Transcript
watching theCUBE's continuous coverage of VeeamON 2026. I'm Dave Vellante with my co-host, Chris LaCasse. Shiva Pillay is here, he's the Senior Vice President and GM of Americas at Veeam. Welcome, mate. Good to see you. How you doing? Welcome, mate. You've got the Australian. I'm an Australian in the Americas. I love it. How'd you end up here? You said eight years you've been at Veeam. Did you start in Australia and then? No, I started in Singapore, so I lived in Singapore for 12, 13 years. I ran the Asia-Packed Japan practice for Veeam, and then I moved over to run the Americas over the last three years. So, two questions. How are things different overseas than they are in the Americas, and how has Veeam evolved over that time? Yeah, so the first question, things are different but also similar. So, you know, in the lack of being in the epicenter of where all the major commerce and transactions happen, you have a higher reliance on innovation. So I look at the LATAM markets and some of the markets I currently serve, but also like in Asia and India, you have this huge appetite for innovation. Australia, where I'm from, is kind of like a mini version of the Americas in that, you know, we adopt very fast. We move quickly with technologies like cloud, and obviously AI is the topic at the moment. So, similar but different. Your second question on what's changed, and this is my eighth VeeamON that I've been at, and this is the first time I think, you know, I've seen this through waves. We were VMware's pride of choice of how we provide resilience for that organization, and then we evolved into the cloud era. This AI explosion is truly like a change in how we're approaching for our customers. Yeah, so, okay, so 2018, you joined. Cloud, obviously, is in full gear. Yep. You know, the container opportunity presented itself. You guys bought Kasten, and then you, through COVID, you had to pivot to that trend. Yep. So, you guys are right there for remote work, and then we exit that into the AI trend. So, there's like three or four major inflection points that you've had to navigate through. How has the architecture had to, the strategy and the architecture changed to accommodate that? Yeah, and I pride VeeamON on our ability to think a little bit ahead of where the customer's going to go next. So, you talked about it. We were VMware. We went to cloud. We were able to provide that cloud technology support. Then containers became a big conversation. So, rightly so, Kasten. The intersection we're in now is, I think, a little bit different. AI represents both a huge opportunity for acceleration and solving for humanity. We're going to talk through a little bit of this shouting to get this conversation going. But what that intersection now means is AI applies horizontally across a business, whereas traditionally, things were in silos. And because of that, we see the requirement to bring the security practice and the data resilience practice together for a data and AI trust platform. So, this is profound. And I just kind of want to repeat it back because I know there was a lot of noise there. But essentially, you guys laid out in the keynote, Anand, we just had him on, is the SAS, the hundreds of SAS applications all have their own data. They have their own business process, logic, metadata, et cetera. And essentially, it's the department that kind of owns that data. And it's not harmonized. The profound change with AI and the opportunity, the promise of AI is we can dissolve all that friction within the enterprise. What you called, I think you called it tribal knowledge. So, humans are constantly interceding. AI can potentially make that all go away. Let's say in theory, and hopefully it will live up to that promise of a single version of the truth. Your job is to protect that single version of the truth, irrespective of all these industry changes. 100%. So, I put it into three categories. Some people are responsible in building AI technology that folks will use to solve the most complex problems in the world. There's a second group, which is most of us as consumers, we're here to consume and embrace AI. And if you haven't, you need to because the world's moving so fast. And the third category is where we sit as Veeam as a company. By buying security AI and putting this together, our job is to make those two journeys, embracing and driving that innovation, safe and trusted. The other profound thing I heard today in the keynote is, if you just protect the agent, you don't know necessarily the shadow agent. You know, how are you going to protect that? You have to go to the data. Explain your philosophy in that regard. Yeah, so I think shadow IT as a concept was a taboo concept. Shadow AI, I wouldn't call it a taboo concept. I would call it a signal. So my prediction is access to AI might become equivalent of a human right, like water, internet, and any of those things. So you can't take the same process of, hey, you can't have shadow IT. Everyone is using shadow AI. There's 70% of organizations that got clawed Gemini usage by employees that are using it outside of the realms of what the company holds. So the traditional method of approaching, like banning or preventing that, doesn't work. Having trust and guardrails in your data, context to who's using the data, what they're accessing, and where that data goes through a security, governance, privacy, and resilient lens is the difference, I think, in a world