Transcript
We're here at Convene, a new venue. We're talking about the AI data and trust layer. One of the things about Veeam, a hallmark of Veeam, has been its partnerships. One of the partners is ExaGrid. Mark Crespi is here. He's the Director of Business Development. He's the Chief Evangelist back at ExaGrid. Good to see you again. Great to be back. Great to see you, Dave. Hi, Krista. Hi, Mark. We talked in 2017, I think it was, in New Orleans. It was a fun VeeamOn. I recall at the time, you were explaining to me ExaGrid's unique value, and you were pointing out that a lot of systems have run out of gas. And you guys had this, I'll call it a cluster approach, where you scaled not only storage, but you can scale compute to have the horsepower necessary to handle the backup, and importantly, the recovery. So I want to start there, because is that same concept, does it apply now to the AI world? So tell us about ExaGrid, what you guys are all about, and I want to sort of dig into that a little bit. Yeah, you got a great memory of 2017, yeah. So those core principles still very much apply, but there's been a lot added to it with AI and with ransomware and the threats that are out there. So we've added a whole host of comprehensive security and so on, and we're really in the business of replacing disks these days. When I started at ExaGrid, it was a tape replacement market, but we're 20 years past that now. So what we find is that customers are usually either backing up to some form of primary storage, or they may have an inline deduplication appliance. And those are really the two categories, except for us. So we're in the tiered backup storage category, a category we own to ourselves. And we took what was good about straight disk or primary storage, got rid of the cons, took what was good about inline deduplication, which is the economics, got rid of the cons, and created a best of both worlds approach with tiered backup storage, with a very comprehensive security framework for this new world. So Mark, I'd love to understand how, the topics we're hearing today at Veeam, right? Sort of the need to be able to have this resilience and recovery, but sort of at AI machine speed, right? And also just the volume of data that's being generated today. So how does that change these kind of data backup and recovery requirements? Yeah, obviously has a profound impact. And in the keynote, Anand talked a lot about the additions to Veeam around AI and threat detection and so on. In the storage side of it, we are the storage of last resort. So we have to have a bulletproof framework, which we have with our S3 support. We have a feature called retention time lock, which basically delays deletes, because we don't know whether they're operational or malicious. So a customer can delay deletes for a period of time till they know for sure it was an appropriate delete. We have immutable objects, so we can't, we won't suffer from a mass encryption. So our role in that whole paradigm is to be the storage of last resort, the storage you can count on when all else fails. So what do you make of this, the intersection we've seen it now for the last several years, the intersection between data protection, cybersecurity, obviously Veeam's doubling down on that pretty heavily. How do you guys fit into that? What are your thoughts on that market trend? Yeah, so cybersecurity or cyber resilience, somewhat of an umbrella term like so many other things, and there's many facets to it. And where we fit into it is, the data has to be there in the event of attack. So we recently added something called AI-powered auto-detect and guard. So we built in a neural network into our product. We actually monitor normal data patterns that occur during backup and recovery and restore and retention grooming and all those sort of things. And if we see a pattern based on what we've learned previously, that does not match what we've learned previously, we will alert the administrator, we alert the security officer in the organization, and then they can take appropriate action. They can indicate this was an expected behavior, and it's going to occur again. So AI, the neural network will learn from that and not alert again. They can say this was a one-time event, but it was us, so we're going to allow it, but don't learn from it. And then the third one is, this is actually an attack, and at which point everything's frozen and the customer can go into recovery mode. So Mark, there are, I'll call macro market changes. Yeah. Cloud, work from home, containers, AI. Okay, those are the sort of the big market vectors that you have to navigate. And there's also these other sort of, I'll call them micro technology changes. And one of the things that AI, you hear about AI factories, has brought in is, and ever since I've been in this business, there's been a storage hierarchy. There's been memory up top and tape down the bottom and everything else in between, and it's actually getting more and more granular. You go to NVIDIA GTC, you hear about SRAMs and high bandwidth memories and low power memories and KV caches and all this stuff. Flash is in there, spinning disk is this, Seagate's going through the roof, their stock price, that's back, tape is still alive. So the hierarchy is getting more granular. What does that mean for ExaGrid? Is all that