RSAC 2026 Conference | Keynote Analysis

Clips
More from RSAC 2026 Conference

Dave Vellante

Co-Founder & Co-CEO

SiliconANGLE Media, Inc.

Erik Bradley

Chief Strategist, Director of Research

Enterprise Technology Research

play_circle_outline ETR‑TSYS Survey of ~1,500–1,800 IT Buyers: Post‑COVID Budgets Peaked, Now Volatile as Spending Growth Falls to ~3.5%

play_circle_outline Security spend resilient; high-tier (> $10–20M) security budgets increasing.

play_circle_outline LLM and GenAI Protection Tops Budgets as Identity, Data Security Rise Amid Prompt Injection and Leakage Concerns

play_circle_outline Agentic AI Adoption Surges, Governance Lags: Only 3% Have Broad Agent-Specific Security Controls; 20% Have None

play_circle_outline Control plane for agentic AI unclear; likely multi-vendor blend, depth-of-defense required.

play_circle_outline Backup and recovery vendors (Rubrik, Cohesity, Veeam, Commvault) shifting toward cyber resiliency.

Info
Transcript

Keynote Analysis

Dave Vellante

Co-Founder & Co-CEO SiliconANGLE Media, Inc.

HOST

Erik Bradley

Chief Strategist, Director of Research Enterprise Technology Research

In this keynote analysis from RSAC 2026, theCUBE’s Dave Vellante sits down with Erik Bradley from Enterprise Technology Research (ETR) to unpack the latest technology spending data and its direct impact on the cybersecurity landscape. The conversation explores how security budgets remain highly resilient despite broader macro IT spending slowdowns, reflecting a decisive shift toward system-level resilience. Vellante and Bradley dive into the realities of vendor consolidation and platformization, highlighting how industry leaders like CrowdStrike, Palo Alto Ne... Read more

explore Keep Exploring

What is the TSYS/ETR quarterly technology spending intentions survey (who it surveys and how it differs from other research), and what do its results show about current macro IT spending trends? add

What are the current trends in security spending, and how does it compare to overall software spending? add

Which security areas do you anticipate will see budget increases or decreases over the next 12 months — specifically LLM/GenAI protection versus cloud security (and other areas like identity and data security)? add

What do your survey results show about how many organizations have deployed AI agents and whether they have appropriate agent-specific security controls and governance? add

Who will own and control the control plane for agentic AI, and what is the likely architecture? add

How are backup and recovery solutions (e.g., Rubrik, Cohesity, Veeam, Commvault, NetApp, Dell) evolving into cyber-resiliency/security offerings, and why is that shift important given AI-related data-leakage concerns? add

bolt Powered by CUBE AI

Keynote Analysis

search

Dave Vellante

>> Hi everybody. Welcome back to day three at RSAC 2026. We're here at Moscone West on the ground floor. My name is Dave Vellante. Kristoff Bertrand is also here as is John Oltsik. He's got his report in his notebook. He runs around all day getting all the insights and we are going to give you a data dump. My good friend Erik Bradley is here from Enterprise Technology Research. And Erik, you've been out talking to customers, talking to vendors, walking the floor, getting the ingestion. But prior to RSAC, what used to be RSA, every year you guys do the drill down on security, state of security. This is I think the third year in a row now.

Erik Bradley

>> It's our third straight year doing the annual state of security. Yes.

Dave Vellante

>> And then you've been kind enough to come on here and share some of that data with us. But I want to start with the macro. So let's explain what you guys do every quarter with the TSYS survey. You survey 15, 16, 17, 1800 IT decision makers. And you are on the pulse of spending and have been for over a decade. So, take us through that. And Brendan, if you bring up the first slide please.

