Why the World Is Running Out of Data Storage

Is your archive built to last 30 years?

In this episode, we sit down with Charlie Gale CEO & Founder of HoloMem, the startup building a commercial holographic data storage drive designed to complement LTO in enterprise backup and compliance use cases.

📩 Want to see HoloMem in your data center?
Get in touch https://www.fortunadata.com/contact-us/
🔔 Subscribe for more infrastructure innovation

We cover:

• Why holographic data storage finally makes sense
• How HoloMem’s drive and cartridges integrate into existing tape ecosystems
• Use cases for immutable media (WORM by default) in compliance, AI reference, and unalterable archives
• Their delivery roadmap through pilots and into production

If you're thinking about the future of long-term storage, this is a practical and forward-looking conversation with the team building it.

Follow the show for more on storage, infrastructure, and data at scale.

Transcript

SPEAKER_00 0:00

Long-term data storage is under pressure. As data volumes surge, the challenge is no longer just storing information. It is preserving it for decades, efficiently, securely, and at scale. Holomem is taking a radically different approach. Let's dive into the conversation.

SPEAKER_01 0:16

Hello, my name is Ray Quattromini from Fortuna Data. Welcome to Smarter Thinking. Today we have Charlie Gale from Holomem. I've met you before, I've been to your premise in London. Tell me a little bit about your background and story.

SPEAKER_02 0:29

I guess I'm a um self-described inventor and uh I've always had somewhat of a love affair with optical technology. So at university, um I developed and patented a holographic road sign, and of course, as a 21-year-old was convinced I was about to change the world. Uh went out uh trying to sell these novel road signage uh uh devices. Um there was no market, uh, but um it really got me interested in what can be achieved with multi-layer optics. It continued throughout my career. Um, I took a job at Dyson in Products Innovation, where I worked for just under 10 years, kind of in the think tank of what can you do, what prototypes can you make, how can we protect these, and how can we bring kind of disruptive technology to the market. But this interest in optical technology led to holographic labels. Um, so uh in the world of holography, all real kind of money has been made in these holographic stickers that you put on DVDs, banknotes, you name it, and these shiny stickers are meant to prove that something is real. But they're analog, you're just looking at a label. What we did was put machine readable data in a hologram, multi-layer machine readable data. So QR codes that are stacked on top of each other in a hologram, and they can be scanned and digitally verified, which is great. We have patented those, we we sell them, we put them on whiskey bottles, you name it. But in lockdown, uh the question was the the team and I that uh had built this how small can we make machine readable holograms and how many layers can we add to them? Because ultimately you're gonna end up with an incredibly high-fidelity optical data storage technique. This is by no means a new theory. Uh, the the kind of the thesis of holographic data storage has existed for a very long time. Uh, there's a Wikipedia page dedicated to it. Um but uh me and my team rather arrogantly thought that we were we've done this before, we can do it again. Um, how hard could it be to make a working, tangible, demonstrable holographic data storage drive? Holo-Mem was born.

SPEAKER_01 2:29

Do you want to tell our viewers a little bit about what Holomem is and how it works? Because let's see, here we have a what is an LTO tape but contains Holomem.

SPEAKER_02 2:39

So the the journey began for us really seeing how how we could record large amounts of data in incredibly small spaces. And I mean, holographic data story, the reasons that people kind of get attracted to this is it's long been kind of heralded as the holy grail of ultra-high density, long life, uh robust archival data storage. And uh the reason for being is that you use polymer-based materials to store very small images of data, so kind of binary dot matrix codes. The first thing Holomen did in our original lab, it was a garage in Acton, was built a large optical table where we had uh lasers, lenses, you name it. Um three meters long, this thing was. We actually had a huge lithography lens uh that had stamped on the cider, made in West Germany, which gives you an idea as to its age. Big rig, but what it made was a small sample of a polymer film that we had recorded the album Random Access Memories by Darf Punk onto. So small one inch square, full album. And um, I took that home, I put it in the oven with my dinner that night, cooked it, bought it in the next day, and we replayed the album uh from that small piece of film. So the whole kind of thesis of having a high-density, ultra robust data storage technique using a low-cost polymer film was was proven by this very big system. The journey from there, I mean, that was some three years ago now, um, has been pretty much operating in total stealth mode as a company. We're just a a team based in West London as a kind of a group of hardware junkies that have taken the kind of the holographic theories and innovations that were built on that table to now a consolidated automated data storage drive. So all of those lenses, lasers have been bought into a a unit. We call it the holo drive, and the film that it writes on that same polymer film, we've taken away from a one-inch square of it and we've put it into a cartridge, a cartridge that's compatible with LTO infrastructure. So within our cartridge, we have a light-sensitive polymer film in that same shape of cartridge where we can store more data, it's longer life. This is guaranteed 50 years of storage, and it's just incredibly robust. We talk about it as it's the cockroach of data storage, right? So when it comes to thermal fluctuation, humidity fluctuation, this material was actually developed for automotive applications. So it's incredibly resilient to thermal change and humidity change. So we're saying no active management of its storage environment is required.