that traditionally was solved in a quite antiquated way. Yeah, it's about enabling, right? As you say, instead of kind of inhibiting. So, but I'd love to talk to you, Sheva, a little bit about sort of where do you think customers are at today? I know obviously there's this appetite to adopt AI, but kind of where do you see customers at from that perspective, in terms of their adoption of AI? And really, what's the crux of what they're struggling with? I know we've talked a little bit about kind of, you know, this governance layer, and how Veeam is kind of bringing this context and these capabilities together across these multiple areas, but I'd love your perspective on that as well. Yeah, and I put it into two lenses, and I sat with a group of CIOs, CISOs, CDOs, CTOs, and CEOs over dinner last week in New York. I was here for a business occasion. There's two things. There's FOMO, so the fear of missing out. It's like, I'm not invited to the dance party, and I don't have the backstage pass, which is wildly intoxicating when you think about, at a board level, board is putting pressure on, get me AI. They're not saying what they want. They're just saying, get me AI access, and make sure I have the latest capabilities to touch my customers. So that's one lens. The other lens, and this interesting survey from BCG came out, which is 65% of customers feel like they're moving way too fast with their AI initiatives. And that's probably from the operator lens of like, I have no control, no visibility. I can't see the repercussions of what's happening. And we're fast going to see that manifest over time. We're not seeing it just yet. I think the downside of AI and the upside are yet to be seen. Now, what that means is this transcends the traditional egos and personalities of CISOs, CDOs, CTOs, chief AI officers. And all of them want to own or don't want to own AI, depending on where you go. Some of them will be like, I'd rather that be someone else's department, and some of them will be like, I'd rather this be my department. Because it's horizontal, that's just not going to work either way. And for us, that's why the data command graph just gives you this unified lens for any role type to look at the data in a trusted manner. Which is important for a company like Veeam. I know you've always been very close to your customers. There's a language of we want to win their hearts and minds like Veeam did back in the day of the VM admin, right? So can you talk a little bit to that and how the command graph can provide that single kind of maybe pain of truth almost for these different personas? Pain of truth is a great one. I might use that one a little later tonight. But think of it as a social network for your data and context. And I should be saying data instead of data because I now am implying an American accent. But when you look at it with the social network or social context, you're now able to see toxic combinations. For instance, my HR information shouldn't be sitting in a shared repository in India accessed by a HR agent and exposed. And that shifts a very different lens to, before it used to be about a threat actor, used to be about someone doing something malicious. That's changed significantly now. Agents aren't trying to do malicious things, they're just working. And so without that visibility and context, you don't have the ability to triage. So I want to, context, the word context is like data management, right? And so I want to sort of put a finer point on it. My interpretation of when you talk about context, you talked last night about it, and we've heard a lot today, is you're talking about the surrounding operational and maybe even technical metadata associated with the data as opposed to what I think about like Palantir ontology. Right. Fair enough? However, at some point, somebody is going to develop a horizontal layer for that type of ontological context, and it's going to capture the state of the enterprise. Do you see it as your job to actually protect that state? It's not here today, but it's coming with these AI factories and this new layer that's coming, absorbing sort of the general purpose computing and all the functions that we do there. Do you see a day where you actually, at a granular level, snapshot the state of an enterprise? Yeah, I won't get ahead of my skis with the product innovation roadmap, but I think you're on to the direction we're heading. We want to give customers the ability to one, see what's happening in their organization, outside of the AI lens, what is agentic systems touching? What is it in context to your security requirements, your compliance requirements? So we'd give you the ability to detect, protect, and then the undo piece is, to me, where this changes, and this is where, why we're so confident in saying this is a new category, because never before have you been able to contextualize your data and then act on that data, leveraging the resilience piece that we've had a long heritage doing. And that category is data and AI trust? Correct. Is that right? Yeah. And it comprises all the pieces, governance, privacy, security, identity is in there as well. Help us understand, we were talking to Krista, you're not competing with an Okta, you're leveraging that good work. Explain what identity means to you guys. Yeah, and I think identity means for us is understanding who has access to what and what they should have access to. The systems that protect identities, Okta