Please welcome to the stage, Melissa Franchi.
Okay. Great. Well, thank you. Welcome to our first Analyst Day. Welcome to Las Vegas for the Forward user conference. For the folks on the webcast, thank you for joining us. I'm Melissa Franchi, head of Investor Relations here at Rubrik. Before we get started, I'll do a quick disclaimer. The slides for today's presentation, which include a reconciliation of GAAP to non-GAAP financial results, are available at our investor relations page at www.ir.rubrik.com. These non-GAAP financial measures should not be considered in isolation from or as a substitute for financial information prepared in accordance with GAAP. These statements are based on what we believe today, and actual results may differ materially. These forward-looking statements are subject to risks and other factors that could affect our performance and financial results, which we discuss in detail with our filings with SEC. Great.
With that out of the way, let me give you a quick snapshot of what we have entailed today. I know it's a super busy season, we'll make the best use of our time. We are very excited to do a deeper dive in the breadth and the depth of our platform, opportunity, strategy, as well as the breadth and depth of our executive leadership team. We'll have a variety of different presentations today. We will have some time for Q&A, both in the morning and in the afternoon. We're going to break for lunch. We'll have a customer and a partner panel. For those on the webcast, that will be offline. We'll continue with our presentations. We'll end the day with Q&A. I do ask that everyone reserves their questions for those designated times.
With that, we'll get started, and I am very pleased to welcome to the stage, Bipul Sinha, CEO, Chairman, and Co-Founder. Thank you.
Good morning, everyone. Hope you had plenty coffee, because going at 10 o'clock in the morning after a very busy first day of a conference is a tough time to get started. Are you guys ready?
Yes.
Awesome. I wanted to start with just a quick reminder of Rubrik's mission. When we started Rubrik 12 years ago, our goal was to build an enduring institution. An enduring institution with the cultural foundation that really creates value for everyone in the ecosystem, including our customers, partners, investors, Rubrikans, and everyone. We wanted to also align all of our stakeholders towards a common mission. Our mission is to secure and accelerate world's AI transformation. As you know, AI will probably be the most transformative technology, at least for our lifetime. Every customer, every business, every government, want to take advantage of this AI transformation. This AI transformation creates significant challenges, too. Rubrik's goal is to give confidence so that our customers can transform their business, roll out trusted agents, and really take advantage of 100 times productivity opportunities that AI is bringing.
With that, let me tell you where we are as a business. There are a lot of numbers. Obviously, we are a fast-growing business at our scale. Lots of customers. What I focus on is our top line growing fast? Do we have great margins? What I'm most proud of, more than 80 Net Promoter Score. This is a strong indicator of how strong our business is. More than 80%, I wanted to remind you, is top 1% of the enterprise B2B. We are in a rarefied air. Mind you, we sell very complex operational softwares. All of the Rubrikans, we get up in the morning and think about how do we innovate on behalf of our customers and not just create customer satisfaction, because satisfaction is not a great benchmark. How do we create delight?
More than 80 NPS shows delight in our customer base. You may ask this question, how did we get here? How did we get here to AI as an industry? Because in 12 years of Rubrik, we have seen also evolution, not just of Rubrik, but the whole industry. If you go back 10, 20 years ago, everybody wanted to be digital as step one. From your handwritten CRM on some spreadsheet, maybe, to adopting CRM, adopting other systems of record, the very, very basic things that we now consider almost legacy or obsolete. That was the rage. Once we got done with automation, which means that our systems of record were digital, the whole world started to think about how do we digitally transform ourselves? Because not just systems of record, but how do we work with each other? What are the workflows?
How does business work end to end? Instead of pushing papers based on the system of record, can we have workflows that make all of us more productive? That really created this whole cloud revolution, SaaS revolution, and that peaked in 2021, 2022 era, and that's what gave birth to the AI transformation. First we did system of record, then we created workflows, and now people have started asking this question, that why do we have people clicking buttons? Why do we even have to do one click? Why can't our systems use AI and do the work? That's the new transformation, and that has the promise of 100x more productivity. That is only the upside, because all of these transformations, different eras, created different challenges and risks along the way.
If you think about it, once you created the data center and had systems of record, you had the issues of flood, fire, crater, and you needed solutions to make sure that your systems are up and running. That was the first era of backup and recovery. That's where Rubrik, we as a business, came in, and we said, "We are going to give you automated, API-driven system of record recovery." That's how we started our business, selling for data centers-oriented application recovery. As cloud picked up, we also advanced our solution and started to think about it's not just a natural disaster is the biggest disaster. Cyber is the biggest disaster. You cannot solve cyber recovery with a legacy product that was designed for flood or fire.
