DNA Drives, Glass Discs & 5D Crystals: The Race to Store Data Forever

by Marcin Frąckiewicz
in Uncategorized
on 7 August 2025

NVIDIA Drive Thor vs Tesla FSD HW 4 vs Snapdragon Ride Flex

Our world is generating data at explosive rates – far more than today’s hard drives, flash SSDs, and magnetic tapes can economically keep up with datacenterdynamics.com fraunhofer.de. Even if we could build enough of those devices, they wear out quickly (hard drives last ~5 years, tapes 15–30 years under ideal conditions) datacenterdynamics.com, forcing costly data migrations to avoid loss. This has sparked a race to find radically new storage technologies that can pack more data into less space and preserve it for centuries or longer. Three of the most promising contenders are:

Synthetic DNA Data Storage – encoding digital bits into the A/C/G/T code of artificial DNA molecules.
Microsoft’s Project Silica – using ultrafast lasers to encode data in quartz glass platters.
5D Nanostructured Glass (“Superman Memory Crystal”) – storing information in microscopic nanogratings inside glass, with five dimensions of data encoding.

Each of these futuristic media claims ultra-high density, extreme durability, and millennia-long lifespans, but they differ in how they work and where they stand on the path to practical use. Below we compare these technologies on how they work, their storage capacity, longevity, environmental resilience, energy efficiency, speed, current progress (including 2024–2025 breakthroughs), and real-world use cases – with insights from experts leading the charge.

Synthetic DNA Data Storage

Synthetic DNA storage is literally about storing digital files in molecules of DNA – the same type of DNA that carries genetic information, but manufactured in a lab. It sounds like science fiction, but research teams and startups have made rapid strides in making DNA “memory drives” a reality.

How DNA Data Storage Works

Data encoding: Digital data (binary 0s and 1s) is translated into sequences of the four DNA bases (A, C, G, T). For example, one simple scheme maps 00→A, 01→C, 10→G, 11→T blocksandfiles.com. This yields a collection of synthetic DNA strands representing the file. At a high level, we convert bits to bases microsoft.com, creating a digital DNA sequence.

Writing (synthesis): Specialized machines then chemically synthesize those DNA strands, essentially “printing” the sequence onto molecules microsoft.com. Early DNA writing is slow and done in small test tubes, but new approaches use microchips with massively parallel reactions to scale up throughput fraunhofer.de fraunhofer.de. For example, a recent project uses a silicon chip with thousands of tiny reaction wells and embedded controls to produce DNA at high speed, dramatically miniaturizing what used to be room-sized synthesizer setups fraunhofer.de.

Storage: Once written, the DNA molecules (each just a few hundred bases long) are preserved in an inert form – often dried and sealed in tiny containers (like microscopic beads or capsules) to protect them blocksandfiles.com. DNA is remarkably stable when kept cold, dry, and dark. In fact, nature’s proof of concept is ancient DNA: scientists have recovered readable DNA from a 700,000-year-old permafrost-preserved horse bone and even from million-year-old mammoth remains techxplore.com. Properly stored synthetic DNA could similarly last centuries or more without degradation.

Reading (sequencing): To retrieve data, the DNA is sequenced – the same process used in biology to read genetic code. Modern sequencing machines can read millions of DNA strands in parallel, outputting the sequences of A/C/G/T which are then decoded back into the original digital data microsoft.com microsoft.com. Error-correcting algorithms (many adapted from telecom and storage industries) handle the fact that biological processes aren’t perfect, dealing with any missing or garbled fragments microsoft.com microsoft.com. Essentially, DNA storage involves a write cycle (synthesis) and a read cycle (sequencing), with the data dormant in molecular form in between.

Storage Density & Capacity

DNA’s data density is astounding. In theory, DNA can pack orders of magnitude more information per volume than any silicon chip or magnetic tape. Researchers have demonstrated data encodings approaching 215 petabytes per gram of DNA en.wikipedia.org – that’s about 215 million gigabytes in a droplet of DNA the size of a sugar cube. In practical terms, “10 full-length movies could fit in a volume the size of a grain of salt” via DNA, according to the industry’s DNA Storage Alliance blocksandfiles.com. And one analysis suggests 1 gram of DNA could hold nearly 1 zettabyte of data (a trillion gigabytes) under optimal conditions blocksandfiles.com.

This unmatched density means an entire data center’s archive could theoretically be held in the palm of your hand. “DNA is extremely dense – it holds far, far more information per unit volume, per unit mass than any storage media we have today,” notes Microsoft researcher Dr. Karin Strauss, who has led DNA storage research microsoft.com. It dwarfs even advanced optical media: a DNA archive could reduce a room-sized tape library to a test tube. Density is DNA’s killer feature, and one reason so many labs are exploring it.

Longevity & Durability

Data written in DNA can last for millennia. As a molecular storage medium, DNA has no moving parts and doesn’t require electricity to sit on a shelf. If kept in stable conditions, it won’t “bit rot” the way tapes demagnetize or DVDs fade. In fact, preserved DNA in fossils has lasted hundreds of thousands of years in nature techxplore.com. For man-made storage purposes, companies are confidently claiming at least 1,000-year durability for DNA data archives prnewswire.com.

DNA’s robustness comes from its very nature: it evolved to store information reliably. It can also be encapsulated for extra protection. Researchers have tested DNA storage under harsh conditions – high humidity, temperature swings, even radiation – and found it quite resilient microsoft.com. One Microsoft study even showed that dried DNA strands, when encapsulated in glass, survived neutron radiation bombardment without data loss microsoft.com, highlighting its potential to weather environmental extremes.

However, DNA is an organic molecule, so it does have vulnerabilities. High heat (beyond ~150°C) or harsh chemicals can damage it. Thus, while a quartz glass platter might survive a fire intact, DNA samples would combust at extreme temperatures. For real-world longevity, DNA archives would be stored in climate-controlled vaults (or at least in cool, dry, dark conditions). Fortunately, keeping DNA cold and dry is much easier than, say, spinning up hard drives every few years. With proper care, DNA might outlast any other storage medium we’ve created – potentially tens of thousands of years or more, truly outliving generations of technologies prnewswire.com.

Environmental Resilience and Sustainability

DNA storage scores high on environmental resilience and eco-friendliness. Since the “media” is just molecules, it is immune to electromagnetic disruptions – there’s no concern about magnets, EMPs, or cosmic rays flipping bits (cosmic radiation doesn’t alter inert DNA significantly southampton.ac.uk). Encapsulated DNA can handle shocks and vibrations that would crash a hard disk. And unlike disk or tape, temperature and humidity fluctuations are not catastrophic (though for extreme longevity, stable conditions are best). Essentially, if you seal DNA data in a dry capsule, it could be buried, frozen, or sent to space and still be readable when retrieved.

Critically, DNA storage can be extremely energy-efficient over the long term. Once data is written in DNA, it requires no power at all to store – you can put the vial on a shelf with zero energy consumption for decades. There’s no need to refresh or migrate the bits periodically as with magnetic media that degrade datacenterdynamics.com. “DNA storage will enable greener data centers with permanent storage that requires no ongoing migration or rewriting, reduced power demands, and minimized carbon impact and e-waste,” says Varun Mehta, CEO of recent DNA storage startup Atlas prnewswire.com. This is a big motivation: data centers currently gobble up around 3% of global electricity techxplore.com, in part due to constantly cooling storage devices and duplicating data for safety. DNA archives could sit safely in a closet, no electricity or air-conditioning needed, massively shrinking the carbon footprint of long-term data preservation techxplore.com.