and the rest, they still have their own purpose in what they're doing, but I don't think anybody's stitching together the concept of, hey, an agent or a human should have access to this shared repository or this file, or similarly, credit card data shouldn't be sitting in this part of the organization. And being able to contextualize that, I haven't seen, I've been in the space for a while, I haven't seen many technologies that can do that, visualize it and give you both visibility and remediation. And that's the key, is your agents will actually remediate based on policy and based on governance and guardrails that you guys establish with your customers, right? Yeah, and I've sat with some of the customers, both from Scrutia and the Veeam Lens, and I think, to me, what's interesting is you could always see it, and then you'd have to go off and find another way to remediate it. Now it's surfacing, hey, this sensitive data should not be existing here, would you like to undo or would you like to remediate? That's super interesting for our customers and makes the work of every person as we evolve to this rapidly changing landscape a little bit easier. So, we love Stackslides, because we're analysts, and Anand showed a Stackslide today, compute and infrastructure, then data analytics, which is like the snowflake, Databricks, BigQuery, and then a missing piece, and then models, LLMs, critical, obviously, agent orchestration, control framework, and then applications and vertical SaaS. And then he sort of did the big reveal, that missing piece is data and AI trust. Okay, I would say, I would come back to the question I was asking about sort of ontologies. I actually see, and I want your reaction to this, is what you're doing as the harness and the scaffolding around that system of intelligence layer, which is not there yet, it's emerging, it's going to be a very high value piece of real estate, and you guys are the harness around that, and that is a critical piece in order to be able to feed agents in a trustworthy manner. Do you buy that premise? I do, I think, yeah, so to your point, we're allowing or going to be allowing your ability to unleash your AI usage to solve real problems. Now, the debate on whether AI usage is solving real problems is a whole different one for businesses, and that's for boards and folks to decide. But if you were to, like, hey, go and solve complex problems that affect customers, sensitive data, and critical outcomes, like I've seen some pretty bad AI outcomes for customers, like order refunding, but some things that can lead to dangerous outcomes, not only money loss, but also health and humanity. So, for us, is give you that trust in that layer so that when you apply your AI models and your foundational systems to it, you're in a position that can either, like, bad things will happen, they always do, they've happened since the advent of IT, but you're in a position to rapidly solve or undo those problems very quickly. When you think about the bell curve of adoption, you guys announced this maturity model. It's a freebie, by the way, so take advantage of it. You mentioned before, a lot of organizations, and it's funny, I've talked to chief AI officers, and my job is to eliminate this role. That should be embedded across the organization, your point about horizontal. For those ones that are on the leading edge, I think of companies like JPMorgan Chase that I've talked to, Walmart, Lilly, they've got the resources to hire the equivalent of a forward-deployed AI engineer. 100%. Right, and so, two questions. What are you hearing from them, and how do you see solutions occurring so the rest of the mainstream organizations can actually take advantage of AI? Because that's the point at which everything you guys are announcing here is just going to get adopted very quickly. Yeah, so I can step back and say AI doesn't differentiate based on your size, whereas previously, systems would differentiate based on size. So if you're an S&P customer, a mid-market customer, and you're deploying AI, the catastrophic outcome, it doesn't differentiate and say, oh, I'm going to take it a little easy on the mid-market customer, because AI is an entity, it ingests information, period. So, you're right, larger enterprises have been a little bit ahead of the curve, they've been thinking about this, but I also think the landscape's changing every day. So, no matter how far ahead you think you are, you wake up this morning, we saw that Anand mentioned there were some new announcements from Chad GPT, next week there's Anthropic, it's changing at a pace. So it goes back to that foundational trust layer. I can be comfortable that things are changing at rapid pace if my foundational trust layer exists, my data and AI trust layer, because it doesn't matter what comes next, I'm able to see, be able to roll back, and undo and have that conversation. But your point about the data maturity model, we had such good success with the resilience version of this. Being able to pivot it now into more dimensions for AI, here's a great story, I've sat with a customer, mid-market customer, I'd say mid to low enterprise, and they sat there and said, we have this all under control, and we brought all their groups together, and through this facilitation process, we worked out that, hey, these departments aren't talking to each other about the same, and it's always a pleasure for us to help solve for our customers. It's not a cost