SPEAKER_01 5:12

That's really interesting. What are the benefits of Holomem after traditional LTO take? Because as far as I can tell, you're going after the true archival market.

SPEAKER_02 5:23

I mean, so we're talking about um cold data storage and archival data storage, right? And as you're probably aware, I mean, the headwinds that are being faced by the industry at the moment are just ginormous. I mean, data storage, uh, original data creation is growing exponentially, all existing storage tiers are being pushed to their limits. So you see these huge price surges taking place across flash, hard disk drives, and magnetic tape because it's a struggle for capacity ultimately. Magnetic tape is seeing somewhat of a renaissance era where it's uh it's uh it was maybe previously considered somewhat a vintage technology, but tape is is incredibly beneficial. I mean, it's ultra low cost, ultra high capacity storage. What Holomem can bring to the table is kind of even more capacity and even more robustness in the media. So we've been intentionally modelling our drive, our cartridge, and our general format around the existing infrastructure that's used with magnetic tape. So our cartridge being compatible with existing LTO infrastructure is a very intentional move because there's so much of this hardware already out there, and we want to be a low-friction kind of complementary companion product to existing magnetic tape. We're in the world of deep archive, our media is right once media, so we are unapologetically worm, we are immutable at a media level. That's perfect for long-term archive. Our cartridges and our drives can work in parallel with existing LTO drives and cartridges within the same library. So I think lots of users that already have large data sets that they're keeping on tape, we can offer an additional tier of storage that works with their existing infrastructure, plug and play, but gives them additional capabilities for their extra long-term data sets.

SPEAKER_01 7:09

The current issue with LTO, which we're all familiar with, is every so many generations you have to upgrade LTO, you have to move to the next generation because of its backwards recompatibility. And your product overcomes that. Yeah, yeah.

SPEAKER_02 7:24

I mean, so magnetic tape, they have uh, I mean, we're on LTO 10 now, right? So it's been running since the early 2000s in its generations, always stepping up capacity. And it's incredible, and their their roadmap will continue to do so. The the challenge they have is by virtue of having a magnetic head that writes and reads and improving that head design and improving the media, the generations of drive and cartridge are intrinsically linked. There was once kind of a uh ability to be somewhat backwards compatible for a few generations, the latest generations are not, so they're they're always linked. The benefit we get at Holomen, because we are an optical data storage system, is our write head and our read head are not going to be locked to the generation of film that we use. So we will always be backwards compatible with our media. It also allows us to be able to offer a bit more flexibility with our drive formats. So traditional LTO is always a write and read system together. Because our holographic write engine and the technique we use to read holograms are separate, we can actually execute a classic write-read drive, but also a read-only drive. The benefit for us is that's a much lower cost system to be able to produce and represent. And it also offers the customers then an opportunity to have much greater read-throughput in their system. You'll know in these uh LTO um libraries, you have so many slots available for the drives that are normally not utilized because the drive cost is so high. We can say you can fill these now with Hollomen read-only drives and offer yourself so much greater read-throughput, which for the high performance compute kind of group and other hyperscalers is such an interesting use case.