On the flip side, the writing and reading processes for DNA are chemical and biochemical, which do consume materials and energy (reagents, enzymes, etc.). There is also a waste consideration: synthesizing DNA uses laboratory chemicals, though efforts are underway to make these processes more sustainable (e.g. enzymatic DNA synthesis uses water-based reactions rather than harsh chemicals). Still, compared to manufacturing silicon chips or running tape libraries with constant power, DNA’s resource use at scale could be much lower. And the medium itself – DNA – is biodegradable and non-toxic, eliminating e-waste concerns. The vision is that future data archives might look more like biological labs than server farms, potentially easing the environmental burden of our data growth.

Read/Write Speed Challenges

If DNA has near-perfect density and longevity, where is the catch? In a word: speed. Today’s DNA data I/O (input/output) is orders of magnitude slower than conventional storage. Writing even a few megabytes into DNA can take hours or days with current synthesis tech, and reading (sequencing) is similarly lengthy. For example, a 2022 demonstration found it cost thousands of dollars and took many hours to encode just 2 MB of data into DNA and then read it back en.wikipedia.org. Another recent report noted that retrieving ~100 MB from a DNA archive could take several days with earlier methods techxplore.com techxplore.com. Clearly, DNA is not plug-and-play fast like a USB drive – it’s more akin to sending your data to a slow molecular scribe and later running a lab test to get it back.

Researchers are tackling this speed bottleneck with parallelism and automation. Microsoft and University of Washington built a prototype DNA storage device in 2019 that automated the full process (bit-to-DNA and back) for a few bytes, using microfluidic chips to move liquids around – essentially a DNA “data writer” robot microsoft.com microsoft.com. New companies like Catalog have designed machines to encode data by mixing pre-made DNA fragments in different combinations instead of synthesizing each bit from scratch, massively speeding up writes. In one prototype, Catalog showed a machine (about the size of a fridge) that could encode kilobytes per second range, which is a huge improvement, though still far off from electronic speeds.

On the reading side, high-throughput sequencers can actually read billions of bases per run, so in theory one could retrieve gigabytes of data in hours. The bottleneck is often computational – sorting out the data from massive parallel reads. In 2025, a team at Technion Israel demonstrated an AI-driven decoder (“DNAformer”) that cut DNA data retrieval time by 3,200×, reducing what took days down to 10 minutes for a 100 MB dataset techxplore.com techxplore.com. This was achieved by training machine learning models to more quickly correct sequencing errors and reconstruct the file. Such advances show promise that DNA readback speeds can improve dramatically, perhaps approaching real-time for modest amounts of data.

Even so, DNA will never match DRAM or SSD latencies – think minutes to hours to get your data, not microseconds. This is why DNA is envisioned mainly for “cold” archival storage – data you stash away and rarely need to fetch. As Twist Bioscience CEO Emily Leproust has quipped, the first use of DNA storage might be as an “offline” service: you send data in, they store it on DNA, and when you need it back it’s delivered somewhat like film processing. Indeed, Twist Bioscience (a leader in DNA synthesis) was planning an early-access DNA storage service by 2025, targeting customers who need to archive huge datasets and can tolerate longer retrieval times (according to Leproust in a 2022 interview).

In summary, speed is the main hurdle keeping DNA storage in the lab today. But with ongoing R&D, the process is speeding up every year. The expectation is that for pure archival scenarios (think: write once, read maybe years later), DNA’s density and durability trade-off against speed is acceptable. A quote often cited: “If humanity ever faces a scenario of leaving Earth, the megabytes-per-gram metric [of DNA] is going to matter” – meaning the sheer density could justify the slower access when moving planetary amounts of data datacenterdynamics.com. For now, though, DNA storage is not about instantaneous access – it’s about writing once, reading maybe much later.

Current Developments (2024–2025)

The past two years have seen major leaps and investments in DNA data storage, signaling the transition from pure research toward commercialization:

Standards and Alliances: In March 2024, the DNA Data Storage Alliance (founded by Microsoft, Illumina, Western Digital, and Twist) released the first official specification for DNA storage, defining how to encode basic file system info in DNA (the “Sector 0/1” spec) snia.org. This is a first step toward interoperability – ensuring different DNA storage technologies can read each other’s data. The Alliance has also teamed with the Storage Networking Industry Association (SNIA) to develop standards, “signaling a maturity of DNA data storage tech” and aiming to build an ecosystem by later this decade blocksandfiles.com.
Big Funding and Spin-offs: In May 2025, Twist Bioscience – a key player in DNA synthesis – spun off its DNA storage division into a new company, Atlas Data Storage, with a hefty $155 million seed funding prnewswire.com. Atlas is tasked with commercializing DNA storage for data centers, combining Twist’s semiconductor-based DNA writing tech with new enzyme techniques prnewswire.com. The leadership includes veterans from the data storage industry, and their goal is to deliver “ultra-high density, secure, permanent data storage” via DNA, targeting hyperscale cloud providers and governments prnewswire.com prnewswire.com. This spin-out and investment show that industry believes DNA storage will soon be more than a lab curiosity. “Now is the time to build on the growing momentum and scale up the ecosystem to meet demand within this decade,” said Twist’s DNA storage director Steffen Hellmold blocksandfiles.com.
Prototypes and Demos: European researchers in the EU-funded BIOSYNTH project unveiled in 2025 a microchip-based DNA synthesizer prototype, shown at the SynBioBeta conference fraunhofer.de. This device integrates CMOS electronics, microfluidics, and micro-heaters to drastically increase writing throughput while shrinking cost and energy use fraunhofer.de fraunhofer.de. It aims to replace room-sized DNA labs with a portable, low-energy chip that can write DNA data, bringing commercial DNA storage devices closer to reality fraunhofer.de. Meanwhile, companies like Catalog have demonstrated automated DNA writing machines (one image shows a prototype that looks like a large printer blocksandfiles.com), and France’s startup Biomemory (backed by Illumina) is pursuing enzymatic DNA synthesis to speed up data writing by 100x. The U.S. IARPA (intelligence research agency) has a program (MIST) funding several DNA storage efforts, further pushing the tech.
Use-Case Trials: Gartner analysts predicted that by 2024, 30% of digital businesses would pilot DNA storage for archiving. While exact numbers are hard to confirm, interest is indeed rising. For example, the US Library of Congress in 2023–24 explored the feasibility of DNA for long-term preservation of critical records (they issued calls for research on it). Media and entertainment companies (who sit on huge archives of video) have also shown interest – a 2022 project with the Eurovision Song Contest archived some music videos in DNA as a proof of concept. All these moves suggest that DNA storage is shifting from academia to industry, with a tentative roadmap aiming for initial commercial services in the late 2020s.

Despite the excitement, there are also skeptics. Notably, Microsoft itself (once a founding member of the DNA Alliance) has recently dialed back its DNA efforts. “We stopped – they’re just not meeting the orders of magnitude gain they thought… Proponents touted the extreme density, but it’s not clear that that’s relevant,” said Microsoft’s storage architect Richard Black in 2024 datacenterdynamics.com. Microsoft concluded that DNA’s advantages might only matter in far-future scenarios (like interstellar travel) and in the near term other tech (like glass storage) could be more practical datacenterdynamics.com. This candid take underscores that DNA storage still faces significant R&D hurdles. However, many others in the field disagree with Microsoft’s pessimism – pointing out that data growth is relentless and that the need for DNA’s density will arrive sooner or later, especially as costs come down. As one researcher joked, “the data storage crisis won’t wait for a supernova” – meaning we’ll need DNA’s capacity on Earth well before we’re building spaceship archives.