exercise, it's not a consulting piece for us, it's a path for us genuinely wanting to give customers a roadmap to say, okay, I'm here, and I want to get to here, and this is what it looks like, and we'll take you through that journey, but it's not necessarily just Veeam's technology that does that. Well, and it is a new buy model for customers, right? They need, to your point, it's not pure consulting, but they need a little bit more hand-holding and a little bit more advice up front, and I know, with the resilience model, I know Veeam found, and we've seen this as well, that most customers almost had these rose-colored glasses, so they thought they were further along than they are, and I'm sure, are you thinking that customers are going to be finding the same in kind of this AI era, or do you think maybe their eyes are a little bit more open to the fact that there is this shadow AI problem, and they probably don't have as much for governance and security in place than they would hope? It's a great question, I'll answer it very honestly. I think prior periods, when you looked at resilience, a huge percentage of our customers and prospects, and anyone we spoke to, overrated where they were. They had a high opinion of where they are, and it's not out of lack of understanding, it's just because ransomware and all these things were changing at such a great pace. I call ransomware Catalyst One, where the CIO and CISO had to meet because backups became a security-grade asset. Catalyst number two is AI. I would probably say customers aren't that bullish to say they've got it under control, purely because I've never seen something so dramatically different in terms of how it impacts systems. Anyone would be foolish to say, hey, I've got this all under control. Well, and it's interesting when you talk to certain startups that call them AI native or AI first, they've got this shared MCP server environment, they've got the tribal knowledge that's a shared knowledge environment. They're doing things with one-tenth the resources that we're used to doing with human labor, and it gives you a glimpse of what's possible, but of course, they're doing it with Greenfield. It's a blank sheet of paper, and so you know that disruption is coming, and I think that a lot of executives are fearful of that, but change management, as you know, is very, very difficult. Yeah, I view it two ways. It's super exciting, and I've personally started, I felt like leaning into watching what these companies are able to accomplish is great, but also I have a pretty good earmark on 20 years of solving customer problems and being number one in the space of 150,000 customers. There's a little bit more rigor than just driving an AI native solution. It's understanding your customers, being able to respond to those customers, and because we're in this early wave of AI native tools, we haven't seen what good looks like yet. Conversely, we haven't seen what bad looks like yet, and so for me, that's why Anand's message and the team's message around being the data and AI trust layer is foundational because either way, if it's going to be exponentially great or it's going to be downside, we're here to protect and solve our customer problems. So category creation is kind of, I think it's somewhat new for you guys. I mean, you've dabbled in other niches, but I remember at one point, Veeam was saying it's all about the backup. Yeah. When everybody else was chasing. I was around for that. Right, everybody else was chasing this fuzzy data management thing. You said, look, we're going to solve the real problem. So category creation is new. How do you think about creating that category? What kinds of things are you doing to help educate the market? Yeah, and I would argue, and back to your comment about us saying it was just backup, I would argue we made one or two mistakes in doing that. One was, and I went back and saw this real life, we didn't realize in support calls and we were triaging, we were solving ransomware problems. Just because our technology was so solid, we didn't even realize we were solving ransomware problems for our customers, and so shame on us for not calling that out early. Category creation is probably the most exciting thing. Right about the time, right now, the world is full of innovation. For us to step up boldly and say this through the acquisition, so it's not a, hey, we're just creating some noise. We've made an acquisition. Those two technologies are already starting to talk together, so intelligent res ops is where this is going to land first, and there's a reason why we picked that to land first. M365 is the estate with the most unstructured and that's where you'll find your raw data to begin with. So we're hoping to show our customers, and I would use not hoping, we're very certain we'll show our customers immediate value, and then with 13.1, a path to realizing the data command graph, and then I love the fact that a customer can come in any way they like. They can come in from the security angle, they can come in from the resilience angle. What they'll ultimately find over time is, hey, the best bet is if I have both of these disciplines sitting in one place. All right, we're getting the hook, Shiva. Thanks so much for coming on theCUBE. Really appreciate it. Great conversations. All right, this is Dave Vellante for Crystal Case. We're here at VeeamON 2026 in New York City. Be right back, right after this short break. ♪♪♪