sort of irrelevant to you guys, or do you have to sort of navigate those changes? What does it mean for your technology and ultimately for customer value? Yeah, it's a great question. We can take a simpler view of it. It's certainly not irrelevant, because in the data protection game, it's about performance. It's about how fast can I do things? How fast can I protect the data? And then how fast can I, more importantly, how fast can I get the data back? At the storage level, it's sort of simple, right? You've got hard disk drive technology. We still lead the industry in performance for that. In fact, we've won bake-offs against other SSD products with our HDD product. However, we also just introduced a whole line of SSD-only products. There are certain customers in certain scenarios where they want that level of SSD performance, they want the green benefits, the rack space power cooling, and we offer that as well. As far as the very granular layers, it doesn't come into play so much in the backup storage realm, so we can talk about it more simply, but certainly not irrelevant. Well, of course, you mentioned it. I mean, obviously, supply chain is something that you have to navigate. The price of NAND is going through the roof. I mean, six months ago, the Neo clouds were like, no, no, we only do SSD or flash. Now they're like, where can I get me some spinning disk? Yeah, well, I have a good plug for that. Because we're such a healthy company financially, and because we lived through COVID, sorry to invoke that word, but we saw what supply shortages can do to our customers and to us by extension, obviously. So we made some very strategic investments up front, bought way ahead on components and so on. So unlike many, we're able to still say to customers that our prices have not gone up significantly. We can still ship product in one to three days from the time you order it. It'll be on site and ready to go. We're hearing from our customers that with other products, the quotes are only valid for a day or two. We still kept our window for quotes, that they won't get product for seven or eight months or longer. And we've been able to navigate all of that. And we'll be able to at least through the end of this year, if not longer. That's great. I think NAND is more volatile than the price of oil. 70% in just the last couple of months. It's incredible. It is incredible. So talk a little bit about your relationship with Veeam. It goes back many, many years. How has it evolved? Where do you guys want to see it go? Yeah, so it's a longstanding relationship with Veeam. And I can remember in the early days, some of our early meetings together, and they started talking about this thing called instant restore. And it was the first time we heard the phrase. And in fact, the NAND today mentioned that they had invented it and pioneered it. We realized at that time that we were really in the backup storage realm, now what we call tiered backup storage, uniquely positioned to support that capability. Because unlike the inline deduplication appliances where everything is deduplicated on the way in and stored fully deduplicated, we have that tier we call our landing zone, which is our restored tier. And that data is not deduplicated. It's in whatever form the backup application, in this case Veeam, sent to us. We also have those copies deduplicated for economics. But when it comes to things like instant restore, we can go ahead and serve that up with very high performance. And then our scalability as well has continued to be a very important thing with Veeam. So as Veeam continues to innovate and evolve, we're always looking for opportunities for us to deeply integrate. Because unlike primary storage and secondary storage, that sort of multi-purpose use it for whatever, we are built for this purpose. And therefore we can deeply integrate with our partners and really highlight the technology they bring to bear that you can't do with regular old disk because the infrastructure is not there. Can you tell us about ExaGrid as a company? You mentioned you're profitable. For those who may not be familiar with the company, little bit about the history. Sure. I think of you as, you know, you're not one of these companies that's always just trying to, you know, ride the next wave necessarily. You ride waves, but you stick to your knitting. So tell us more about the company. Yeah, and I've been there pretty much from the beginning. So I can give you a bird's eye view, bird brain's eye view, however you want to look at it. But so like I said, we started when tape was still roaming the earth, dinosaurs and people were, you know, just itching to go to disk. But we made a couple of fundamental decisions at the beginning that other vendors didn't make. One was this whole idea of deduplication. You have to have it for economics. It's a must have. Otherwise disk is too expensive. But it can be a performance hit if you don't do it appropriately. So we avoided monolithic inline deduplication and said, we need something adaptive that favors the backup window, favors the restore time, but still does things in parallel so that customers have all the deduplicated data stored, replicated, et cetera. The other thing we got fundamentally right at the beginning was scalability. We recognized that this was a high throughput game. It wasn't