In The Arena by TechArena - Dell Leads the Entertainment Industry’s Storage Revolution

Episode Date: August 8, 2025

Dell and Solidigm explore how flash storage is transforming creative pipelines—from real-time rendering to AI-enhanced production—enabling faster workflows and better business outcomes.

Welcome to the tech arena, featuring authentic discussions between tech's leading innovators and our host, Alison Klein. Now let's step into the arena. Welcome in the arena. My name's Allison Klein, and we're here for a Data Insights episode. And I'm so excited because I've got two. great guests with me. Alex Tim's senior business development manager of media and entertainment at Dell. It's Scott Shadley, leadership marketing director at Solidine. Welcome to the show,

guys. Thanks, Alison. Thanks for having us. Great to be here. So Alex, why don't we just get started with quick introductions? You can go first and then we'll get to Scott. Sure. Well, my current role is business development manager for Dell Technologies focusing on media and entertainment. But prior to that, I spent 15 and a half years working for Animalogic, now known as Netflix animation. And there are many animated and visual effects films I worked on there. I had the privilege of working with the company and growing from, I think it was around 80 people when I started at the beginning of Happy Thie. And when I left, I was about halfway through Peter Rabbit 2 and the company globally

was over a thousand people. So it was an amazing phase of growth and made some amazing films and worked with some amazing people. And Scott, why don't you go ahead and introduce yourself? I can't really follow that very well, but I've been in the semiconductor industry for close to 30 years now. I've done a lot of work on building all the technology that supports all the really cool stuff that Alex has done in his career. A lot of it focused around NAND and storage and DRAM and things like that. But definitely know the ins and outs of everything up the full food stack and how we can have a lot of fun with this different technology. No, I just want to start.

And I'm still thinking, Alex, about the fact that you get to say happy feet as part of your introduction for your career, which is pretty cool. But I do want to start with just talking about AI. And I know that Dell knows a ton about what's going on in enterprise. Obviously, AI is clearly transforming the landscape for that computing. How does that translate into specific storage demands from speed to scalability to even reliability? So what I've seen personally is that AI workloads, and for that matter, real-time workloads for things like game engines, demands a lot from the GPU. specifically V-RAM, and so it needs to fill and flush that V-RAM really quickly. At the same time, we've seen the amount of V-RAM and GPU going up substantially.

We used to see 24-gig, for example, being adequate for a lot of workloads. We're now seeing a lot of customers hitting that 96-gig and wanting even more as the resolutions for these workloads go. So in order to fill and flush that V-Ram, you need to have very fast storage. And so I've personally seen an industry go from Flash, it would be really nice to have for certain workloads to you must have Flash. It's sort of the default choice for one of these workloads. And so I think that would be the single biggest driver of storage shift that we've seen.

And it doesn't need to be huge. We've seen customers that are small to medium size, starting off with very small deployments of Flash storage and doing quite amazing things. And I'm sure we'll probably cover some of that later on today. And we've also got larger customers that are on their journey and start. in a very large clusters of flash storage. And it's not just the V-RAM that benefits from this, in my own experience in these media pipelines.

You get this serendipitous performance improvements throughout the pipeline when you start adding flash. Things that you could have never quantified or never expected to happen. And they might be 30 seconds here or a minute there, opening and copying things, loading extensions for applications such as Maya, for example. All these things contribute to hours of performance improvements

across the production, man hours, every hour, for example. And so you start getting these compounding value propositions that you didn't quite account for. So this serendipitous performance improvements and the improvements in human collaboration as a result of being able to do things faster is something that starts to have huge impact on the ability to iterate and to collaborate a lot more effectively. Now, what role do you see Dell playing in terms of shaping this innovation way with your customers. And how are you bringing technology and customer need together in practical ways?