Potential Use Cases and Outlook

Given its characteristics, DNA storage is poised for very-long-term archival applications. Some envisioned use cases in the next decade include:

Cold Archives for Cloud & Enterprises: Large cloud providers or institutions could use DNA to store “cold” data – backups, compliance records, video archives – that rarely need access but must be kept potentially forever. For example, a hospital could archive patient scans and records in DNA knowing they’ll be safe for 100+ years; media companies could save master films or TV footage without fear of digital decay. Microsoft’s research showed DNA is well-suited for archival data that is written once and read infrequently microsoft.com.
National Archives and Cultural Repositories: Government archives, libraries, and museums are interested in DNA to preserve historical and cultural data. Imagine national libraries encoding all their digital books, or UNESCO saving world heritage documents in DNA capsules to bury in a time vault. DNA’s small physical footprint means even petabytes of national records could be stored off-site (or even off-planet) as insurance copies.
Scientific and Medical Data: Projects that generate massive datasets (genomics, astronomy, particle physics) struggle to store all the data long-term. DNA storage could allow archiving everything rather than discarding data. There’s a poetic fit for genomics labs to store genomic data in DNA itself. In fact, in 2021 an art/science project archived the *full text of the Wikipedia and COVID-19 research data into DNA strands, then physically integrated them into synthetic fossils for future discovery.
Disaster-Proof Backups: DNA’s durability makes it attractive for doomsday-proof backups. There are concepts of storing DNA archives in deep ice cores, caves, or even launching them into space. For instance, one startup is exploring storing data in DNA, encapsulated in tiny glass spheres, and sprinkling them into the stratosphere – an “aerosol library” that could survive a global catastrophe. While far-fetched, it underscores the appeal of DNA as a truly long-term memory of civilization.

The future outlook for DNA storage depends on solving the cost and speed issues. The cost per megabyte is still very high (though plummeting each year as DNA synthesis technology improves). Experts are optimistic that within 5–10 years, costs will drop enough that DNA storage becomes economically viable for specialized archiving (perhaps on the order of $100 per terabyte stored, which would compete with tape libraries considering DNA needs no maintenance). If those trends hold, by the 2030s we may see DNA data vaults as part of mainstream data centers.

As Dr. George Kadifa of Atlas Data Storage puts it, “I’m confident that the data storage technology Atlas is creating will enable storing billions and billions of terabytes at low cost, power, and waste”, which would have “an immense long-term economic and national security impact” prnewswire.com. In other words, if DNA storage reaches its potential, it could fundamentally reshape how we think about preserving information – we might finally have a medium where nothing ever has to be deleted due to space or decay. That is the revolutionary promise driving synthetic DNA data storage research forward.

Microsoft Project Silica (Quartz Glass Storage)

While DNA storage takes inspiration from biology, Project Silica is building on one of humanity’s oldest materials: glass. Microsoft’s Project Silica is developing a system to store data inside tough quartz glass plates using ultrafast lasers. The vision is to create “etchings in glass” that can hold immense amounts of data, be read optically, and remain stable for many thousands of years. Microsoft refers to it as the first archival storage technology “built from the ground up for the cloud” – essentially, a replacement for magnetic tape in cloud data centers, with huge longevity benefits microsoft.com.

How It Works: Lasers Writing in Glass

Project Silica uses femtosecond lasers (lasers emitting incredibly short pulses, on the order of 10^-15 seconds) to inscribe microscopic patterns inside a thin slab of quartz glass. When these focused laser pulses hit the glass, they produce tiny “voxels” (3D pixels) – essentially nanoscopic damage points or gratings in the glass structure. By varying the laser’s intensity, polarization, and position in the X, Y, and Z axes of the glass, each voxel can encode multiple bits of information. This is often described as “5D” storage because there are 5 degrees of freedom for encoding data: three spatial coordinates (width, height, depth in the glass) plus two optical properties (the orientation and size of the resulting nanostructures) southampton.ac.uk 5dmemorycrystal.com. In simpler terms, they are etching digital data in multiple layers through the thickness of the glass, with subtle differences in how each tiny spot is formed carrying additional bits.

Each piece of silica glass is like a small transparent disc or card (current prototypes are roughly 7.5 cm wide and 2 mm thick, about coaster-sized). Data is written layer by layer: the laser starts near the bottom of the glass and creates a pattern of voxels, then moves up a micron and writes the next layer, and so on – “like pouring layers of cement” until the top datacenterdynamics.com datacenterdynamics.com. A single piece can contain hundreds of layers of voxels. Early versions stored data in two layers, but newer versions greatly increased this count (Microsoft hasn’t disclosed the exact number in current prototypes, as it’s part of their recent breakthroughs datacenterdynamics.com).

Reading the data is done by shining light (e.g. a polarized light microscope) through the glass and observing the patterns. The nanostructures created by the laser change how light passes through (via birefringence – they alter polarization of light) 5dmemorycrystal.com 5dmemorycrystal.com. A camera sensor captures the diffraction or polarization image, and machine learning algorithms decode the patterns back into bits geekwire.com. Essentially, the glass, though transparent, has millions of tiny “scarred” spots that a clever optical system can interpret as digital ones and zeros (and more, since orientation can allow multi-bit symbols). Microsoft has developed software that uses AI to decode the images of the voxels quickly and accurately geekwire.com.

One important aspect: these laser inscriptions are permanent and immutable. The glass remains transparent (you can’t see the data with the naked eye), but the physical changes inside are locked in place unless the glass is melted or shattered. Richard Black, the research lead on Project Silica, explains that the ultrashort laser pulses make a “permanent, detectable, and yet transparent modification to the glass,” so the data is “as durable as the piece of glass itself.” geekwire.com In other words, once written, the data will live and die with the glass – it won’t fade or leak away.

Storage Density and Capacity

Project Silica’s storage density is high, though not as astronomically high as DNA’s. Microsoft’s early public demo in 2019 stored 75.6 GB (gigabytes) of data on a single glass platter – famously, they archived the 1978 Superman movie onto a glass about the size of a drink coaster. That was a proof of concept. By 2023–2024, their prototypes vastly improved: Microsoft has confirmed that newer Silica glass platters can hold multiple terabytes each geekwire.com. Black noted they provided one test unit that stored 100 GB, and later had versions storing 4 TB and 7 TB on a single piece of quartz datacenterdynamics.com. This suggests they significantly increased layer count and bit-per-voxel efficiency. In terms of density, Silica is now “competitive with Linear Tape-Open (LTO) on density,” according to Microsoft datacenterdynamics.com. LTO-9 tape holds about 18 TB per cartridge, which is roughly comparable to a few Silica glass pieces.

Crucially, Microsoft realized past a certain point they didn’t need to push density much further. “We pushed the density a lot. Eventually, the Azure [cloud] business asked us to stop… ‘stop pushing density, push other aspects’,” Black recounts datacenterdynamics.com datacenterdynamics.com. In a data center, ultra-high density is less important than cost and reliability – you can always add more glass platters if needed, since warehouses are big. So, while Silica’s density could theoretically go higher (research from University of Southampton suggests up to 360 TB on a DVD-sized glass disc might be possible with very fine voxels datacenterdynamics.com southampton.ac.uk), Microsoft is optimizing for a balance of density and write speed. In practice they found encoding 2–3 bits per voxel was ideal, rather than chasing 7+ bits per voxel which slowed writing exponentially datacenterdynamics.com datacenterdynamics.com. “Around two to three bits is where you want to be… it’s only a small amount of energy to make each one, and you can use all that other energy for doing hundreds simultaneously,” Black explains datacenterdynamics.com. By writing many voxels at once (splitting the laser into parallel beams) and keeping the per-voxel data modest, they achieved better throughput.

For real-world context: a Silica platter of a few terabytes is equivalent to a hard drive or LTO tape in capacity, but in a form factor the size of a coaster that can be easily handled by robots. For archival use, this is plenty – especially since you can have an optical jukebox of thousands of such glass pieces. Microsoft envisions racks of Silica platters stored in library systems with robotic arms (“crab-like robots on rails,” as one description put it) fetching the right piece and presenting it to optical readers datacenterdynamics.com datacenterdynamics.com. The concept is similar to tape libraries, but with much longer-lived media.