Erik Bradley

>> Yeah, sure. First of our ETR has been around 15 years. We have a long history of data and analysis that can go along with it. But what we're looking at right here, if I believe I know the slide we're taking a look at, yeah, this is basically our quarterly technology spending intention survey. What differentiates us from a market research firm is everything we do, as you know, Dave, is based by the buyers themselves. These are the people making the decisions, spending the money. So what we're looking at here is overall macro budgets. As you could see coming out of COVID in 2021, we were at a peak. Then we had that inflationary rate period where we thought the economy was crashing. Never quite happened, hit recession. We came close. And then, now we've been kind of kicking a can down around bouncing around. Now, we did see a very optimistic amount of spend up above 4% in January, but unfortunately geopolitical again has now taken us back down to 3.5%, a huge drop. Even more concerning for everyone out there is it's the world's largest organizations that are seeing even lower. So if I take that to global 2000, we're talking about 2% growth now this year.

Dave Vellante

>> It's unbelievable if this is still up. So coming out of COVID, I mean, it was like boom times, seven plus percent and it hung there for a while. Almost inversely proportional to the two year. So as interest rates rose, spending declined, and then it's just bouncing around based on headlines now. And so, it's like every quarter we get encouraged or discouraged. It was tariffs.

Erik Bradley

>> If you look at the chart there, you could see even the prior year. We were coming out of the election cycle. I think organizations were ready to spend. They knew the plan. And then, we get hit with the tariffs. So yeah, it's been a really tough environment for practitioners and to be quite honest for the vendors.

Dave Vellante

>> It messes up our predictions, too, every year because you and I, we pin on these numbers because we say, okay, let's use the data to make our predictions. And it just becomes kind of unpredictable when you're-

Erik Bradley

>> I don't know. I think we still do all right with the predictions.

Dave Vellante

>> Oh, the predictions, we kill it. We do kill it. Okay. So that's good. That's a good sort of framing, Erik, for what we're going to get into in terms of just what you're seeing in security. What are you seeing out there? Where do you want to go next?

Erik Bradley

>> So although the backdrop right now is declining spend, not the spend growth we were hoping for, security spend itself is much, much more resilient. What we're seeing right now is the middle tier is staying the same, but the really high tier of security spending, above 10, 20 million, it's really increasing. That's the bucket that increased the most. We're seeing only 5% of all people saying they plan on decreasing security spending at all. Absolutely minimis, hardly noticeable. So right now, if we take a look at all software areas, which we track by subcategory, software spending is at about I think 7.5% right now, if that's the chart shows, I believe. And if we take that and compare it to like DevOps for instance, our HCM is at 2%.

Dave Vellante

>> Okay. So actually, Brendan, before you bring that up. So basically what I'm hearing from you is that security, as has pretty much been the case in software has always been somewhat insulated from the downturns.

Erik Bradley

>> Yes.

Dave Vellante

>> Security does a little bit better and sometimes a lot better than the broader software base. And now, security has been taken down because everybody puts it in a SaaS bucket. You've seen CrowdStrike get hit, Zscaler, Okta kind of bounces around, but generally speaking, it's Salesforce, it's ServiceNow. It's, oh, you're software? You got annual recurring revenue? You have great marginal economics at volume? Sell." Which makes no sense. Now, if you bring this slide up, Brendan, so this emphasizes that point, Erik, that you just made. Is this security significantly, meaningfully higher in terms of the growth rate than these other ones?

Erik Bradley

>> Very much so. In the valuation side, as you know, Dave, the rest of the people might not, but I was on the equity Wall Street side for many, many years. The valuation concerns there were, hey, listen, software did have high valuations coming in and it's not just a disruption that Claude announces every day, but it's also the fact of, you have to look at three, four, five years out. And are people going to take that three, five year contract in this tumultuous time? So I think there's also some concern about much longer term range on the valuation side. And that's why we've seen all equities get hit on the software side. I agree with you. There is less impact in that security software bucket than there should be in the core ERP modules of HCM and ITSM and CRM.

Dave Vellante

>> But so, it's a multiples thing. And when somebody looks at software and does a discounted cash flow, they're discounting the effects of whether it's Anthropic or yesterday Databricks came in.

Erik Bradley

>> Exactly. Or whoever's coming next.