You have to fundamentally understand that a clean data recovery is what is needed to recover from digital transformation risk, which was cybersecurity. We built a purpose-built platform for cyber recovery. In fact, Rubrik created cyber resilience as a phrase. The whole world has adopted cyber resilience, but we were the first one who said cyber disaster is the biggest disaster. You have to assume breach, and for you to assume breach, you need to have cyber resilience. Cyber resilience has two parts. Do you understand the risk of cyberattack when you assume breach, and do you have the ability to remediate? We created a beautiful product across complete cyber resilience solution. We have entered a new era, and this era is different.
This era is different, the AI era, because in this era, you cannot practically do any prevention or most of the prevention or any detection. Anybody who says, "I stop anything," is living in the digital transformation era, not in the AI transformation era. In the AI transformation era, attacks are inevitable. You heard it from Mikko, that vulnerability chaining, toxic combination of vulnerability discovery makes attacks inevitable. Cyber resilience is very, very critical. The question is: Is your yesterday's cyber resilience relevant even today? Because it's not just somebody is getting into your organization and trying to steal your data. That was yesterday's problem. Today, somebody sitting in North Korea can run your whole business, can write your code, do customer service, do customer refunds, and make you bankrupt. Businesses face unprecedented risk with AI.
All the prevention and detection solution is not going to help in the AI era. As we have gone from automation to digital, Rubrik is reinventing itself one more time for the AI era. Here is my fundamental assertion. The cybersecurity industry, as we have all known and grew up with, is dead. You may think Bipul is making a bigger statement and trying to shock us. It's not about shock. It's about accepting the reality. Because we were completely focused on preventing and detecting attacks with 100 or 200 solutions sold by thousands of vendors. Think about it. How good are they? If they were so good, why did we have a panic with MITRE? This is what the change is. We used to have human attackers and humans who were working within your companies.
Job of the human attacker was to somehow get in and figure out where your critical data is. Human worker had identity. In the worst case, they will steal their identity, steal their data, or do ransomware attack by going through human identity. That was the world of cybersecurity. That world doesn't exist anymore, because in the new world, you have AI attackers. AI attackers don't sleep. AI attackers don't have to figure out, how do I go find the right data and chain the vulnerabilities, because they're reasoning. They're using AI to reason and find the right toxic combination to attack you. At the same time, more and more workers that run business processes are not human beings. They are agents.
These agents are vulnerable to hallucination because they are probabilistic systems or cyber compromise, and should they get compromised, you have a massive insider attack. You have two huge problems. AI attackers that you cannot detect, and they are in, and doing tremendous damage at machine speed. You have agents, should they get compromised, you have a massive insider attack. When the AI speed breaches happen, and if you have yesterday's cyber resilience that does human speed recovery, there is an impedance mismatch. You can't respond by human speed, which is at best 10 bits per second, to an AI attacker, which is going million to billion bits per second. You will be out of business. That's why everybody's so worried about these modern AI models.
Second problem, these agentic operators, these operators that are agents, if they get compromised, they will do 10x more damage in one tenth of the time. If you have human processes, human set of rules to figure out what is allowed and what is not allowed, or somebody managing these detections, again, you have huge impedance mismatch. Not only your agent, but you will be out of business. The question is, you have AI speed compromise or attacks and human speed response. That's the world that we have lived in. Everybody talks about, "Let's fight AI with AI." The question is, how do you fight AI with AI? You can't fight AI with AI with the same thing that you have been doing or use the same old software.
To fight AI with AI, obviously, you have to deliver AI-speed recovery to make sure that you are in the business after an AI-speed attack. The second piece is, you don't let your agent do bad things in the first place. If they do bad things, you also recover at AI-speed. The question is that, how do you deliver these two capabilities with your same old software? To solve that problem, you need to deliver these capabilities also as AI agent. You need AI agentic orchestration to deliver cyber resilience, to deliver agentic guardrail. That is the requirement. If you're not following these requirements, if you're not building capabilities on these requirements, you're not relevant in the AI era. That is challenge to everyone in the cyber industry.
We started Rubrik 12 years ago as, at that time, even today, the best software that you can deliver in the marketplace with API, with automation, but we still needed human beings to be able to deliver these capabilities of automation, ease of use, beautiful UI. You needed human beings to do the recovery. That changes today. We are making Rubrik an agent. An agent can only respond to AI-speed attacks that are coming in its way. How does this manifest? Rubrik AI. That is Rubrik agents for all aspects of Rubrik, that never sleeps, that monitors, collects information, understands what's going on, creates recovery plan, asks human for permission to not work and creates outcome.