Dell Technologies as a whole is approaching the AI boom with something called the AI factory. And the AI factory as a full stack, fully validated solution that brings together best-a-breed software, hardware, and invalidated manner, and delivers that with certainty. So customers have an easy button, I suppose, for AI. And this is not necessarily, as the word would indicate a huge factory and a big investment, you can start very small with small t-shirt sizes, as I said before. And we have a range of storage solutions that underpin that, starting with our power scale F210 cluster or node type at this point in time, all the way up to the F-19 for those much larger deployments, much higher capacity densities. And in doing so, we're trying to make the

journey for customers an easier one for AI. And we also have a number of partnerships that we bring together with everything from disk providers to make sure that we have the fastest, most reliable underlying SSDs, all the way through to invidia relationships, ISB partnerships, and making sure that we're able to deliver a choice for our customers across the entire staff. Now, Scott, let's bring you into the conversation. Alex has provided a pretty clear view of how storage is evolving to meet all these needs, and Soledime obviously sees a role in enabling vertical and

through high-capacity SSD delivery. How do you see that manifesting, whether it's AI analytics or real-time workflows that customers are tackling? Yeah, actually, I really like the way that Alex introduced the fact that it used to be Flash could be used, and now it's Flash must be used, if you will. It's a perfect way to think about at one point in time it was the idea of Field of Dreams. If you build it, they will come is how this whole industry started when it came to the Flash and what Solidime is doing.

And now we're at the point where it's not so much that. It's what does your customer? So Alex is famous for the why question, right? And our job is to help with that why. And by having a plethora of products that all serve specific workload use cases and needs, from a capacity point to a performance metric, allows us to see the fruition of our joint customers success and that human interaction, the ability to cut little bits here and there

and to shift that bottleneck into a place where it can be better managed is something that the storage platform that we're working with and developing to support Dell and others is just mind-boggling to see how well it's working to look at it from a customer point of view back. And that's where our focus is back. Now let's pivot into some real-world examples. This has been great so far, but we want to get real. The media and entertainment space is one area where storage innovation has been moving at a very fast clip and has had a massive impact in terms of what people can do. What kinds of storage challenges to companies in this vertical face, Alex?

And how has that shaped your solution strategy at Dell? Well, the industry as a whole has generally been what I would call bleeding edge. The industry is always looking for the fastest, best way of approaching and solving problems. But at the same time, it's confronted by being for need of survival, being very frugal. They have to do things in a very clever way. And so you quite often see a lot of innovation coming out of media entertainment. They're often the first to try things, to try new ways of doing things, to try new technology, which includes storage, and try to ensure that it's used in a way that doesn't overcapitalize,

that's scalable, that's modular, that has reuse potential in other parts of the pipeline, and that benefits as many interactions or connections or collaboration, if you will, as possible throughout a pipeline. And we've got great examples of the actual impact that this can have on media pipeline, specifically Flash storage. Now, going back to my career specifically, back when Flash was optional to put it one way, I remember a time when we introduced the first Flash solution, which was a small amount of Flash that we're introducing to be a cache.

And that cache was going to have a number of data sets pinned to it. Data of the type, what we used to refer to as production cache, which is things like textures that are used across the pipeline by lots of different processes by all the artists. And in addition to that, we identified network software, which is all of the versions of software and plugins that were proprietary and unique to the pipeline that were loaded for any process within the pipeline. And just to give you an example, when an artist opens a 3D application in a company such as Animalogic, there would be an additional 300.

plus extensions that were loaded in real time when they opened that up. And when they jump between applications, you can imagine how much is this happening. The same thing would happen with render processes. All of this is pulling from network software, and the reason you want it centralised is because they're writing maybe seven plus changes to software every 24 hours. It's a constant process of change. And so we identified production cache and software that was pinned to this small, I'm talking, a couple of hundred terabytes worth of flash storage.

And we got immense gains. This is back to that serendipitous performance gains that we used to speak of. And this accelerated everything throughout the pipeline. The users that used to go in and launch their environment by clicking on a button sometimes used to take 15 minutes on traditional storage without that cage. That went to a minute or 90 seconds after adding this. So you imagine the over hundreds and hundreds of artists that they're not necessarily cheap people. and time is money.