a game where you could have a single point of entry into a product and that that thing was going to do all the work no matter how much capacity is strapped to it. So we came up with scale-out as a way of always adding performance with capacity. So every executive appliance in a scale-out system is a full server with network, processor, memory, disk, et cetera. So if my workload doubles, my performance profile doubles. And then more recently, we focused very early on the coming ransomware threats and security and developed a very robust comprehensive security framework that makes sure that all of our customers can recover from a ransomware event from executive. Okay, so I was going to ask you what you optimize for. Obviously, you optimize for resilience based on that last statement. Right. You optimize on getting the data off fast into your system and recovering quickly. Those are the three big optimization points? They are, and it's shown in our numbers. We're really the only vendor in this space growing double-digit at this point, year over year over year. And the other thing is we're really the only ones with a purpose-built product in this category that's still aggressively innovating. You know, primary storage innovates in its zone and so on, secondary storage. The inline deduplication products seem to have slowed down. There's just not many blockbuster new capabilities coming out, yet year after year after year, we're innovating very aggressively. We've got a great engineering team, and because we're built for this purpose, we're not a general-purpose product, we're constantly innovating in this space and have big investments and long-term roadmaps planned so our customers know we're going to keep up with the market. It's not going to be a dead-end product that doesn't innovate. And obviously, Veeam is a huge partner, but you're software agnostic, is that right? Yes, we are agnostic. We support over 25 different types, and many customers have multiple tools in their environment. They may be using Veeam as their primary, but they have a couple of specialty tools they may use for a specific purpose, so we're able to sit behind all of it, yes. So Mark, to me, it's a really clear value proposition when we talk about resilience and sort of, you know, addressing data protection requirements in this AI era. Here at VeeamON today, you know, a big headline discussion is the acquisition of security AI by Veeam and sort of, you know, how they're integrating capabilities like DSP and AI security posture management, you know, into their capabilities. What does that mean for ExaGrid as a partner? You know, are there opportunities to maybe, you know, kind of leverage your landing tier with faster performance to be able to serve some of this more sensitive data, for example? Yeah, I think, you know, it's twofold. One is because of our performance optimizations, some of the functionality data inspection things require very strong performance from the storage, and we deliver on that. But we also stay in our lane. You know, those are things that really are uniquely capabilities that should live in the software. So we don't look to duplicate or complicate things by having the storage step out of its lane and do things that really the software was meant to do. Our role is to provide the best performance, scalability, and security for the data that's being sent to us, and to not do anything that interferes with any of that functionality by, you know, putting deduplication in the middle of it, or not scaling well, or those sorts of things. So we're a very diligent partner to make sure we showcase all of these wonderful new capabilities that are coming down the road. You took some time out of the industry to do the, I think it was the Lionheart Foundation. Yes, yes. Hello, Lionheart folks. Explain what that was. Why'd you do that? What brought you back into the industry? Yeah. A little personal. Yeah, great question, yeah. You know, I have been in technology for a couple of decades, and I just wanted to see if I could try something completely new that I didn't understand. So I saw this wonderful organization, and they needed an executive director. I didn't think, honestly, that they would give me a second look, but I wrote a letter to the founder, and I was there for two years, and it's a wonderful, wonderful organization, the Lionheart Foundation. And I'm on the board now of the organization, and the work they're doing is so incredible and so important. I'm honored to have been a part of it and to remain a part of it. And it focuses on mental health, yes? Yes. Is that right? Yeah, it's for, it deals with the fact trauma impacted populations. So, you know, I think we overlook the role trauma plays in society and the dysfunction that we see in society. So it's about programs that help people to heal, recover from trauma, and lead productive lives. And I miss technology. I love both, but I, you know, when it came time, talking to Mr. Andrews, our CEO on the phone, he said, well, why don't you come back here? And I'm like, I'd love to. Well, welcome back to the arena, Mark. Thank you. Thanks so much for coming to theCUBE. Appreciate it. Yes, thank you. And thank you for watching. This is Dave Vellante for Chris D'Cascio here at VeeamON 2026 in the Big Apple, right back, right after this short break. ♪♪♪
