In The Arena by TechArena - AI, Flash, and Data Protection: Scality’s Vision with Paul Speciale

Episode Date: November 25, 2025

Scality CMO Paul Speciale joins Data Insights to discuss the future of storage—AI-driven resilience, the rise of all-flash deployments, and why object storage is becoming central to enterprise strategy.

Welcome to 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 Cori, and this is another Data Insights episode, so that means Janice Norowski with Solidime is with me. Welcome back, Janice. How are you doing? Thank you, Allison. It's great to be back. I'm doing well. How are you? I'm doing great. I am so excited for this interview. This is a company that I caught up with at a tech field day, and I knew we needed to bring them on data insights. Do you want to introduce who our guest is? Yes, I am equally excited. We actually have Paul with us, Paul Special, our CMO of Scality. Welcome to the show, Paul. Thank you very much, Janice. It's great to be here with you and Alice.

So, Paul, do you want to start with an introduction of Scality at the first. time you guys have been on the platform and your role is CMO. Of course, yeah. So Scality is a vendor in the software-defined storage space. We've actually been around for 15 years. The company first came out with its flagship product, the Scality Ring, in 2010. I actually joined four years later. So I've been with the company now over 11 years. Originally, I was the product manager, then the chief product officer. Now I run marketing for the last five years or so. It's been a great journey. And again, we're kind of this pioneer early vendor, but now one of the vendors providing solutions for big petabyte scale storage within the industry. Awesome. So with that, Paul, with a focus

on big pentascale storage in the industry, can you tell us a little bit about data protection? Why is it top of mind for every enterprise today? And how do you really see the priorities of backup and ransomware protection evolving in this environment? Yeah, yeah, absolutely. This has changed a lot in the time I've been here. And let me say this, we've seen the need to backup and archive very large amounts of data for all the time I've been here, right? So this is a continuing need, especially you've all heard about the growth of data and especially unstructured data, right? It's the bulk of new data that gets created in the form of big files, images, videos, lots of sensor data. So there's no shortage of it. People absolutely have wanted to back that up

and protect it against some form of disaster or the need to recover it for years. I think the big switch now is obviously these cyber threats. And this has been something that's been with us for years, but you can't go a day, you know, now without seeing some form of ransomware event or some other cyber attack. And I think that's really switched things in a lot of CIO and enterprise mines. They're more thinking now about not just how to back up and backup quickly, but it's how do I get the data back. So it's almost like a reversal from thinking backup primary to thinking restore primary. They need this backup to really protect them

if and when there's an event to get their operations back up and running and that really is going to depend on the capabilities of your backup system and the data. Now, obviously, the elephant in the room is AI. It's really shaping how organizations rethink their data. From your perspective,

what role is AI playing in the modern data protection strategy and how are you thinking about implementing that with your customers. Yeah, we've thought about this kind of multiple ways. Obviously, AI as a new data source is there and people are going to think about how to store it, how to access it for all the different phases of data preparation, model training, model fine tuning, et cetera. But for data protection specifically, we're really tuned in on this, right? Because if you start thinking about the threat of these cyber events against AI data, things get really scary. Imagine training your AI on data,

that's been tampered with or has some kind of integrity constraints. So the need to be cyber resilient on the data that flows into AI is super critical, right? So that gets you thinking two ways. One is how can we use AI to protect the data protection system? Can we embed AI to make the system smarter to look at things like access patterns that might be suspicious or is there some kind of payload coming into the system that we shouldn't trust. So the use of AI in the storage system to really enhance data protection is one. And I think the other one is to think about more active defenses, right? Because it's not just

the good guys that are going to use AI, but the ransomware actors are also thinking, right? You might have seen some reports in the industry about how these ransomware actors are really very organized. Guess what? They're organized to the extent of using AI themselves for fishing attacks, access to personal information and impersonations. So you really kind of have to think about it both ways. But I would say for us, it's starting to think about how to be much more on top of protecting the data that we're storing using AI technology. So with that, Paul, you know, shifting gears a little bit into the tech. Skillity has built a really strong reputation in enterprise storage. But how is Flash and specifically high capacity SSD's kind of changing the

landscape for backup data protection? Yeah, that's a really good one. And I should say that we've been focused on this very high capacity storage solutions space for all these years, right? So people don't come to us and say, I have 10 terabytes or 20 terabytes to protect. They have petabytes, tens of petabytes, hundreds of petabytes, and now we have people with actual exabytes. So it's kind of a different scale. And to answer your question, for a long time to do that cost effectively, we use the highest density media we could find, which for many years, of course, was HDD, hard disk drives, right?