SPEAKER_01 9:06

One of the things that is interesting about LTO is obviously we've got LTO 10, which currently holds 30 terabytes natively. We've got LTO 10E that gives you 40 terabytes natively. But your what where are you going with your capacity roadmap? Because you mentioned it's capacity, everyone needs to store something sometime.

SPEAKER_02 9:26

One of the core virtues of holographic is is such great data density because you can do multi-layer storage. What HoloMem's looking to do is bring out its first product in 2028, which will be a 50 terabyte cartridge. So looking to kind of go toe-for-toe with the latest generation tape, but offer that greater long life in storage and much more resilient and robust media. From there, our roadmap to higher capacities is really strong because we're adding more and more layers to our data density. So we deal with multiples rather than kind of organic growth. So we see line of sight to hundreds of terabytes per cartridge, um, if that's what the market is interested in.

SPEAKER_01 10:02

In my career, I've I've come across many, many organizations that are storing data. Well, many years ago they stored them on NetApp. One springs to mind was the Coal Authority. They bought uh an optical archive, um, I think it was DVD at the time, and they kept it on there, and then they moved it to the cloud. But there's a lot of noise and a lot of movement at the moment about the uh US Cloud Act, about where GDPR and data sovereignty begins and ends. Because if you read the Cloud Act, where it ends in terms of how secure is your data, and as far as I'm concerned, this could sit on a shelf and no one can access it. From my perspective, your product is a very long-term archival device that is offline, so can't be hacked, can't be tampered with, and it's worm, so you can't even edit it or change it or anything. So, your device, in terms of the cost and what it brings to the market, especially for archives, like I've worked with the BBC archives before, I've worked with very many um archival organizations over the years. Yours is like the panacea. I know there's some other if you want to say about other technologies that are out there, then what are your thoughts in terms of where the market will go from an archival perspective?

SPEAKER_02 11:27

Original data creation is is not going anywhere, right? It's growing exponentially. And our desire to store our existing data for ever longer periods of time is also not going anywhere. And you even look at the National Archives that are still digitizing uh lots of their collections, the desire for ultra-long life, ultra-low total cost of ownership storage is really increasing. So it's maybe the lesser um told part of the industry when everyone's getting kind of very excited about ultra high performance flash storage. What we're really looking to cater for is kind of um, we sometimes call it the uh you talk about the AI boom and uh data creation skyrocketing. We're here for the hangover of all the data creation that comes up. Gotta put it somewhere. So this air-gapped, uh ultra-resilient, ultra-low-cost, stable data storage is where all that data goes eventually.

SPEAKER_01 12:21

Another question that always interests me is data archivists, they've got petabytes of data, uh, the BBC Archives being one of them, National Gallery being another, Imperial War Museum another. And the this data doesn't need to be altered. It's history. So how can HOLOMEM help those organizations in terms of data migration from what they've got? Because they're not all on LTO tape, but how can they help them move to HoloMem, for example?

SPEAKER_02 12:54

I think there's two parts to it really. I mean, we're talking about these kind of um really important data sets. Uh, we we we need to retain this data uh and to lose it is unthinkable. And with existing generations or or tiers of data storage, the anxiety about losing this information and that you hear the term data rot being used in the industry a lot, um, it is a real concern and uh results in a much higher total cost of ownership for managing these data sets. Large amounts of cooling are used to maintain archival libraries. There's a general dichotomy of you want to check that your data set is okay on magnetic tapes. So you want to read the tapes to verify it's still there, but reading the tapes accelerates the data rot. Um, and I was once given the analogy of uh checking your archive is a bit like checking if your red wine has been caught. By the time you find out, it's too late. What Holom M offers is a much more robust media, so it doesn't require any active management in its archival state, there's no data rot, and reading the cartridges doesn't damage the data set at all. So it's far more kind of risk-free long-term data storage. What's more, I think archivists face this challenge that there will often be on existing generations of magnetic tape, right? So they have their library, their cartridges, and their drives. They need additional capacity, but to move to the latest generations of tapes, they have to rip out all of their existing stuff and bring in the new because the cartridges and the drives are linked in generations. Whereas Holomem, we will be able to bring additional cartridge capacities to customers without having to change the drive. So the upfront CapEx will be that much lower. So it's a future fit data storage tier.