If human says, "Yes, do the recovery," it will fully orchestrate recovery, bring the business back to its operational state, make sure that your agents are back into operating state without human doing the work. Think about humans are now a manager. They give prompt, "I want to get this work done." If Rubrik AI needs to ask for permission or a judgment call, they will involve the human in the loop, otherwise, they'll create the outcome. That's the new world, and that is the new Rubrik. We are manifesting this strategy as agentic cyber resilience. That is one more time, Rubrik is redefining what it means to have cyber resilience in the AI era, just like we defined cyber resilience itself for the digital transformation era. You may ask, "How does it work? What are its foundations?" Obviously, agents can't run on ethers. It requires data.
It assumes identity. Rubrik platform brings together all data and all identity from all the business process, business systems across on-premises, cloud, SaaS, identity everywhere onto a single platform. On top of it, we have built Rubrik agent orchestration. These are the set and hierarchy of agents that is built into the Rubrik platform to create agentic outcome, where these agents are doing the work, not humans. You may say, "How about these agents? Can they hallucinate?" Yes, they are also probabilistic systems. We have also built agentic governance and security onto our platform, built into our platform. This agentic governance layer not only governs the Rubrik's own agents, but we are also selling it to our customers to make sure that we create runtime guardrails for customers' agents for sales, marketing, support, and other systems.
In some ways, we are drinking our own champagne to ensure that our own agents are governed and secure, and we are also giving the same capability to our customers to govern and secure their other agents. You will ask this question that, "Bipul, why did you change Rubrik from software to agent? Obviously, the speed of recovery, but what is a higher level governing thesis?" If you think about it, our customers are deploying marketing agents, sales agents, coding agents, customer service agents. When they come to IT, they don't want to manage the software. That is not only creating more work for them, but it's also not ready for AI-speed recovery. Rubrik now fits in the AI framework that our customers are creating across the whole enterprise. Rubrik is an agent for cyber resilience, agentic governance and security.
It's all built into our platform. This platform delivers two product suites, Rubrik Security Cloud for cyber risk assessment plus recovery, and Rubrik Agent Cloud for agentic security and governance. How does Rubrik Security Cloud deliver AI-speed recovery? It preempts the work needed for recovery, and that preemptive recovery engine does the work in peacetime to be able to do AI-speed recovery in wartime. It automatically understands what identity systems should be recovered or rolled back, and then autonomously creates minimum viable business plan all built into the platform. Rubrik Agent Cloud not only enumerates and monitors both sanctioned and unsanctioned, which is shadow IT agents, but also creates runtime security guardrails, so that should your agent take action or attempt to take action that is not commensurate with the policy, it is stopped. We are applying AI to actually create the guardrails.
Finally, rewind is what rewinds the whole agentic work. If you come to it, agentic cyber resilience is delivering two outcomes for our customers. We are telling our customers to assume breach, which means that in the AI era, attacks will come to you, and it'll be continuous. How do you deliver AI-speed recovery, as well as assume agentic overreach where agents will misbehave, and how do you ensure that your agents are not doing damage to your business? As I was preparing these slides, our Rubrik Innovation Lab came to me and said, "Bipul, remember Rubrik Annapurna? We have been working fast and furious on Rubrik Annapurna, and we were trying to figure out where does Rubrik Annapurna really create the maximum value. Because Rubrik has all sorts of data, but where does we can really make an impact?" This is the updated Rubrik Annapurna.
What we found was that vast majority, 80%-90% of enterprise data is unstructured data sitting in their file system or in their S3 or object store. It is very hard or very expensive and very slow to understand the classification and nature of this data. It is such a high volume data that if you load all the data before you can figure out what you can use and what is the classification and categorization of this data, it's a very expensive proposition, and people don't do it. These data is like your X-ray data, your semiconductor design data, it is your manufacturing images data to see that the batch manufacturing is going all right. Lots of sensor data that all comes in.
Rubrik not only breaks the law of physics but also economics to deliver unstructured data AI by creating catalog integration with Unity Catalog and providing catalog before loading data, so that people can figure out what is the right set of data they want to actually apply AI to. Once they apply AI to the right data, and we actually deliver data in the tabular format so that they can create an AI outcome and also have the provenance and history on what data they created the AI solution. We believe that it's a very important area. Could be an important market opportunity for us. Early days. It's still in the innovation lab, but very excited.