Dave Vellante

>> They don't believe, and of course we were GTC last week and Jensen saying the market's got it wrong. These software companies are going to inject AI into their business and bring agentic and agents into their platforms and they're going to be driving new incremental revenue, new value streams. We'll see. I mean, my feeling is... I'd love to get your thoughts on this, to the extent that software companies actually do that and invest and make those R&D investments. Like for instance, I do see Salesforce doing that. I see ServiceNow doing that. And obviously, a number of other SaaS companies. Why wouldn't they? Yes, we know the seat pricing is under fire, but there are new potential revenue models that are going to generate tokens and generate value for customers. Your thoughts on that.

Erik Bradley

>> Yeah, I could riff on this for a while, but I will say a couple of things. We're not necessarily sure that seat pricing will go down. It's probably the number of licenses that will go down. And I think it'll be offset by higher pricing actually because they're more value, but it'll be less seats on that side. The other thing is we actually have studied this at a great detail. We recently did AI SaaS disruption study that we could dig into really quickly, and it again went through all different areas. So security specific, what we're seeing is any kind of vendor replacement or swap is actually due to other vendor competition, not AI. It's actually the lowest area of actual replacement happening from AI. However, when we get into analytics and BI, super high. There we got 35%, 36% is saying, "No, we're replacing our vendor with AI." And it's very, very different. We're not seeing it in security. It's again, the most resilient in that study.

Dave Vellante

>> Well, you might have seen George Kurtz's little tweet on this. He basically said to AI, "Build me a CrowdStrike clone."

Erik Bradley

>> Right, exactly.

Dave Vellante

>> And I did the same thing, when IBM got hit, I said with the Z, I said, "Build me a Z clone." And Grock came back and said, "Ha ha @DVellante," and then explained why that's impossible.

Erik Bradley

>> It does show the headline world we're living in, right? The IBM one was great. All of a sudden it comes out and says, "Hey, Claude can read Cobalt." So IBM goes down 10%, but this is a very headline driven valuation world we're in right now.

Dave Vellante

>> Well, and a lot of it I guess is automated, but I mean, when you look at the market these days with the headlines, it's like, "Oh, we're going to end the war," and the market goes up. And, "Oh, we're going to keep escalating the war," and the market goes down."

Erik Bradley

>> Not an easy time to be an asset manager right now.

Dave Vellante

>> Right. They're just playing the headline for now.

Erik Bradley

>> They're exhausted.

Dave Vellante

>> You want to get into consolidation and platformization?

Erik Bradley

>> Yeah, let's do it.

Dave Vellante

>> So let me set it up. I mean, I believe Nikesh Arora kind of coined the term platformization. The whole premise was instead of buying all these bespokes tools, the average customer has 60, 70, 80 tools in their stack. We want to consolidate those and ultimately platformize them, two different things. Consolidation, reducing the number of vendors is one thing, but it's sort of a sign toward platformization. You've done research now for the past three years. Brendan, if you bring up the next slide, when we first looked at this, we said the marketing narrative doesn't match the practitioner reality. And I remember I talked to John Oltsik about this a couple years ago and he says, "Yeah, duh Dave, no kidding."

Erik Bradley

>> When does it ever?

Dave Vellante

>> You said the same thing. But look at this data. I mean, we're seeing a different story now, not necessarily in terms of the red decreasing the number of vendors in the stack. That's up a little bit, but the other two staying the same and increasing is dropping, staying the same is growing. So there's light at the end of the tunnel here, isn't there?

Erik Bradley

>> Yeah. It's a two part story. When we first started tracking this together, by the way, Dave, it wasn't just me, it was also Dave, but we weren't seeing any signs. So marketing as always was leading the reality, but now we take a look at the fact that vendor expansion, the amount of tools you're adding has stopped. I mean, that has dropped significantly, as you can see from the two years ago until now, back to 35%. Meanwhile, the amount that are staying the same, stable. So people have the vendors they think they need. And now if we take a look at the second part of this, which isn't shown on that slide, when we ask people why they're adding vendors, the best of breed rationale has hit a two year low. So when we first started this, people were saying, "I have to add new vendors because I want best of breed." That stopped. It's the lowest it's been in three straight years. So when you combine those two, they're not expanding vendors, they're staying the same, and no longer best of breed is the biggest driver, and we're finally seeing signals that platformization or expansion of vendors in your stack is likely happening.