SPEAKER_01 14:38

One of the other questions that a lot of organizations face is where to put their data. Currently, a lot of them moved away from tape because of its capacity and the incremental changes and upgrade of drives and next generation, everything else. And they put it in the cloud. And we all know storing archived data in the cloud, whilst it might seem a good idea, is you know, weighs heavy on your pocket. You know, every month you've got to pay, and if somebody ups their price by 10%, guess what? You're paying an extra 10%. And on top of that, you've got the ingress charges of putting it up there, the egress charges of getting it out, and that adds up. So, what are your thoughts on long-term cloud archiving?

SPEAKER_02 15:22

Part of the context here is that hyperscalers also use tape, right? It's uh it's it's not often declared or acknowledged, but yeah, all of the hyperscalers will have the same three tiers of storage at their disposal. Um, and in fact, AWS just announced that they're building a 20,000 square foot tape-only data centre here in Hemel Hempstead in the UK in the coming years. So um, it's called Glacier. Uh, they won't call it tape uh upfront, but that is exactly what they're doing because tape is still in a class of its own when it comes to this low-cost, long-life data storage. What Holomem offers is just a again, a companion product that uh that plays nice with that existing infrastructure, but offers lower total cost of ownership, lower OPEX costs, and hopefully then represents to the customer a lower ingress and egress fee as well.

SPEAKER_01 16:12

Exactly. Because, you know, a lot of organisations they have not terabytes but multi-petabytes of data, and where they put that data, you know, someone say oh it's gonna be good for 20 years or 10 years, whatever, and I'm not gonna be around. That's that's typically what they say, you know. And it's probably true, or you know, they they get a data breach and they get fired and they move on, and then somebody else is left to pick up the pieces, or and no one really has from my experience a clear strategy about data archiving. And I and I think HoloMM as a technology solves a lot of those dilemmas that companies face, and also from a from an OPEX cost, you know, for it's very green, you know, LTO is very green, it can sit on a shelf, you don't have to keep it in a library with fans spinning all the time, and it doesn't take any cooling or heating or whatever it happens to be, you know. So that's where I see Holomem taking a front foot, if you like. So, from your perspective, what other competing technologies are you seeing in the market as a replacement for tape, for example?

SPEAKER_02 17:25

It needs to be said that we're we're not the only show in town. Uh, we're also not the only company that believes the future is optical when it comes to long-term archive. Um, since Holl and M has formed, I think there's at least a half a dozen new new uh players that have entered the scene. And um uh we we share their view and their vision that there needs to be a new tier of long-term archival storage technology that tackles some of these issues about capacity, about robustness, and about kind of operational costs with long-term archive. The difference that Holomem brings to the table is this plug-and-play kind of low-friction approach to working with existing infrastructure and protocols. You'll see projects that are talking about long-term glass storage, kind of flat platter storage. Silicon, yeah. And they've achieved incredible things. They present as research projects using research grade equipment. So uh whilst there might be an amazing result in a lab, how that actually results in representing a benefit to a customer at cost and being deployable to an existing operational environment is yet to be seen. I mean, HoloM has very intentionally looked to develop a product, not a project. So the Holodrive is using off-the-shelf, low-cost components. Our drive and our cartridge are compatible with the existing infrastructure. And to really demonstrate this, we are now being quite front-footed with pilot deployments of our system. So late last year we took the holodrive, we built it into an existing tape library, and we deployed it to a data center. So working fully autonomously in an operational environment as a working system for long-term archive.

SPEAKER_01 19:03

So you can get some NTPF figures.

SPEAKER_02 19:05

Correct.

SPEAKER_01 19:06

And are you, you know, to our to the viewers looking at this video, are you actually looking for pilots, large pilots, scroll your product?

SPEAKER_02 19:14

Yeah, we we love getting out there, we love learning. Uh the fact that the holodrive is live now. We have these systems, uh, they just about fit in the boot of my car. And so we're we're looking to be um yeah, really engaging with the with the community. So we have archivists, we have the high performance compute industry, um, we've got some fantastic connections, but we're always looking for more uh to yeah demonstrate our use cases, demonstrate how credible and robust our technology is.