To sum up, Project Silica’s capacity per unit will likely hit multi-terabyte scale in its first iterations, with potential to grow if needed. It may not rival DNA in bits per gram, but it doesn’t need to – it’s already dense enough to replace tapes and optical discs for archival stores. And because the pieces are durable, one can also physically ship them or store them off-site easily for disaster recovery (like one might do with tapes).

Longevity and Environmental Resilience

One of Silica’s biggest selling points: it can last basically forever by human standards. Glass is an extremely stable material – it doesn’t corrode or degrade in normal environments. Microsoft has tested Silica glass samples under torturous conditions to ensure data survivability. They report that Silica can “happily survive being baked in an oven, microwaved, flooded, scoured, demagnetized, or exposed to moisture”, all without losing data datacenterdynamics.com. It can also withstand being dropped or struck to a significant extent – while it is glass and can crack under extreme force, the data voxels are inside the material, so minor surface scratches won’t affect it. In the University of Southampton’s tests on similar 5D nanoglass, the discs endured temperatures of 1,000 °C, pressures up to 10 tons per square centimeter, and even cosmic radiation exposure with no data loss southampton.ac.uk. In short, no practical environmental factor (heat, cold, water, magnetic fields, radiation) will erase or corrupt the data. The only real threats are shattering the glass or melting it down.

The longevity is measured in millennia. Microsoft often says Silica’s storage would be good for 10,000+ years datacenterdynamics.com datacenterdynamics.com. Southampton’s researchers go further: in theory the nanostructured glass could last billions of years without material degradation at room temperature southampton.ac.uk 5dmemorycrystal.com. In practice, no one is waiting around to test that, but the confidence comes from glass’s inertness – unlike plastic or magnetic materials, glass doesn’t chemically break down easily. It’s why museums have glass artifacts from ancient Egypt still intact after 3,000+ years. So if you store family photos or, say, a copy of Wikipedia on a Silica platter and keep it somewhere safe, it could still be readable in the year 12025 and far beyond datacenterdynamics.com.

This extreme durability is why Microsoft is comparing Silica to a “golden record” for the cloud. Just as the Voyager spacecraft’s Golden Record was meant to preserve sounds of Earth for eons in space, Project Silica aims to preserve digital data for future generations. They even collaborated in 2024 with a student-led project to create a “Golden Record 2.0” on glass – encoding images, music and literature about humanity onto Silica platters, complete with an engraved instruction key for aliens or far-future humans on how to read it geekwire.com geekwire.com. Each platter can hold several terabytes and remain readable for many millennia, the project noted geekwire.com. For the Golden Record platters, Microsoft also etched a visual user manual on the glass itself (in symbols) explaining DNA bases, the concept of a microscope, etc., to help unknown finders decode it southampton.ac.uk. This demonstrates the level of confidence in Silica’s longevity – planning for scenarios thousands or millions of years ahead.

Additionally, unlike most current storage, Silica requires no special climate control. Hard drives and tapes need controlled temperature/humidity to meet their lifespan specs (tape, for instance, is only rated if kept in cool, low-humidity archives). Silica, however, doesn’t need a constant cool environment – it doesn’t rust or mold. And because it’s unaffected by humidity swings, you don’t need power-hungry air conditioning or dehumidifiers to protect it datacenterdynamics.com. “Once the data has been inscribed, [its] cost is basically warehouse space,” says Black. “It doesn’t need any energy-intensive air conditioning… we’re talking true cold storage.” datacenterdynamics.com. This is a huge practical advantage: data centers could store Silica archives in normal warehouse conditions – even unpowered warehouses – with no risk of decay, dramatically cutting storage operating costs.

In terms of resilience to disasters: Silica archives could survive fires, floods, electromagnetic pulses, and more. For example, if a data center burns down, the Silica media inside might be scorched but the data would likely remain (quartz melts at ~1700°C, so typical fires won’t destroy it). Similarly, a Silica archive could be dunked under water or left in a salt mine for centuries and still be intact. This makes it attractive for heritage and backup archives (some have floated the idea of using Silica to store data in the Svalbard Global Seed Vault or other deep preservation sites).

Energy Efficiency and Cost

Like DNA, Silica is a true “write-and-shelve” medium – once written, it draws no power to maintain. No spinning, refreshing, or periodic rewriting is required. “Like tape, but unlike HDDs and SSDs, Silica is also true cold storage. It requires no power to maintain it in its rest state,” notes Black datacenterdynamics.com. This means massive energy savings for archival data. Currently, even “cold” data on hard drives consumes some power (drives need occasional spinning and climate control). Silica could eliminate that: a cloud provider could fill racks with glass archives and literally turn the power off for those racks until a read is needed. Over years, this is a huge reduction in energy usage and also reduces hardware maintenance.

From a cost perspective, the raw material – glass – is extremely cheap. A blank silica wafer costs virtually nothing compared to, say, an LTO tape cartridge or a hard disk. The major cost in Silica is the writing process: femtosecond lasers are currently expensive and slow. However, Microsoft expects these costs to drop as the tech matures (just as DVD burners were costly in their infancy). “The main cost remains the femtosecond lasers… [but] we’re hopeful the technology will follow nanosecond and picosecond lasers in dropping price as it scales,” Black says datacenterdynamics.com. Even if lasers stay pricey, Microsoft has a plan: since they offer storage as a cloud service, they can amortize the laser cost by running it at full capacity. Users of Azure won’t pay per laser use; they’ll just pay for archival storage as a tier, and Microsoft will manage how data is migrated to glass behind the scenes datacenterdynamics.com. Essentially, Microsoft can write data to Silica during off-peak times, keeping the laser busy, thus maximizing efficiency and lowering per-byte cost.

Black even suggests that in the cloud context, customers wouldn’t even know or care that their data moved to glass – Azure would simply have an “archival tier” and might silently shift rarely-used data from disk to Silica over time datacenterdynamics.com. Because Silica lasts forever, once written, there’s no ongoing replacement cost. Over decades, that could make it far cheaper than rotating out tapes or HDDs every few years. “Cost savings over the long term will drive societal change,” Black argues, noting that when incremental storage cost goes near-zero, organizations will keep data longer and regulations might mandate longer retention since the tech no longer limits it datacenterdynamics.com datacenterdynamics.com.

In summary, Silica promises very low operating cost (power, cooling, maintenance) and longevity that defers replacement costs indefinitely. The initial capital expense – the writing hardware – is the main economic hurdle, but one that can be managed at scale. If Microsoft’s plan works out, glass storage could actually undercut tapes in TCO (total cost of ownership) for archives. The team already claims Silica is on par with LTO tape in cost per TB and density datacenterdynamics.com, and because you never have to replace a Silica platter, the lifetime cost should be lower. As evidence of their confidence, Microsoft has stored all of Datacenter Dynamics magazine’s archives (along with the article describing Silica itself) on a Silica platter as a permanent record datacenterdynamics.com – a fitting demonstration that the medium is ready to carry historical data forward.

Performance (Throughput and Speed)

The read/write speeds of Project Silica are better than DNA, but still not as fast as conventional drives. Writing data in Silica involves scanning a laser through the glass – it’s a sequential optical write process. Early iterations were slow (writing tens of GB could take many hours). However, Microsoft achieved several breakthroughs to speed this up:

They developed a method to split the laser beam into multiple beams (e.g., 8 beams) so that it can write multiple voxels in parallel datacenterdynamics.com. Think of it like writing 8 bits at once instead of one at a time. This parallelization directly multiplies throughput.
They use a rotating mirror polygon (like those in laser printers or supermarket barcode scanners) to sweep the laser rapidly across the glass without physically moving the glass datacenterdynamics.com. This allows writing at high speed line by line.
By optimizing the bits-per-voxel and spacing (as discussed, ~2 bits per voxel to avoid too many pulses per voxel), they keep the laser firing efficiently rather than overworking on one spot datacenterdynamics.com.