Dave Vellante

>> And at the macro level, there's sort of a parallel trend. I don't have data on this, but you take Wiz, for example, they were the hot, shiny new toy a couple years ago. The party that Wiz would throw, the line was out the door and you could see it in that. The whole posture management thing, people started just chatting up posture management, it became the hottest thing. Then you see the consolidation, $38 billion consolidation folded inside of Google. The joke, by the way, is G Wiz, which is pretty good.

Erik Bradley

>> I like that.

Dave Vellante

>> It's very good. I heard that the other day, but the point being, you're seeing certainly industry consolidation. At the same time, you're seeing a lot of new funding going in to security, but to your point, they're not chasing that sort of gap filler like they used to, and your data suggests that.

Erik Bradley

>> Yeah. I think that more so now than ever with another yet new frontier, and you and I were talking earlier, it doesn't matter if it was computer or then IP and then internet and then cloud security. And right now with the emergence of AI and agentic AI, we're changing the threat landscape yet again. And now more than ever, there's not one shiny object that's going to work. You need depth of defense, you need hygiene, we need to get back to the basics. Everyone's going to plant a flag in the ground and say, "Nope, I'm the right one," but it's not true. The truth is you need real security hygiene, you need depth of defense. So, yeah, cloud security was the big thing two years ago as rightfully it should have been and now maybe it's sort of shifting again. And again, our data will show that, that I don't know if we're ready for that or not, but-

Dave Vellante

>> But before we go there, I think I just want to put a finer point on that. I mean, to me, the two biggest consolidators and platformizers are Palo Alto and CrowdStrike, and they've got pretty different philosophies. They're both acquisitive. CrowdStrike tends to make smaller acquisitions than say CyberArk with Palo Alto, but Palo Alto had Anand Oswal on yesterday and I asked them, I said, "Look, the criticism of Palo is that they're doing what Cisco used to do. They're just buying a bunch of companies for growth." And I said, "How do you respond to that?" And he said, "Look, we're an engineering company. We're a product company. We do deep integration. We're not expecting our clients to just... We're not bolting it on and expecting our clients to have to traverse through a veneer. We are doing the deep integration." So okay, I'll give them credit for that. I think Palo Alto is a deep engineering organization. CrowdStrike has a different philosophy. Yeah.

Erik Bradley

>> They buy them early.

Dave Vellante

>> They buy them early. They integrate them in. So there's some similarities, yet differences, but they are, to me, the two sort of gold standards of that consolidation trend. And I think continue to make the R&D investments necessary. CrowdStrike obviously is founder led. Nir Zuk has sort of, I guess, stepped aside now, retired, hard to believe that Nir is going to retire. So that's something that we watch, but I think Nikesh is driven. So there's two very interesting companies in a sea of organizations, technology companies.

Erik Bradley

>> Completely agree with you on those two leading the platformization. I think you do have to throw Microsoft's hat in the ring, although they don't get everything end to end, but they do a lot. They really do. And with their adding agentic into their security on the now E7 license, it's just going to continue. So there are plenty of people where that's good enough. And I would also add Cloudflare, what they can do with the edge, and I think the edge is going to become more and more important as we're doing agentic computing right where the customers are. I think they're another one to watch, too. But for right now, if I had to pick 200% agree with you, it's CrowdStrike and Palo.

Dave Vellante

>> But I'm glad you brought up Microsoft because you have other data. We're not going to show it.

Erik Bradley

>> We've got so much data.

Dave Vellante

>> You are the Data Kings. There was a question in there that you guys have asked frequently, and I would look for who's the leader. If you had to do it all over again, which vendor would you choose to standardize on? And Microsoft jumped way, way up, which is quite remarkable.

Erik Bradley

>> Yes.

Dave Vellante

>> And it says a couple of things. I mean, CrowdStrike was obviously pretty prominent.