SPEAKER_01 19:41

How should industry leaders talk about or think about data permanence, especially organizations that are creating sensitive data, it could be nuclear data, it could be intellectual property data. How how do you think they should be tackling this challenge themselves, you know, five or ten years from now or even 20 years from now? Because if you know, if I had a crystal ball when I first started in this industry, I would never have imagined it to do what it has actually done and become. So, what are your sort of thoughts on that?

SPEAKER_02 20:13

Many archival data sets are not uh just a nice to have, it's a requirement. So that there are regulations around the the long-term storage of certain data sets. Um, and you'll know LTO actually comes in two flavors at the moment. There's the the rewritable cartridge and there's the premium worm cartridge when you are uh required and regulated to store an immutable data set for a long period of time, you use the worm LTO cartridges. With Holomem, you get the worm for free. So all of our cartridges are worm. We're unapologetically worm. And so to have that data set immutable at a media level means that it's so relevant for long-term archives. Um, we're talking about banking records, medical records, um, all forms of regulated data sets.

SPEAKER_01 21:02

One of the interesting um subjects that's come to light, and I I deal with a number of organizations, but one in particular springs to mind is everyone's talking about the AI boom, you know, oh, we're looking at AI, we're looking at AI. But something uh something resonated with me a few weeks ago, whereas this guy who I deal with, he said the problem for us is we don't have our data in one place, it's on spinning disk, it's on RAM, it's in the cloud, it's all over the place. And trying to get all this data together in one place is quite important. So, so from my perspective, your technology would be able to store that original AI data set unaltered. So if something happened from a compliance point of view, a legality point of view, a data breach point of view, then you have a reference.

SPEAKER_02 21:57

AIs need to feed on large. Models and large data sets to inform their decisions. And I think having an immutable data set and using our technology as an active archive for AI to draw into and feed from is incredibly powerful. And you're right, the fact that it's truly immutable and non-modifiable actually works as an advantage in that use case.

SPEAKER_01 22:22

So, your technology, what's your roadmap to deliver? So, for people looking at this, they're obviously planning five, ten years in advance. What's your delivery roadmap? Obviously, you've mentioned that you're looking for active pilots. Um, what where are you in terms of your actual deliverables?

SPEAKER_02 22:41

Holomem's a deep tech hardware startup, right? And uh since forming, we've been building uh building systems, building drives to prove that what we're saying is real and that it works. Um, we finished last year with our first public deployment of our system, as we say, to a data center here in the UK. That is our our proof-of-concept holodrive system with the cartridges that work with it. Our journey now is to continue these pilots. So throughout 2026, beginning of 2027, we're going to be tapping into all of our end users and kind of scenarios where we think the holodrive really shines through uh with its performance to build that credibility and build these relationships in the market before ramp up and entry to the market in 2028.

SPEAKER_01 23:26

One of the things that has always fascinated me to a certain extent is in order for your holo drive to work, you obviously have to have a SaaS interface or a fiber channel interface, and it's the software vendors, the ISVs, that you need to get certified with. Or are you gonna is that something that you're currently engaged with as well in getting ISV certifications from all these vendors?

SPEAKER_02 23:54

We we realise we're part of an ecosystem here, right? Exactly. We very much want to uh play within that uh rather than redefine it. Um, we want the holodrive and our HoloMem cartridges to be usable in the same way that people use existing LTO infrastructure. So we're partnering with the existing software vendors, the existing library manufacturers, so you can really use HoloMem as an alternative to LTO or even in companion with LTO. This is um evolution, not revolution.

SPEAKER_01 24:24

Yeah, and I I can see mega benefits in this because obviously you've got your LTO tape for backup, for fast restores, for your Air Gap, you know, and then you've got in the same library, effectively, you've got your long-term archive.

SPEAKER_02 24:40

Exactly, and and it's horses for courses. Um there's there's real benefits to what LTO can currently offer with its ability to be rewritten as well for backup. Yeah, that will always be a good product. We're here for the long-term archival data sets, which is actually the largest growing amount of uh data requirement there is. Um, but no, to have these as companion products is actually incredibly valuable.