So our scalety system as we deploy it, we'll use traditionally HDD for most of the data storage capacity. But we've always had a flash layer in the system. And you can think about it as sort of a 1% or 2% of the overall capacity being flash. Why do we do that? Because we can accelerate a lot of operations in flash, things like metadata operations. You need to do lookups on the data. You need to list the data. You need to do update.

on small metadata tags. Flash is great for that, right? So we've had these optimizations in place for years, but now you have the shift happening to high-density flash media and especially the cost curves coming down on Flash. And I think the industry has always been hopeful

that Flash will come to parity with respect to HDD pricing on a dollar-per-gigabyte basis. We see it converging, but I would say for us now, the need to have super-high performance, low latency, high-throughput. In many cases, people want us to be able to use full, all flash deployments. So it is happening more. I would say

it's more driven by use cases that need that kind of performance. And I'm starting to see a lot of it in analytics and AI. Those are the specific areas. Thanks a lot of sense. I have a follow-in question for you, Paul, because you mentioned high-density storage. He also talked about high performance. How should we view the demand leaning and the trade-offs between high-density and high performance and how you're seeing that impact the deployments that you're delivering to customers today. Yeah, I think as the use cases start moving more toward these arenas where object storage is used

for high performance access, you're going to see more and more flash, right? I actually had a conversation this morning with one of the big analyst firms, and I actually asked them, do you see the shifting over time, the use of flash for different mixes of workloads? And they said, if you think about it,

Already AI models in the cloud are trained on object storage, and a lot of that is optic storage on flash, okay? We're seeing the same pattern for on-prem deployments. I would say that fast backup even can benefit from all flash. Why? Because again, it's back to this restore equation. If I have flash, I can do a faster job of restoring. I can give you a shorter restore window.

So I think it's going to shift and it's going to become predominantly flash, all flash. but I think we can't ignore the fact that HDD still lives. There's still a lot of deployments of it in the data center. And you'll remember 20, 30 years ago, we all said tape is dead. It's not dead. We still see people asking us for cold storage on both the cloud and on tape. But I think it's leaning more and more over time toward high-density flash.

Yeah, I was just going to ask, Cal is Scality really leveraging Flash, which you kind of answered that. So you do see a world where you might just be using fall flash. in eradicating hard drives? I do. Yeah, maybe just to give you a quick comment on that, what we see is customers have a range of needs. And one of the things that we need to address that range of needs is to be able to bring the right solution for each use case.

And that might mean all flash for some. It might mean this mixed hybrid, but mostly HDD model for others. But I really see the world moving to this hybrid model where data management and data life cycles span across media. And so data's hot for a while, that's Flash, right? You're going to need the performance of Flash to deal with hot data. Then as it cools down, I would say over time, maybe it doesn't make sense anymore.

Now you can put it on spinning disks and then over time maybe even on tape. So I think what we see as one of the values we can bring as an agnostic software vendor is to be able to do the right thing at the right time. And yes, I think it'll move more and more to all flash deployments, but there's still this sort of data life cycle that can maybe be optimized a bit by different qualities of storage media. Paul, one of the things that I've been most impressed about with you and your leadership team, sharing the vision of scalety and your impact on the market,

is your intimate knowledge of large-scale customers and what's on their minds. I know that you're going to have unique perspective on this in terms of the biggest challenges your customers are facing right now, terms of protecting and managing their data and how this is changing and how you're working to address these things as they emerge. Yeah, I think we're really focused on what's next. And obviously, you probably talk to every tech vendor in the space. They can't have a conversation without saying AI, us included. It's coming at us fast, right? And I think the thing to understand is there's different forms of AI. And we need to be innovating for all of the different sorts. So

Gen AI is all the rage today. It's going to move to a general. Gen. Take AI, which is kind of this evolution of gen AI. But you know, we've had AI for five or 10 years before with what you might like a computer vision. These are like multimodal AI. Maybe you would call them analytics applications that need to do pattern recognition, image recognition, video, medical image, pattern recognition to see if you can spot some kind of anomaly. That kind of AI we've dealt with for years. And I think we see it continuing. And I actually think it's going to be the bulk of data. But for us to innovate, we need to think about that spectrum. And really, the innovation for us comes in, how do we make it most efficient to manage that data life cycle?