Collaboration is key to successful outcomes, creative outcomes. And so this had a huge impact. So that was my very early experience with the potential for Flash and how quick the ROI was on that. And we have other recent customer examples that we may want to go into more detail later. We had examples of customers going again from traditional storage solutions to Flash

and accelerating their pipeline, in their own words, I found out this week, in their own words, 100x, what they were previously able to do, 100 times, not 100%, 100 times. So that means they could do 100 times more work in that given time, which meant they could iterate more, which means they could do more creative versions, which means they can get a better creative vision of the director down so that you get better creative outcomes. Or they could do a lot more work in that given time with the same number of people. So these are really impressive numbers, just simply from a tweak,

the type of storage that you're using. No, I know if we fast forward in time, it's not just the media itself that is creating change in these environments. It's also AI enhanced workflows and how the entertainment space is navigating production work, for example, and trying to compress that to save time and money. Can you talk about how Dell has worked with customers in this perspective and what your storage technology contributes to their success? Yeah, sure. So, I mean, AI is all the rage. And as I said before, it's not everything. There's also real time, which is having a huge impact, I think, it has for a while. And this includes things like in camera VFX, virtual production, but also the use of Unreal Engine is a means of getting immediate creative feedback along with use of AI and emerging agentic AI use cases. So we're very focused on ensuring that customers are walked along that journey. You know, it can be quite daunting to,

consider what might be required to deliver on that. But customers are starting to see, and we have references that what others are doing and the potential benefits of that, the ROI, if you will. And so we're better able to talk to how to get to an outcome with that. And customers such as Kennedy Miller Mitchell that are a public referenceable customer, customers such as Heretic Foundation, which is Alex Prois's company, they have very real and quantifiable returns on investment that we can help use as a means of educating customers on what the potential is and therefore walking them through what some of this technology may achieve for them and some of the strategies that they may put in place that are based around the Dell AI factory and how that can help

them with low friction get to the outcomes that they want. That's awesome. Now, you know, one thing that I've been thinking about while you're talking is let's just break it down a little bit. When you look at high-resolution content, like large-scale video graphics, for example, you have stated such a compelling performance message in terms of the breakthrough year. But what makes Flash so critical in these environments and what is the difference got in terms of comparing what it can deliver from a performance perspective versus traditional technologies?

It's an interesting thing to think about people are all, you know, the hard drive spins and the flash doesn't type of thing is the basics, right? But at the end of the day, the ability to get concurrent information. So Alex brought up the fact that you load one thing and several people needed or a whole bunch of other things behind it have to open at the same time. And the ability for Flash to have that concurrency and the density of being able to launch all of that information in real time is valuable. Because it may be a large single thing that you're looking at, but all the metadata, all the little things behind it that have to come alongside it at the same time are very valuable. And that's something that you can really only do with an environment that you can get

that level of concurrent and random access for a very large sequential file, if you will, to some extent. No, Alex, I know that customers need to strike a balance between performance and capacity in this space. How do you work with customers to figure out what is the right blend between high performance SSDs and maybe higher capacity traditional solutions? Understanding the workflow, I think is critical. So you need to be in the conversation stream with the customer. DEL also brings to the table, you know, not people like me, but there's a lot of others in my team.

So we've got over 100 years of experience in M&E across broadcast, game development, and the effects and animation. And so that subject matter expertise means that we bring to the table a deep understanding of the workflows, the type of data, what some of the challenges are, and strategic ways of dealing with some of those challenges. and it isn't always flash for every use case. Sometimes it makes sense for spinning disk. Flash is like an F1 car, and spinning disk is like a delivery truck.

One's really utilitarian and can carry a lot more, and the other one's really fast and gets you there a lot quicker. So you don't want to use an F1 car necessarily to do a bulk delivery of something to use this analogy. And so sometimes spinning disk will make a lot of sense for a lot of capacity that's much slower as a means of staging something use case. near line storage, for example.

And so we see a lot of hierarchical storage approaches where customers will put the active data set on Flash and then they'll stage down to a spinning disk or hybrid disk solutions, which are a combination of flash and spinning disk. And that may even go further off file down to object and that can be on-prem in the cloud adjacent to cloud. And so it really comes down to where the data needs to be,

who needs to access it, and how fast that access profile is. And then understanding that will help tease out what the best architectures are for the customer in those situation. Now, Scott, I know that Solidim does a lot of work with your partners and your customers to ensure that the deployments of solutions that feature your products balance architecture from a performance and scale perspective.