SPEAKER_01 25:02

Yeah, and and the software vendors that you mentioned, they will there they should be able to take advantage of that's for backup, that's for archive. Totally. Should be able to differentiate between this because it's quite important. If you had one piece of advice to give to leaders today, obviously we can't prevent data growth, but what advice would you give to them? I've got I've I've got a little opinion on this, but you can, you know, what how do you think they should be looking at the future?

SPEAKER_02 25:28

I mean, uh, I guess Holomone's here to say uh rock on with data generation. We're here for it. Um, I think um I think data uh is one of our greatest assets as uh as a civilization, right? And um it can be so beneficial and so powerful. What we need to make sure is the way that we're managing and storing that data is sustainable uh and practical. And currently the infrastructure is simply not sustainable throughout the entire stack, right? So you yeah, you look at uh day-to-day bandwidth operations, how data centers are struggling to keep pace. Holomem in its sector, in this kind of long-term archival sector, is here just to breathe new life and new capacity and ultimately present a future fit solution.

SPEAKER_01 26:11

Yeah, I agree. But also a lot of organizations, especially in light of what's happening with with market forces that are outside our control. So, for example, WD have announced that they're not selling any, or all their everything they make now is sold out to 2026, 2027, which is ridiculous, you know, and the and the and part of the part of the issue is I think a lot of business leaders think, oh, we could just go and buy it. But what they you know, even if you place an order today for an MVME drive, you probably may, you may or may not get that drive for six months. So the pressures on storing data, everyone's creating it every day, this video's creating data, you know, everything's is becoming increasingly complex, and having something that you can actually free up that tier of storage makes a lot of sense.

SPEAKER_02 27:10

Heard anecdotally that some flash storage is worth more than its weight in gold these days, so uh data really is the new gold. Um, I think what Holomem really is presenting here, given that we are using a low-cost, abundant polymer to store data, it really kind of it takes the sting out of these uh out of the the cost and the energy required to store these ginormous data sets.

SPEAKER_01 27:32

Yeah, absolutely. And it's you know, we've got 8k videos, they they consume hundreds of terabytes. You know, there's a lot of applications for an archive that people may not consider or may not think of. Yeah, but I think, you know, from from my perspective, your technology could make this the next step in terms of archiving data and moving it off very easily, you know, and without a lot of complexity. You know, we're not talking about reinventing the wheel, are we?

SPEAKER_02 28:04

Totally. And I I think there's um somewhat of a stigma associated with archival uh infrastructure and data that it's somewhat dusty and old school. Um, I I view it totally differently. That uh I think this is actually some of the most kind of interesting uh parts of the ecosystem, and actually there's gonna be more innovation happening here than anywhere else. I mean, archival data sets and the utilization of existing archival hardware is growing like never before. And so I think Holomem's really here to uh uh yeah to bring uh innovation back to archival data storage.

SPEAKER_01 28:38

I think a lot of organizations, when I've spoken to them, they say we back up everything. But that's all well and good, but that costs some money, whether it's on hard disk, whether it's on tape, whether it's in the cloud, they back up everything. So they're not actually looking at it from a point of view as how can I actually save some money long term, you know. Data that what's the statistic? Data after two years, 80% of it's never accessed again. Yeah, but we'll still back it up, you know, and that's where I can't understand the mindset of organizations where they should be looking at data as an asset. So if you're accessing data daily, that needs to be on primary storage. If you're accessing it once every six months, maybe you've got that on hard disks. But if you're accessing data once every 10 years, then why are you keeping it in the cloud? Why are you keeping it on spinning disks? Why are you backing it up?

SPEAKER_02 29:31

And I I think if we if we can offer functionality and kind of flexibility to allow this kind of the decision makers to really trust an additional tier of storage that offers these lower costs and offers these reliabilities and reassurances, then hopefully we can start to inform change.

SPEAKER_01 29:48

Brilliant. Well, thank you, everybody. I hope you enjoyed that from HoloM. Charlie, thank you very much.

SPEAKER_02 29:53

Thank you, Ray.

SPEAKER_01 29:53

Take care.

SPEAKER_02 29:54

Cheers.