While exact figures are confidential, Black hinted that the team made “a number of breakthroughs in the last few years” on density and speed, which they are keeping under wraps for now datacenterdynamics.com. It’s telling that Azure asked them to stop increasing density, implying focus shifted to improving throughput. We do know writing still isn’t instant: the Silica lab setup occupies a large table of lenses and sensors, and writing a multi-terabyte glass still likely takes on the order of several hours at least. But Microsoft plans to refine this into a smaller, production-ready machine (perhaps leveraging expertise from their HoloLens optical team) datacenterdynamics.com.

Reading speed is quite favorable: since you can read an entire layer of the glass at once via imaging, retrieving data could be very fast – potentially on the order of Mbps to Gbps once the system is optimized, because it’s just taking pictures of each layer. The bottleneck is moving the glass in position and focusing, which a robot can do relatively quickly. Microsoft hasn’t published numbers, but one can imagine retrieving a few terabytes could be done in minutes to hours, which competes well with tape (which typically streams data at 100+ MB/s but has slow loading times).

One constraint: random access vs sequential. Silica, like tape, is more sequential – you’d likely read large chunks or entire platters at once, rather than randomly seeking a 1 MB file. But with clever indexing (and because the cost of reading is just time, not wear and tear), it might be feasible to still retrieve individual files by reading only the relevant layers or sections.

Given Microsoft’s use-case (cloud archival), they are content if Silica reads/writes in bulk with latency of maybe a few seconds to minutes. It’s for backups, not live databases. So the performance, while lower than spinning disks, is tuned for high throughput archival workloads.

To put it simply: Silica is faster than writing DNA or burning Blu-rays, but slower than writing to SSDs or even HDDs. It sits in a new tier of the storage hierarchy – slower to write and read than “hot” storage, but not unbearably slow. And since no one will use Silica for active frequently-accessed data, that trade-off is acceptable. The big win is that once written, you don’t have to touch it again for centuries, so speed is mostly a one-time concern.

Development Status and Recent News (2024–2025)

Project Silica has moved from a lab experiment to a near-product within Microsoft’s cloud division:

Azure Integration Plans: By late 2024, Microsoft indicated it is in the “early stages of thinking about rolling out” Silica in its Azure cloud datacenters datacenterdynamics.com. They envision offering it as an archival storage tier. This suggests the project is far enough along that deployment strategies are being planned (how robots will handle platters, how data will move between Azure’s standard storage and Silica, etc.). No exact date is public, but this could mean pilot use in data centers as soon as 2025–2026.
Prototype Achievements: Microsoft has demonstrated Silica’s capabilities with several notable prototypes:
- In 2021, they showed off a glass storing the entire Microsoft Flight Simulator 2020 game (~2 petabytes of world data in the cloud; obviously not all on one glass, likely a representative chunk) datacenterdynamics.com datacenterdynamics.com. Black was literally holding a quartz piece with Flight Simulator on it as a testament to progress.
- In 2024, they partnered with the UK’s National Archives to test Silica for preserving historical records (Microsoft stored some archived documents on Silica as a demo for the Archives). Although not widely publicized, tags in the DCD article suggest National Archives interest datacenterdynamics.com. This aligns with the idea that archives are eager for ultra-long media.
- The Golden Record 2.0 project in 2024 (mentioned earlier) – Microsoft encoded student-curated content about humanity onto Silica platters to simulate a far-future time capsule geekwire.com geekwire.com. They even worked with original Golden Record creators like Jon Lomberg, proving Silica’s appeal for cultural preservation.
Partnerships: Microsoft’s first big partner was Warner Bros., which collaborated on the Superman movie-in-glass proof of concept in 2019. Warner was interested because studios have vast film archives to preserve. That partnership likely continues, aiming to eventually archive Warner’s content vault on Silica. Additionally, Microsoft has been working with researchers at University of Southampton, who pioneered 5D glass storage. The DCD piece notes Microsoft initially partnered with Southampton for early iterations of Silica datacenterdynamics.com. Since then, Microsoft brought the work fully in-house, but that academic collaboration jump-started the effort (Southampton’s 2013–2016 research essentially proved the concept that enabled Silica datacenterdynamics.com datacenterdynamics.com).
Competitive Landscape: While Microsoft appears ahead in glass storage, others are looking at similar concepts. In 2022, a startup Cerabyte announced work on femtosecond-laser storage but using ceramic materials instead of glass (claiming multi-thousand-year life as well). A company called Piql in Norway uses a different approach: optical film designed to last 500–1000 years (somewhat analogous concept, though lower density) datacenterdynamics.com. Microsoft is aware of these but seems confident Silica is the frontrunner. They’ve stated Microsoft will likely “solely offer Silica as a cloud service”, owning the tech and not licensing it out widely datacenterdynamics.com datacenterdynamics.com. This proprietary approach indicates they want a unique Azure advantage.

Recent news did throw a bit of shade on DNA solutions, as noted earlier, with Microsoft’s Richard Black openly saying they stepped back from DNA to focus on Silica because it was yielding more practical results datacenterdynamics.com. That was a significant moment in 2024: a major tech company basically choosing glass over DNA for its future archival strategy. It underscores Project Silica’s momentum and Microsoft’s confidence in it.

As of 2025, it’s expected that Microsoft will soon announce a more formal timeline or productization of Silica. Perhaps a preview in Azure for certain customers (like those with big compliance archives) could come by 2026. The hardware still needs to shrink and become robust – presently the writing system is table-sized, but Microsoft envisions a production version “much smaller” without all the diagnostic sensors, akin to how early HoloLens prototypes shrank to a headset datacenterdynamics.com. The reading system (microscope+camera) might also be integrated into a drive-like unit.

If things go well, within a few years we might see Azure advertise an “Immutable Storage” tier where data is stored on quartz glass for ultra-safe keeping. It’s quite possible that the first commercial glass data storage service in the world will come from Microsoft Azure, thanks to Project Silica’s head start and Microsoft’s cloud scale to implement it.

Real-World Use Cases and Future Potential

Project Silica’s target use cases overlap with DNA’s in the archival domain, but with some differences given Silica’s faster access and slightly lower density:

Cloud & Enterprise Archiving: The immediate use case is replacing tape in cloud data centers. Cold storage services (like AWS Glacier, Azure Archive) could use Silica for storing backups, compliance data (e.g., financial records that must be kept 7+ years), medical records, etc. Customers would benefit from higher durability (no bit rot) and providers benefit from lower maintenance (no tape robots cleaning and cartridges wearing out). Essentially, any scenario where LTO tape libraries are used today is a candidate for Silica libraries.
Media Archives: Film studios, broadcasters, and content creators have petabytes of video sitting in archives. Silica could be used to archive master copies of movies, TV shows, audio recordings for preservation. The Warner Bros. test storing Superman on glass demonstrates this use. Unlike film reels or magnetic tape which decay, glass could keep original footage intact for future remastering or cultural heritage. Similarly, national broadcasters could preserve historical broadcasts on Silica (BBC archives, etc.).
Scientific Archives: Large scientific experiments (like particle colliders, telescopes, climate data logs) produce data that may be needed many decades later for analysis or verification. Silica could store raw datasets or processed results that are rarely accessed but must not be lost. Its radiation hardness even makes it feasible for storing data in space or on other planets (e.g., deploying a Silica data archive on the Moon or Mars that can survive high radiation and temperature swings).
Legal and Institutional Records: Governments and institutions can use Silica to satisfy long-term record retention requirements (centuries-long). For example, a country’s birth/death records, land registries, or parliamentary archives could be sealed in Silica so that even in 500 years the data is readable. The UK National Archives’ interest hints at this. Also, critical documents like constitutional texts, treaties, or religious scriptures in digital form could be preserved on Silica to guard against loss or alteration over ages.
Heritage and Time Capsules: Beyond functional data, Silica opens possibilities for digital time capsules. Museums might create “digital Rosetta Stones” on glass, storing information about our civilization for the far future. The Golden Record 2.0 is a perfect example: a cultural snapshot stored on Silica and perhaps kept in a deep cave or launched into orbit, with confidence it will persist when current digital formats are long gone.