Erik Bradley

>> And Palo.

Dave Vellante

>> Palo was there as well, but Microsoft jumped out at me. And so, that suggests to me... But what's interesting there is the hyperscalers will do well within their environments. AWS is not really trying to compete with the endpoint and XDR.

Erik Bradley

>> They don't have to. They don't have to. They're no matter what.

Dave Vellante

>> But Google is getting increasingly aggressive there with a couple acquisitions, Mandiant for threat intelligence and Wiz and so forth. Microsoft has always essentially competed in that space and one could criticize the degree to which they were effective, but they're really starting to improve, which is, I guess, always the case with Microsoft. And because they're so ubiquitous, I think they've got a better foothold than say, even a Google or an Amazon.

Erik Bradley

>> Yes.

Dave Vellante

>> And so, they are right up there and I'm glad you brought that out. That was really interesting data.

Erik Bradley

>> In our data across all the different surveys we do, all the different data silos, whether it's at the product level, the company level, or these annual ones we do, Microsoft is always at the top. They are always there. They're in the conversation. Whether it's a rebuild portfolio, or even this is the surprising part, if we ask people who they think the most innovative vendors are, Microsoft is still up there. And that's interesting to you and I who have been doing this for so long, because we're used to Microsoft just saying, "I'm going to throw the product out there even if it's not good, and then we're going to make it better."

Dave Vellante

>> And the joke used to be Patch Tuesday means Hack Wednesday. And so, now we're moving it . And so, they've got to... Right, but they've got to deliver for the clients. All right. Let's bring up the next slide, which is, do you anticipate your budget to increase or decrease in the next 12 months for any of the following specific areas? Take us through this and look what's happening with AI.

Erik Bradley

>> Yeah. So this is not going to surprise everyone. We know the buzz, but this is one of those instances where the data is actually leading or matching the buzz. So this is the first time that we've seen LLM and GenAI protection be the number one player. Cloud security had been the number one player for the last two years. And not only do we see a massive almost 10 point jump in LLM GenAI, we're seeing it in tandem with an almost 5% decline in cloud. So I know you had a theory and maybe we would have liked to see stronger numbers to prove this out, but all the other areas are rather even. They're stagnant. They're flat. So the only area that's decreasing is cloud. So we are seeing some evidence of a shift from cloud security budget going directly into LLM GenAI budget. Now this is not tracking agentic. We can get into that in a minute. This is just the first layer around your LLM near GenAI. And we're seeing that it's also not only that, it's the fastest mover in all the rankings, not shown here, but we have another one where it's not the actual dollars, but it's priority. So what's your number one priority? In that one, identity security became number one. Data security became number two, but AI, LLM was number three and the fastest rising. It rose 15 percentage points in the last two years.

Dave Vellante

>> So to the discussion we had earlier, I was saying, when the cloud was sort of at its peak, it was the first line of defense. And now, you were sharing with me that CrowdStrike put a stake in the ground saying that endpoints are going to be the first line of defense. We have Jeetu Patel coming on later on, and I'm sure he's going to say networking and security are coming together. And that's a first line of defense. Guys like Okta are going to say that identity is the first line of defense. Others said to me, somebody said to me yesterday, "Look, we had a practitioner on it." It's like the TSA. When you go to the airport, you have to go through multiple layers of security because one can miss it and that's the way the world is. But somebody else said to me, well, I think John Oltsik said agentic agents are going to be the first line of defense. So you've got the layers, you've got multiple first lines of defense, and this is what CISOs have to balance.