I know you've talked about the fact that you've got an entire portfolio to yield. What are those collaborative? look like? It's a really important thing. And when you think about it, a lot of people see that storage is the lowest common denominator, at least a lot of these platforms. So it's an afterthought or it's something that I know I'll need at some point. And when you start getting into a transactional type of environment, you can't really get the most benefit out of it. So when you have Alex as a partner or Dell as a partner and you want to go talk to an end customer, we need people on our side just like he does with the subject matter expertise. We've got a lot

of people that are solutions. Architects, solution experts, work in the software stack to understand which products fit. And he's absolutely right. A flash drive is not, or an SSD is not the perfect fit for every workload. And we understand that. And we're not in a position to try to upsell or do anything like that. We're not playing the game of the, you know, spinning disks are dead again. But it's just talking to the customer, understanding the data, the workflow, what they're expecting to do with it, and what that stack needs to look like. So we can put the right match together and make sure that we're not just overselling or trying to make the fastest thing out there. It's about right sizing a solution for the customer. So the customer gets what

they need out of the platform. To Alex's point, the right cost structure as well. Because at the end of the day, you've got to be able to make someone be able to afford and take advantage of what you're trying to build without just saying, here, this one looks good. You need to actually be able to do that. So we actually have a Solutions Lab at Solidime where we can bring those types of workloads in and do some analysis on behalf of the customer before they even have to buy, kind of like a try and buy. And we're also outfitting an AI lab where we're going to actually have a place where customers can log in directly and utilize different layers of architectures all the way up to the CPUs and the GPUs on Dell platforms and the like to be able to solve those problems

in a way that's collaborative and cooperative to make sure they get what they need and not just what I might want to sell them, if you will. It's really cool. I love the idea of being able to utilize the technology before you deploy or even bring something on program. That's very cool. Alex, I know that you're navigating a lot of very complex operations with your customers. What advice would you give to a customer looking at complex storage decisions across multiple locations or departments, some of whom might have different needs? How do you get your arms around that?

I would say understand your data as best you can, but most importantly, particularly in media and insane, but I think it applies to any organization, any vertical, to be honest. This is fundamentally about understanding your connections. It's about understanding how people interact where the information needs to flow. And that includes systems, things that aren't people, but it's how everything's connected. And in this sort of post-COVID world where there's more geo-distribution than ever, that's never been more important. And understanding those connections, understands where the data needs to be, how it flows,

how fast it needs to be. And as you start to unpack, that'll give you indications about what you may need at edge versus what you may need at the core. So you might need a really perform an edge at multiple locations, but you might only need a much slower core. It'll help inform your capacities. The available, even things like data set, understanding your data center capacities makes a big difference. Can you host additional infrastructure in your data center? Do you have a cloud strategy? Do you have an offsite data center partner that you need to work with? What are the capacities there? And as you start to tease that out, to appreciate that I think the ID

see numbers still pretty much valid. I've checked it with other customers, but it's 23% I think compounding growth per year, which is huge. And then we're getting even more expected as the outcome out of AI workloads. And so this is only going in one direction. So think beyond just the actual storage cost, the initial investment, start thinking about the serendipitous performance gains you're going to get by saving 30 seconds or minutes here and there across the board every hour, you know, in the way you collaborate with people, start thinking about potential power savings as the density of the flash affords you. You might be able to reduce racks and racks of spinning discs down to half a rack of flash or less. And because there's no

appreciable change in the actual power draw, depending on the capacity of flash drives, as I understand it, it's marginal in terms of the amount it goes up. As these capacities increase, your power savings and the value you're getting there, are only going up and up and up and up, which is a major consideration as we're running out of data center space, power and sustainability is also consideration. So start thinking beyond just the initial investment, think about the density, the power savings, and obviously the performance and the value that delivers to your pipeline. I'd say that would be where I would recommend customers start to unpack it. And also think about today, not what your immediate requirements around storage are, but where

they're going to be in five years because they're only accelerating in terms of the demands. You know, you might be working in a pipeline that's doing grade in, for example, at 4K, but in two years' time, you're going to be working natively at 8K, an uncompressed 8K for DIY. Think about how you're going to be able to service that. Don't buy something today that won't scale in terms of performance and capacity for tomorrow. I have a couple drill down questions for you, Alex. You did mention AI, so we're going to go there first. I think that we've seen the implementation of a lot of gen AI integration into enterprise apps and media and entertainment is no exception there. When you think about AI in terms of