Looking ahead, Project Silica could evolve to higher capacities (through even more layers or larger glass pieces) and possibly wider adoption outside Microsoft. If Microsoft proves it out, others might develop similar glass storage systems. It’s not hard to imagine future archival libraries, whether in government or enterprise, adopting glass storage appliances (even if via Azure’s model or eventually on-premise units sold by Microsoft). Over decades, Silica might complement DNA storage too – DNA might win when physical size is at an absolute premium, whereas Silica wins when faster access is needed and a bit more volume is acceptable.

Interestingly, Microsoft’s work also suggests a shift in how we think about data permanence. If Silica truly makes data effectively immortal, we may need new policies. Black mentioned that once cheap ultra-long storage exists, people may be required to keep data longer datacenterdynamics.com – for instance, companies might have to retain environmental or medical data for the public good for centuries, since the old excuse of media decay no longer holds. “If people internalize what it means, it’s going to be a step change in how we think about data preservation… We don’t need to have so much loss,” says Dr. Ioan Ștefănovici, one of Black’s colleagues datacenterdynamics.com datacenterdynamics.com. The existence of Silica and DNA storage could usher in an era where digital forgetting becomes a choice, not an inevitability.

In summary, Project Silica represents a compelling marriage of ancient material (quartz glass) with cutting-edge tech (ultrafast lasers and AI) to tackle the modern data crisis. Its current trajectory shows it’s very close to practical use, and within a few years our most precious data might be entrusted to tiny glass slabs, safe for ages.

5D Nanostructured Glass Memory Crystals

Often nicknamed the “Superman Memory Crystal” (after the fictional crystal data stores in Superman’s Fortress of Solitude), 5D optical data storage is another glass-based technology that closely relates to Project Silica but emerged from academic research. The concept was pioneered by the University of Southampton’s Optoelectronics Research Centre (ORC), which in 2013 first demonstrated using nanostructured glass to record data in multiple dimensions datacenterdynamics.com. This technology also uses ultrafast lasers to encode data in quartz glass, but with some differences in technique and a focus on maximum density and durability. Think of it as the academic cousin to Project Silica, exploring the outer limits of capacity and longevity in glass.

How 5D Optical Storage Works

5D storage uses a very similar idea of creating tiny laser-induced changes in glass. The “5D” refers to the same five degrees of freedom (3 spatial + 2 optical) described earlier southampton.ac.uk. However, the Southampton team’s method emphasizes creating nanogratings – extremely small periodic structures – whose orientation and strength encode bits via birefringence (splitting light into different polarizations) 5dmemorycrystal.com 5dmemorycrystal.com. In simpler terms, instead of just “bubbles” or voids, they create little patterns in the glass at the nanoscale that alter how light passes through. By analyzing the light with a polarizer (like polarized sunglasses) and microscope, they can read multiple bits from a single spot depending on the angle of those nanogratings 5dmemorycrystal.com.

The writing process uses a femtosecond laser moving through a disc of fused quartz (pure silica glass), focusing pulses to write layer by layer, similar to Silica. Each spot (voxel) can store up to 8 bits (one byte) of data by using different orientations and retardance strengths of the nanograting 5dmemorycrystal.com. In lab experiments, they achieved up to 7–8 bits per voxel (which is where the term “5D” often is illustrated – e.g., 2^8 possibilities per voxel). However, as Microsoft found, pushing to 8 bits per spot slows writing. The Southampton team, unconstrained by commercial considerations, has pushed towards the theoretical max density even if it’s slow.

In terms of capacity, they published that a standard DVD-sized 12 cm glass disc using 5D nanostructures could in theory hold 360 TB of data datacenterdynamics.com southampton.ac.uk. This assumes very tight voxel packing and high multi-bit encoding per voxel. It’s essentially the upper bound if you filled a disc with hundreds of layers of densely packed nanograting data. In practice, early prototypes were much smaller: the first demonstration in 2013 stored 300 kB (kilobytes) on a tiny piece of glass datacenterdynamics.com. By 2015, they stored a few MB. Progress accelerated, and by 2016 they had stored text files like the King James Bible (over 1 MB) to showcase capacity and retrieval.

Most impressively, in September 2024, the Southampton ORC team announced they had stored the entire human genome (over 3 billion DNA base pairs, ~0.8 GB) on a 5D crystal southampton.ac.uk southampton.ac.uk. This was done on a small glass disc about 8 cm in diameter, and it utilized redundancy (each genome letter stored 150 times to ensure accuracy) southampton.ac.uk. The achievement was symbolic: preserving human genetic code on an “eternal” storage medium. They then placed that glass disc into the Memory of Mankind archive – an ultra-long-term time capsule in a salt mine in Hallstatt, Austria southampton.ac.uk southampton.ac.uk. This effort shows the technology working for a reasonably large dataset and being aimed at ultimate longevity use cases.

Capacity and Density Potential

As noted, 5D optical storage’s potential capacity is enormous – up to hundreds of terabytes per disc in theory southampton.ac.uk. In practice, the current prototypes have used smaller data volumes (gigabytes) as proof of concept. The focus has been on demonstrating extreme durability rather than maxing out capacity in one go.

However, in a technology perspective:

A single 5D disc can have thousands of layers of data, since the laser can refocus deeper into the glass. There isn’t a hard limit like with Blu-ray (which had 3–4 layers) – it’s more about how fine you can position the laser in depth. The ORC team has mentioned the possibility of “stacking layers with no limit” except physics of diffusion, etc. Practically, a few hundred layers have been shown, and more is feasible with high NA (numerical aperture) lenses.
The voxel density within each layer can approach that of Blu-ray or better (voxels ~microns apart). Using shorter wavelength lasers or tighter focusing, they can shrink voxel size to e.g. <200 nm, similar to Blu-ray pit sizes 5dmemorycrystal.com. This yields very high bits per layer.
Multi-bit encoding per voxel adds another multiplier (if you reliably store, say, 3 bits/voxel, that’s 3x the capacity of simple binary per voxel).

Combining these, the oft-quoted “360 TB on a disc” is if you exploit all dimensions fully. That would be like compressing 1000 standard archival Blu-rays into one glass disc. It hasn’t been reached yet, but even a fraction of that is huge.

One advantage of 5D approach is rewriteability. Surprisingly, researchers have found that by using certain laser conditions, the nanogratings can be erased or re-written 5dmemorycrystal.com. This is still experimental, but if achieved, it means these glass crystals wouldn’t have to be write-once; you could update data on them, which is normally not possible on immutable media like Silica or WORM discs. An ORC paper in 2016 noted the nanostructures could be selectively erased by heating or additional pulses, then rewritten, albeit with some limitations. If perfected, this could be revolutionary: a nearly eternal, rewritable storage medium.