Erik Bradley

>> Yeah. And if we're just talking about the bigger picture while we're here... I'm going to give credit to Ben Horowitz. Yesterday I was at an event where both him and Ali from Databricks were speaking and he said something that was startling to me and it was the real time to exploitation has now gone from two and a half years to two and a half days and that's exponential. So that's going to continue and continue. So the only way to combat that is exactly what you said, it's the agents. The agents are going to be doing the infiltration and the agents are going to be trying to stop it. It's going to be a chess match between two agents. Now, before we get there, you still need hygiene. You still need all those layers in the depth of defense, which I completely agree on. Yes, CrowdStrike came out yesterday and said, "I already have the endpoint so I can give you the visibility. That's where it's going to be." I'm sure the networking guys are going to say, "Nope, you're still actually using the network to move around and get your workloads done, so that's going to be us." I firmly believe that truthfully, you can't follow anything until you tag it. So I do think that identity layer is first. We have to put a non-human identity on this so we can know where they are, what they're doing. There has to be governance around it. There has to be compliance around it. Once we have that, then yes, we can do all the other things to monitor them. Now what about the observability players? Can they keep up? Is this Second Life or Splunk maybe? They could still see everything, or is it too far behind the scenes? What CrowdStrike was saying was, "We're going to see it first. The logging and monitoring guys, they'll be useful later, but we're the ones who can see it first, send that out so it can get remediated by the right place." Every year it's interesting, but this one's really interesting to watch.

Dave Vellante

>> So of course, I'm coming off of GTC, which was a massive Kool-Aid injection, but with real substance, deep, deep technical innovations and quite impressive. At the same time, I made the observation yesterday. I feel like the technology industry is in the third inning. And when you're here at RSA, I feel like the customers are in the first inning. I hosted a panel at MWC, Mobile World Congress, like, 23 business technology executives and really sort of tried to probe on where they are in their AI journey. And they're deploying single agents. They're being ultra, ultra careful. They're rightly paranoid about what could happen. So I say the big gap there between agentic deployment and the ability to govern and secure those agents. So why don't you share some of the data that you-

Erik Bradley

>> Yeah, massive gap. And you're right. We come to these type of events, the talking heads like myself and then the vendors out there trying to sell what they have. And we think of security in these pillars, right? We're talking identity, networking, this. CISOs don't think like that. CISOs have a business to secure, right? They're worried about how do I get the information to my knowledge workers? How do I make sure we stay out of the headlines? They don't think of it in these pillars. They have real problems to deal with. So I think there's always that disconnect. But to your point, we have thousands of CISOs and security executives in our community. They are incredibly cautious. It's happening. They are deploying it and they're actually deploying it at a rate that's higher than they're securing. So right now we've seen in this last study, 37% of organizations already have AI agents deployed or in active testing at this moment. That's up 10 percentage points just from last year, just from... And we did last years in April because they changed it. So in 11 months, that's already up 10%. So almost 40% are already out there. However, when we talk about security controls, 20% admit they have no agent specific security controls at all. Only 3% of them, and we're talking 500 CISOs in the survey. Only 3% said, "Yeah, yeah, we have broad controls in place." So they're 40% deployed and nobody has any controls in place. And then, the survey gets better. When we ask them to unify on what they think the right governance model would be for agentic, the data's useless. Every single answer option's 23, 24%. There's no clear view. They don't have a view. So we are really way ahead in deployment than we should be from a governance perspective and control perspective. It's going to need to catch up fast.

Dave Vellante

>> I had this conversation with John Oltsik yesterday at the end of the day in The Cube and I said, "Well, what's the answer?" And he came back to exactly what you just said. He said, "Dave, it's governance. You have to start with governance."

Erik Bradley

>> You have to start there.

Dave Vellante

>> Okay, but that's cultural, that's controls, there are edicts, there's education that has to take place. That's the starting point. And companies are moving so fast, it's not like you could just kind of bolt on governance. It really is cultural and deeply... It has to be deeply embedded in the organization.

Erik Bradley

>> And then, the other part that the vendors don't want to talk about too though is the fact that not everyone's buying this. Many people are building it. And as you know, we have our own separate AI series. Twice a year we do the build versus buy. So where are they spending? What's the budget? What's the stack? Where's that happening? How do you secure it? And what we're seeing right now is due to compliance and regulations and just data sensitivity, a lot of people are choosing to build this stuff themselves. And then there, you have a little bit more ability to govern that and wrap security around it. But who are the winners in there? When we're looking at the names, and this is coming out on Friday by the way, I'd love to show you a glimpse of it, but we're looking at the same names there again. We're looking at Microsoft to enter ID. We're looking at Okta, SalePoint, we're looking at CyberArk. But CyberArk, it's interesting to me that acquisition that Palo made. They're known for privileged access, right? When you think about the thousands and millions of agents that are going to be out there, and some of those agents are going to have 10 of their own agents doing work, how are we going to figure out privileged access for this many agents at that scale? I think we're so early in this and the vendors are going to keep pushing that they have the answer. They don't have the answer. They don't have the answer.