storage and you're seeing storage requirements changing, depending on whether or not we're working with training or inference, how do you advise your customers on how to navigate that? Well, trainings can be very different. If you're training a model from scratch, that's a vastly different paradigm than if you're inferencing or creating a fine tune using a law or some other approach. And so most customers, unless they're the very large enterprises, are probably not going to be building a model from scratch. If you are building that model from scratch, again, it comes back to that AI factory discussion from a Dell language perspective where we'll have the ability to unpack that with the customer. And if you need highly parallelized high performance access

from thousands of GPUs to be training a model from scratch. That is a very different conversation, probably with a very different customer than you would be having with, say, a customer that wants to use an existing model with a laura attached to it to be able to produce some novel images that are aligned to their creative vision.

And so these are very different storage decisions and architectural decisions that would be taken. And that is drawing it full circle back to the AI factory conversation. That's what it's about. It's about helping customers understand the t-shirt size that fits their use case and taking them on that gym. I love that analogy.

There's one other question that I wanted to ask you. Media and entertainment is also known for having a lot of needs at the edge, maybe not as much in animation, but also when we're considering live action filming, often people are collecting that data in a very large variety of locations. How are you seeing Edge evolve, especially within the realm of AI implementations? at the edge. And how do you see storage requirements evolving at the edge, especially with all of that data creation in that space? So firstly, there's the ingest. So there's higher and high resolution cameras that are becoming cheaper and cheaper, more accessible. So you can have

more of them. So instead of having two cameras that are on a shoot, you might have 10 cameras at 17K or higher that are doing these shoots. So they're generating immense amounts of data. And you want to be to get that data off those cameras at the edge. So right at the edge, the extreme, on a shoot, on location, customers want to be able to suck that data off the cameras. Once you've got it off the cameras, you want to be able to potentially do select. So you want to start actually making creative decisions with the director about which camera you want to take, which particular cuts you might want to put in there. There's a really rough things, but you want to start distilling it down to get some value. Then you've got these new workloads to your point, which is maybe I

want to do a bit of in-camera VFX in close to real-time or in real-time on set. So while I'm shooting, I actually want to have a CG background included in the monitor that the director is looking at. So he or she can make creative decisions about where the actors are sitting or what props or visual effects maybe needed where. And so that requires high-performance flash storage to be on set, again, to be able to take that stream of data off cameras, to be able to do some kind of work on it and to then be to stream that back out to a monitor. So we're seeing more and more performant edge and potentially

performant core as well, but certainly mixtures of architectures where I think most of the amazing things are happening at the edge these days. You want to bring that capability to where that data is being captured, acquired or processed. And that can be on set. It could be in multiple locations simultaneously. We're seeing some customers that are wanting to have an acquisition on a camera or multiple sensors outside of just camera, could be light hour, et cetera, and they want to stream that, capture that at the edge onto fast storage, and then immediately stream that to another edge point or a core, and immediately start working on that data stream, either for things like real-time broadcast scenarios or for defense applications. You want to be able to take all of that

sensor data and telemetry, et cetera, metadata, captured at the edge, and work on that in close to real time. And I don't see that going away. I see these workloads, the acceleration only increasing and the desire to be able to work on data immediately rather than having to wait is something that we're seeing a lot of in media entertainment and other protocols. Scott, you know, I think that what we can say is that media and entertainment is a fantastic use of SSDs after this interview. Dell is doing some amazing things to deliver our innovative technology. And the question that I wanted to ask both of you, before we wrapped, is what are the most exciting critical innovations in this space that you're seeing on the horizon? And how are you preparing for them? Scott, why don't you go

first? It's a very solid point. Definitely, the effort further from what we classify as a traditional core is a focus and understanding what that system level looks like. To Alex's point about camera type or size or dimension, I remember way back when I first got involved in media entertainment, I was providing solid-state drives for red cameras, and it was in the gigabytes of density per camera. And now we're talking terabytes per camera per node, per everything else. I've even had a little brother who works in the industry who was doing the monitor work that Alex was talking about. And he had a nice server set right there on his little cart where he was doing that real-time collaboration for the producers and the directors to be able to see that

kind of work done. So being able to continue to emphasize the ability, to go further away from our, quote, traditional data center into these environments where we've artificially and or just naturally have less resources available to us. We have to be able to figure how to do more with less. So less power, less space, less weight. Think about the guys that are running around with the cameras on their backs. That data's got to get somewhere, but it's streaming at such high resolution. So these kinds of innovations in media entertainment as it shifts into everything else that we're dealing with and the creative things that are happening in the health care and what you're trying to capture from the new MRIs and x-rays and all that kind of stuff.