Currently, the project remains in the research phase and is not productized. However, a company called Sphaera Photonics (SPhotonix) appears to be a spin-off or collaborator aiming to commercialize 5D memory crystals. The website 5Dmemorycrystal.com (branded with SPhotonix) talks about the technology and even has an option to “order memory crystal” 5dmemorycrystal.com 5dmemorycrystal.com. This suggests small-scale offerings or custom inscriptions might be available (perhaps for novelty or special preservation projects). Indeed, some enthusiasts have commissioned, for example, a 5D crystal with certain famous documents stored, as a way to own an object that literally holds the text internally.

Longevity and Stability

The 5D glass crystal holds the current Guinness World Record for the most durable data storage material southampton.ac.uk. The researchers boast an estimated thermal stability of 13.8 billion years at room temperature (effectively, data could survive until the end of the universe barring cataclysm) 5dmemorycrystal.com. This number comes from extrapolating the measured deterioration at high temperatures – basically, at normal conditions the decay is so slow as to be negligible. In a more digestible claim, they often say “virtually eternal” storage.

As described earlier with Silica, these quartz glass crystals are resistant to extreme conditions:

Heat: withstanding up to 1000°C without losing data southampton.ac.uk. (So you could drop it in boiling water or a fire and it’d be fine, though the disc itself might deform if it truly melts.)
Cold: no effect from freezing or extreme cold.
Chemical stability: Quartz is very inert. Short of dissolving it in hydrofluoric acid, most chemicals or solvents won’t harm it.
Pressure: Tested against 10 tons per cm² – which is an enormous pressure – and it remained unchanged southampton.ac.uk.
Radiation: The disc in Hallstatt is exposed to natural radiation and possibly cosmic muons; it’s expected to handle that fine. The press release explicitly states it is “unchanged by long exposure to cosmic radiation.” southampton.ac.uk.

Prof. Peter Kazansky, who leads the ORC team, highlighted that unlike other formats, these crystals “do not degrade over time” – they can keep data “for billions of years, even at high temperatures.” southampton.ac.uk This was proven enough to get the Guinness record in 2014. It’s fair to say 5D crystals and Project Silica share the same fundamental longevity, since they’re both quartz-based; the difference is ORC is pushing it as far as possible.

To truly future-proof, the Southampton team also considered the information accessibility over time. Knowing that after millennia, someone might find the crystal with no knowledge of our technology, they engraved visual “clues” or keys on the disc itself explaining how to decode the data southampton.ac.uk. For the human genome disc, they etched symbols of atoms, DNA structure, etc., to guide a future intelligence in interpreting the binary code of the genome southampton.ac.uk. This mirrors what was done with the Voyager Golden Record (which had diagrams explaining how to play it). They even included an image of a human figure like the Pioneer plaque, as a nod to space-bound records southampton.ac.uk. Such measures show the creators anticipate these crystals could outlast our civilization, so they come with instructions for whoever (or whatever) discovers them in a distant future.

In terms of maintenance, again, no environmental controls needed. A 5D crystal can be left on a shelf in a cave for eons (as they did in Hallstatt). There’s no requirement of periodic re-copying. It’s truly a “write and forget” medium – hopefully not forget to retrieve, but forget in terms of upkeep.

Current Stage and 2024–2025 Updates

As of mid-2025, 5D memory crystal storage is still primarily a research project, but with growing real-world demonstrations:

The September 2024 announcement of storing the human genome was a major news event for the field southampton.ac.uk. It garnered media attention (being covered by Phys.org, Smithsonian Magazine, etc.), not just for the technical feat but the philosophical one of preserving humanity’s blueprint in such a durable way technologynetworks.com reddit.com. It proved that multi-gigabyte data could be handled (the genome with redundancy was several gigabytes of encoded data).
The Memory of Mankind (MoM) partnership: By placing the genome crystal in the MoM archive in Austria, the team signaled a use-case for their tech – joining a repository intended to preserve human knowledge for the far future. The MoM archive primarily consists of ceramic tablets with micro-printed text, but now also contains this high-tech glass disc. This may open the door for more contributions: perhaps future MoM deposits will include 5D crystals containing libraries or artworks.
SPhotonix / Commercial Angle: There is an indication of trying to commercialize on a small scale. SPhotonix (possibly run by ORC alumni) offers “Memory Crystal” products. It’s not clear if they are selling large data storage services yet or if it’s more experimental. One could imagine early clients might be museums or wealthy individuals who want to preserve certain data in a novelty eternal format (for instance, offering a service to store one’s family photos or personal data on a crystal for the grandchildren of your grandchildren). We don’t have evidence of large orders, but the website being up and referencing the tech suggests at least outreach.
Collaboration with Microsoft? In the earlier years, Microsoft worked with Southampton ORC as mentioned in the DCD article datacenterdynamics.com. That collaboration ended as Microsoft pursued its own design. But it means the 5D tech and Silica have cross-pollinated ideas. The academic team continues to push limits (like reaching 7 bits/voxel, or showing rewrite, or doing the genome archive), while Microsoft focuses on engineering a deployable system (with 2-3 bits/voxel and robotics). In 2022, a DCD profile of various long-term storage noted that “others are working on alternative approaches (see box)” and cited Southampton’s work datacenterdynamics.com. So, Microsoft acknowledges them as a parallel path. We may see a divergence where Microsoft’s Silica becomes the enterprise standard, while the 5D crystals become the ultra-endurance option for special archives.

In 2025, we haven’t seen a 5D storage sold as a product like “store your data forever for $X”, but given interest, it might not be far off. Perhaps a startup will offer it as an archival service similar to DNA pilot services. It’s worth noting that reading the data from these crystals also requires a specialized optical setup (a polarizing microscope with camera and decoding software). So any commercial offering would include providing the read solution or guaranteeing future readability.

Potential Applications

Many of the use cases overlap with Project Silica’s, since both serve ultra-long storage needs. However, 5D crystals might carve out niche uses where maximum longevity or data density is paramount, possibly at the expense of speed or cost. Some envisaged scenarios:

Time Capsules & Civilizational Archives: This tech is almost tailor-made for things like the Memory of Mankind project, or a future “Library of Earth” that we send to the Moon or bury underground. The combination of density and durability means you could archive entire libraries, museums, scientific archives, etc., in a few shoeboxes of crystals and be fairly certain they’ll outlast humanity. For example, storing all of Wikipedia, Project Gutenberg, and a snapshot of the internet’s knowledge on crystals and sealing them in a vault for a future civilization. The human genome crystal is a prime example (they even suggested doing similar for endangered species’ genomes as a backup of biodiversity) southampton.ac.uk southampton.ac.uk.
Space Missions: If we ever create an interstellar probe akin to Voyager again, 5D crystals would be a great medium to include. They can endure space radiation and time, and hold vastly more than phonograph records or golden plaques. A “Golden Record 3.0” might be a silica crystal with a compendium of Earth’s culture. Even for more near-term missions, say a lunar base storing data in situ – crystals would survive the harsh lunar environment (temperature swings, radiation) far better than electronics. In fact, some have whimsically suggested sending a time capsule to the Moon with humanity’s knowledge on 5D glass, just in case.
Ultimate Backup of Backups: Organizations might use 5D storage as the format for disaster recovery archives. For instance, a nation’s critical data could be written to crystals and stored in a secure bunker (or multiple bunkers). This would be the “in case of apocalypse, break glass” data vault – pun intended. Tech companies might use it to store source code or encryption keys in an immutable form offline.
Museum and Cultural Heritage: National archives or museums could offer exhibits where, say, a 5D crystal holds all the museum’s digital collection – art scans, documents, etc. – as a symbolic cornerstone for preservation. It’s both functional and educational: visitors can see the “eternal archive” and know that those works are safe for ages. Already, we see interest: during the genome crystal event, Prof. Kazansky noted it “opens up possibilities to build an everlasting repository of genomic information… from which complex organisms might be restored should science allow” southampton.ac.uk, hinting at storing other complex data like possibly digital blueprints of organisms or critical knowledge.
Military or Secure Archives: Militaries and intelligence might store sensitive data on crystals because they’re tamper-proof (you can’t alter data without leaving evidence) and can survive EMP or other attacks. For example, nuclear launch code archives or treaty documents could be kept on a medium that even a nuclear blast wouldn’t destroy, ensuring some continuity of info.
Personal Legacy Data: This is more speculative, but perhaps in the future individuals might be able to store their personal data – family photos, journals, family tree – on a small crystal to hand down as a heirloom. It’s like a digital locket that great-great-grandchildren could still read. While not mainstream yet, the concept of a “digital legacy” stored for centuries might appeal to some, especially as costs come down.