Dave Vellante

>> So that gives me to the other point that we want to talk about, and it's a good point to close on, who owns the control plane for agentic AI? We were talking earlier, we sort of touched on it. CrowdStrike put their sort of endpoint stake in the ground. We talked about Palo and the platformization play.

Erik Bradley

>> Databricks.

Dave Vellante

>> Databricks now is disrupting this business. You've got the hyperscalers, which was the first line of defense. ServiceNow has been playing in this. McDermott's going to say, "Wait a second, what about me?"

Erik Bradley

>> Yeah.

Dave Vellante

>> So what is the architecture? Do you have thoughts on that, data on that? What's your final word?

Erik Bradley

>> We don't necessarily have the data on that right now. I truly believe from talking to the CISOs in our community and here on the ground that the vendors are going to do what they do. They're going to go put their flag in the ground, say, "Look at us." They're going to overstate their claims, but the truth is I don't think anyone has the control tower at this moment. I think it's going to be a blend. That's why I keep saying, again, depth of defense. It's a term we used to use 10, 20 years ago that I think is more important now than ever. But eventually, we know that it's going to be agent versus agent, but you still need to make sure that you don't have prompt injections, you don't have data leakage, that these things are governed, that there is an identity attached to them so that you know when something goes wrong. And also, you know what's interesting to me? No one thinks about this because it's kind of a little bit taken for granted. But what about the backup and recovery? What about the guys controlling your data, storing it every time? If something happens, you need to remediate it. You need to know exactly when to go back to that time and when to remediate it. So I'm looking at Rubrik. I'm looking at Cohesity. I'm looking at the smaller mid-market. I'm looking at Veeam. These guys are in a really interesting position to move from what used to just be storage and backup into real security.

Dave Vellante

>> I love that you brought that up because our next guest is actually a partnership between NetApp and Commvault. Oh, okay. And Commvault at Shift in New York City this past year, it was in the fall or early winter, announced a term, they coined a term called res ops. And the whole idea is cyber resiliency. So backup and recovery is moving to sort of cyber resiliency. You're certainly seeing... You mentioned Veeam, there's Commvault, there's Rubrik, Cohesity. Dell is a player in that space.

Erik Bradley

>> Dell is still a player in there.

Dave Vellante

>> So it's been a massive pivot into security. Why? Because A, it's a fundamental component of the security stack, even though it's maybe a small slice, and it's vital for, especially in the AI ever, to have some kind of business resilience strategy.

Erik Bradley

>> I think it's more vital now because when we ask people about... We didn't get into this, but the survey also asks about concerns that they have around agentic AI and the concerns really all have to do with data leakage. Whether it's through prompt or through an agent being overprivileged, it's, they're really concerned about data leakage. That's the number one concern they have. So I think it's actually going to become more in the forefront.

Dave Vellante

>> The battle continues, Erik. Great stuff. I really appreciate it. Check out etr.ai if you want to access to this data. It's absolutely unbelievable. There's a lot of free stuff on there. And when you get the priority access like we have, it's just mind-blowing. And the thing is, it's not just a snapshot in time, it's ongoing. So Erik Bradley, thanks so much for coming on The Cube.

Erik Bradley

>> Thank you, David. Great seeing you, buddy.

Dave Vellante

>> Yeah, you too. All right. Keep it right there. We got wall to wall coverage here. This is day three. We're live, by the way, here at Moscone West. I'm Dave Vellante. Kristoff Bertrand is here, as is John Oltsik and the whole Cube team. Keep it right there, right back right after this short break.