It's an amazing opportunity just to see this focus on outside of that comfort zone. I spent a lot of my career helping people outside their, quote, comfort zones into these new and in and way of ways of doing things and making sure that they're successful because it's not about me being successful. It's about the customer success and be able to see them go thumbs up to what we were able to help to work with. So you've had a very interesting ride in the media and entertainment landscape. What are you excited about on the horizon? Yeah.

So I think fundamentally the last few years has been very much in media entertainment specifically, has been very much about risk and it's about a reduction in risk. So I think these technologies and the immediacy of the context they bring is about reducing risk. And I always try to draw it back to the shift that I expect. I think other people, everyone else did, that shift from analog film cameras, SLRs, to a digital camera. I remember the anxiety I used to have about setting up a photo because there was a real cost of failure there. If you took a whole reel of film that were all bad shots and none of them

were great, you'd spend a lot of time setting those shots up. You'd had anxiety waiting while you were taking that film to get developed. You'd spend some money getting that developed, only to then come back days later to have a look at these photos that didn't work out. There was a lot of risk in that. There was a lot of uncertainty. And I think these technologies, much like the digital cameras brought to the industry, it's the ability to keep your finger on the trigger. And I know one of those photos is going to be okay. So it means you can take a lot more risk and therefore it reduces the creative and financial risk associated with getting an outcome. And it means ultimately we can deliver better outcomes, more of them, and empower creative people to make

decisions much quicker. So I think that the future is all about immediacy of the data, understanding it, having context of it, being able to analyze and distill it, whether that data is text or whether it's images or sensor data, it's about bringing all that together for you to be able to make an in-context decision. And so I expect to see more and more of this happening at every level in media entertainment and for that matter every other vertical. You mentioned healthcare, et cetera. It's all about making the most informed real-time decisions possible. Alex and Scott, thank you so much.

This interview was fantastic. I learned a lot, and I'm sure our listeners did do. Where can folks go if they find that they would love to learn more about the Dell solutions in a space or learn the latest about SST technology innovation? Alex, go first. Yeah, certainly. I mean, the Dell website's the go-to site for our solutions. Obviously, if you are in the media and entertainment vertical, we are present at a number of the large trade shows globally, and so certainly encourage customers to come visit us there. And quite often we have some of your drives on display there, which is great. Customers like to be able to see those incredible capacities and talk about why Solidon, for sure. We are also available as a team, the subject matter experts team. If you want to reach out to me, if anyone watching this wants to reach out to me, feel free. It's Alex. Scott Tims at Dell.com.

You know, if you want more information, it would be connected to other people. Happy to do that. And we also have a range of solution briefs, which are aimed at explaining how our solutions are able to address some of the common workflow challenges in meter and assessment and other verticals. And we also have any account manager or Dell should be able to talk to the Dell AI factory narrative and also take you on the journey, if that's your desire. And Scott?

Yeah, and for Solidime, it's pretty straightforward. solidime.com forward slash AI gets you right to that page if you're interested on the AI side for the rest of the portfolio is also there we have a great opportunity to reach out and talk about all those different things we have solution briefs and product information as well as some of the collaboration work we've done with Dell in the media and entertainment industry specifically as well some other landmarks and engagements around edge deployments you can go read about our friends the hedgehog we really do have a story that's based on a platform with a software partner on a Dell box saving the hedgehog

So not quite Sonic, but it's out there. And it's interesting. And if you want to find me personally, I'm on socials at SM Shadley or Scott.shadley at solidime.com. Fantastic. Well, thank you so much, guys. It's been a real treat. Thanks for the interview.

Nice, Alice. Thanks for joining the tech arena. Subscribe and engage at our website, thetecharena. All content is copyright by the tech arena. Thank you.