One thing to note: reading a 5D crystal in the far future assumes the knowledge of how to build a polarizing microscope and understand binary. But the researchers have tried to mitigate that with the engraved guides, as mentioned southampton.ac.uk. They even express that “if we can explain to extraterrestrials how to read it, it should be obvious we can explain to our human descendants” geekwire.com. So they are conscious of the need for accessibility.

In terms of next-gen versions, the 5D concept itself is arguably the next-gen of optical storage. It leapfrogs Blu-ray and holographic storage attempts by adding more dimensions of encoding. If any “next-next-gen” improvements come, it might be:

Using different materials: maybe doped glass or other crystalline materials to improve contrast or allow even finer voxels.
Multicolor lasers: possibly encoding with different wavelengths of light to add even more data dimensions (this is theoretical, not currently done).
As mentioned, making it rewritable reliably, which would be huge.
Automation: building robotic systems to handle these crystals in libraries (similar to Silica’s plan).

For now, 5D memory crystals remain a fascinating blend of science experiment and forward-looking archival solution. They capture the imagination because they tick all the boxes of the dream storage: extreme density, eternal longevity, no maintenance. As Prof. Kazansky mused, “We don’t know if memory crystal technology will ever follow [the Voyager] plaques in distance traveled, but each disc can be expected with a high degree of confidence to exceed their survival time.” southampton.ac.uk In other words, these might be the closest things to time-proof digital libraries that mankind has created.

Conclusion: A Future Where Data Truly Lives Forever

The quest for next-generation data storage has led down some radical paths – from synthetic DNA in test tubes to sapphire-hard glass platters etched with lasers. Synthetic DNA storage, Microsoft’s Project Silica, and 5D nanostructured glass each offer a solution to the impending “data storage apocalypse” where our needs far outstrip the capabilities of conventional media. They share common goals of sky-high density, longevity measured in millennia, and minimal energy use, yet they differ in maturity and ideal use cases:

Synthetic DNA leverages nature’s information molecule to pack unbelievable amounts of data in tiny spaces. Its appeal is the ultimate density (potentially millions of GB per gram) and proven durability over geological timescales microsoft.com techxplore.com. DNA storage could theoretically shrink the world’s archives into a few refrigerator-sized biovaults. However, the technology is still gestating – writing and reading are slow and expensive today, meaning DNA is likely a decade away from routine use. Early deployments will target deep archives where access time of hours or days is acceptable. Thanks to heavy R&D investment, we are seeing rapid progress (e.g. automated DNA storage machines and error-correcting algorithms), and startups like Atlas aim for commercialization by the late 2020s prnewswire.com blocksandfiles.com. As costs drop, DNA could become the go-to for cold storage of massive, rarely-accessed datasets – imagine cloud providers offering DNA storage for backup and compliance data that you only retrieve in emergencies.
Project Silica has made the concept of “glass data archives” tangible. Microsoft’s focus on a real-world product means Silica might beat DNA to market. By compromising a bit on density but simplifying read/write, they have a system that’s nearly ready for cloud-scale introduction datacenterdynamics.com. Silica’s strength is being practically implementable: it works with existing data center models (robotics, optical drives) and promises immediate savings by never needing replacement or power datacenterdynamics.com datacenterdynamics.com. With a few terabytes per platter and extreme robustness datacenterdynamics.com, it’s poised to disrupt the archival storage tier soon. Its likely role is in enterprise and cloud archival, replacing tape libraries with racks of glass that can sit for centuries. Microsoft’s own experts are now evangelizing Silica as the solution that could allow us to “store the world’s data” indefinitely and cheaply datacenterdynamics.com datacenterdynamics.com. We may soon see a paradigm shift where data that used to be deleted for cost or obsolescence reasons can be kept forever on glass – changing attitudes toward data preservation in science, history, and business.
5D Nanostructured Glass (Memory Crystals) represents the bleeding edge of what’s physically possible with inert storage. It pushes the boundaries – aiming for hundreds of terabytes on something like a DVD, and survival until the end of time southampton.ac.uk 5dmemorycrystal.com. While not yet a commercial solution, it serves as an existence proof that we can make storage devices that laugh at the limits of decay and capacity. The 5D approach will likely find its way into specialized uses: think archival vaults of last resort, time capsules, and scientific or interstellar archives. It may complement more practical solutions (like Silica) by offering even greater density for those willing to trade off speed and production cost. If future research achieves things like rewriteability or easier mass production, these memory crystals could step out of the lab as ultra-capacity archival media for cloud providers too. Even now, the fact that a small team encoded an entire human genome onto a tiny glass disc and asserted it could survive billions of years southampton.ac.uk southampton.ac.uk is a sign that “eternal” storage isn’t fantasy – it’s here, at least in prototype form.

Finally, it’s enlightening to hear what experts say when comparing these technologies. Dr. Richard Black of Microsoft, having worked on both DNA and glass projects, opined in 2024 that in the near term “there’s not a clear use case for DNA storage”, whereas glass was meeting the needs datacenterdynamics.com. He noted proponents of DNA emphasized density, but for now data centers care more about cost and integration datacenterdynamics.com. On the flip side, DNA storage champions like Twist’s CEO Emily Leproust remain confident that DNA will ultimately win for future-proofing huge volumes – she projected offering DNA storage services by 2025 and beyond, highlighting that biology’s advancements (cheaper synthesis, enzymatic writing) will close the gap. And as a venture capitalist might point out, both approaches can thrive: the market for long-term storage is so vast and varied that DNA and glass may each fill a niche – one for absolute density in smaller vaults, the other for faster access in large active archives.

As for 5D crystals, we have the enthusiastic words of Prof. Peter Kazansky, who called them “virtually everlasting, indestructible storage allowing hundreds of terabytes per crystal” phys.soton.ac.uk. After preserving the human genome on one, he said it “opens up possibilities to build an everlasting repository of genomic information” for the future southampton.ac.uk – suggesting these crystals could safeguard the essence of life and knowledge itself.

In conclusion, the emergence of these technologies signals a coming era where data can truly be kept for posterity. We are nearing solutions to the age-old problem of information loss. Imagine a future Library of Alexandria that cannot burn down, because its books are encoded in DNA or glass, replicated around the world. Or a cloud storage account where your photos and documents, if you choose, can be handed down across generations without degradation. With synthetic DNA, glass platters, and 5D crystals, our digital civilization may finally have archiving tools as resilient as the information is valuable. The next few years will be critical to watch as labs turn these prototypes into real systems. One thing is certain: in the race to store data forever, humanity now has some powerful tools on its side – at the crossroads of biology, optics, and nanotechnology – ensuring that our bytes and bits won’t just vanish with time.

Sources: Synthetic DNA storage press releases and interviews prnewswire.com microsoft.com; Microsoft Project Silica technical reports and media articles datacenterdynamics.com geekwire.com; University of Southampton 5D memory crystal announcements southampton.ac.uk southampton.ac.uk; Expert commentary from researchers and industry leaders datacenterdynamics.com southampton.ac.uk.