Again, you have to collect data for AI. You have to process it for model training, fine tuning, inferencing, and then you archive and see that there's a whole life cycle just underneath the AI pipeline that we need to innovate in. Yeah. And, you know, AI and all the different workloads emerging, right, has really put pressure on the storage market and it's become super competitive, as you know. But I guess my question is, with the storage market becoming even more competitive and customers having so many options available, what would you say really differentiate skill at ease approach? And where would you say your biggest kind of strengths are for your customers? Yeah, that's a great question. And it's often the hardest question, right? As a tech

vendor to say how you're different. I'll go back to one thing and then I'll add a new twist. Number one, we'll stay really true to our software focus. We're not a hardware vendor. We leverage the best of breed like Solidime brings for SSD for Flash. That's our value is in being able to be agnostic, but still optimize for each type of media. If there's something special we can do for your flash drives to make it the best performing or most optimal or most energy efficient, we'll do it. But that is a value, just in being able to be sort of that layer across the different types. I would say the second part to your question is, how can we differentiate?

You're seeing the market move up now. So storage is storage. Then we have data management above it. But I think you'll see the industry and us included moving up to data platforms. And what is a platform? A platform is something that has a series of services. It might be a search service. It might be a data cleansing service.

It might be a vectorial database for AI, right? So that's what I would call the platform vision. and that's absolutely something we think. And as a software vendor, we think we have a real advantage in doing that. We're not totally tied to a specific server or specific form factor or specific technology. We can leverage it all. Paul, you know, you've described the migration from Gen. AI to Agentic AI, the huge challenges in storage.

You just gave us a fantastic definition of platform, the platformization, I guess, storage systems. When you look ahead and you see how this next wave of innovation is shaping, how do you and the rest of the Scalady leadership team position yourself to best meet these emerging means with the most competitive solutions possible? Yeah, I think I said that we'll stay true to our software-only routes. That is one. But I think we have another thing happening now, and it took a long time to get here. I think we're finally seeing object storage hit the mainstream.

And we've been a proponent and a champion and a pioneer in that for now 15 years. Right. Even before my term at Scalady, I was with another object storage vendor back before people couldn't spell object unless you were using Amazon S3. I think there's a real opportunity now because object storage is ideal to hold all of this data. It's super scalable. We all know that.

But moreover, it can provide these sort of services around it. It understands metadata. It understands the data. I think we can stay to that and really be perceived as the singular, pure software-defined object storage leader going forward. That's what we want to be seen as, and that's what we want to represent. And I think that market's growing, right? If data protection is now centering around immutable storage, object storage does that well. And you already have the hyperscalers proving that Amazon S3 and Azure blob, which are object stores, are used a lot in AI.

The opportunity is huge as an object storage player in that arena. If you're looking to do object storage, I think Scality has really done it right. So I really appreciate your clarity, Paul. This has been an awesome conversation. I think a lot of our listeners want to continue this conversation with you and would love to learn more to where they can go to either contact you or just get a download around Scality. Where should they go?

Perfect. I'll actually give you the usual answer and then a little sub-answer. Obviously, our website, scalety.com, is a great hub for all of our information. If you're interested in just data protection backup solutions, we have a micro site for that, and it's represented by our second product, Arteska. So arteska.com is that mini site that will give you much more specific information on just backup solutions. Well, I was excited about this episode, and it totally exceeded my expectation. So, Paul, thank you for the conversation.

It was such a pleasure to have you on the arena. Likewise, Allison, I really appreciate talking to you and Janice, and I enjoyed our time together. And Janice, this wraps another episode of Data Insights. You keep bringing the heat with your guests. Thank you so much. Thank you for having us, Allison, and thank you again, Paul. Thanks, Janice. Bye-bye. Thanks for joining Tech Arena.

Subscribe and engage at our website, Techorina.a.i. All content is copyright by Tech Arena. Thank you.