plate written using a femtosecond laser that encodes data into nanoscale structures. The information is stored in five dimensions: three spatial coordinates (x, y, z) and the orientation and intensity of the nanostructures, which are read optically using polarized lithe. SPhotonix claims that a single 5-inch glass drive can hold up to 360TB of data, and the media is designed to be stable for 13.8 billion years – the estimated age of the universe – assuming no external mishaps occur along the way.
According to SPhotonix, its current prototypes achieve write speeds of around 4 MB/s and read speeds of around 30 MB/s. These numbers place the technology well below current archiving systems, but the company has published a roadmap that assumes sustained read and write speeds of 500 MB/s over the next three to four years.
The company estimates that initial system costs will be approximately $30,000 for the writer and $6,000 for the reader, with the reader expected to be field deployed in approximately 18 months. SPhotonix says it has raised $4.5 million to date and is currently working to advance from Technology Readiness Level (TRL) 5 to TRL 6, which typically involves validation in appropriate operational environments rather than controlled laboratory settings.
In terms of durability, the company describes the media as inherently isolated, requiring no power to store data and suitable for archives where access latencies of 10 seconds or more are acceptable.
SPhotonix is not alone in its pursuit of non-magnetic frosty storage. Microsoft has publicly tested glass media through Project Silica, while other startups such as Cerabyte are promoting ceramic-based alternatives for robotic library systems. What distinguishes SPhotonix’s approach is its focus on licensing the media and optical platform into existing data center architectures, rather than building an end-to-end data storage service.
Whether SPhotonix’s 5D glass can go from impressive density demonstrations to competitive system-level performance will determine whether it becomes a niche archival medium or a viable data storage solution in newfangled data centers.