In summary, we have one platform solving for two imperatives, assume breach and assume overreach, across two suites of products built on a very strong foundation of the Rubrik platform, and that is agentic cyber resilience. What is our market opportunity? Where we can and who we can sell this product to? Rubrik sits at the intersection of data, identity, and now AI. You look at the data piece, that has been our heritage. We help our customers secure their data, we help our customer understand their data, we help our customer recover their data and restart their businesses. We are doing across data dimension, both risk of the data and remediation of the data. You look at identity piece, we have a strong business in identity resilience across two vectors also, identity risk and identity remediation.
With AI and with the Predibase acquisition in creating this runtime guardrail engine, we are doing three things, understanding all your agents, creating guardrails for your agents, and then doing agent rewind should anything go wrong. A complete platform end-to-end. You may ask, "Bipul, why are you doing AI?" We are doing AI because agents assume identity and work on sensitive data. Rubrik has full enterprise context on data, full enterprise context on identity, so it's very natural for us to give full security for agents. We are doing all of it on a single platform with multiple products. That is where the Rubrik's huge consolidation opportunity is. You look at our customers. Just across data dimension, whether it's enterprise on-premises or cloud or SaaS, buying numerous point solutions, these solutions don't talk to each other.
Should anything happen to your data, you have to turn 30 knobs, one for each service within AWS, to be able to recover. Rubrik has a single policy engine, single platform, single way to bring your business back into action. Now we are doing two agents, so at AI speed. You look at identity dimension, again, our customers have multiple point solutions they are buying for identity posture management, plus identity recovery, and if you have different kinds of identity platforms, they have one solution for each. Rubrik is one platform, one solution across all your identity. Similarly for AI, you may hear there's a lot of companies trying to do AI agent security. Some people are only doing observability. Some people are only doing identity risk. Some people are only doing runtime. Some people are talking about agent rewind.
Rubrik is one platform where all four capabilities are built into the platform with full enterprise context around data and identity. That's why we believe that we have a unique opportunity to not only solve the data issues, but identity, plus AI. That's what creates significant TAM for us. From data protection, expanding into security and resilience pieces, another large market opportunity, expanding into identity resilience and identity recovery solutions, and then expanding into a massive AI deployment and security. If you combine all of it, we are not opportunity constrained. Over $125 billion market. Rubrik didn't get there with this number to start with. We always build solutions and expand their TAM.
Every single product, every single vision, every single market that we enter, we think about how do we get bigger share of the wallet, how do we consolidate for our customers, create a single solution, a single platform that is logically and intellectually, for our customers, connected, and deliver a complete end-to-end solution. It creates a significant flywheel for our customers and us, because our customers also want to deal with fewer vendors. They want to adopt more of Rubrik solutions, they can start with Rubrik in one of the six or seven ways. They can just start with the AI solution. They can just start with the identity solution. They can just start with the cloud solution. They can just start with the SaaS solution. Just start with on-prem data center solution. Any product they buy, we instantiate Rubrik platform for them.
Single policy, single platform. They just click more buttons and expand the solutions. That's the power of our platform, and that's the power of flywheel. That is the reason that when we introduce new products, we are able to scale it quickly because they don't have to buy yet another product and think about yet another way to manage it. It's fully and completely integrated onto our platform. This is a practical example of how it manifests for Rubrik as a business. When we first introduced cloud or cloud solutions, it took four and a half years to get to $100 million ARR, because this was our first attempt to the multi-product thinking. When we introduced our M365 solution, it went faster because customer understood that it's a single platform. You have to flip a switch and adopt a new platform.
It went from four and a half years for cloud to three and a half years for M365. When we introduced Enterprise Edition data security products to detect ransomware, to do data classification, to do threat hunting monitoring as a package, it took only two and a half years to get to $100 million. Again, it shows that the effect of the platform. Our platform is based on complementary network effect. What is complementary network effect? When you adopt more solutions, the value of all your existing solution go up. That's why it accelerates. When you introduce new products, it goes faster because then value of all other products increases. When we introduced Identity, in just over a year, we went to $50 million from cold start, organically built product. That's the platform effect, not platformization, real platform.
Our customers are also saying, "You guys have lots of products. We want to consolidate. Make it easier for us to adopt Rubrik." We are launching Rubrik Flex, which is truly a platform contract, one contract, one commitment. It helps our customers adopt our products faster, adopt new products faster. They can pick what products they want to consume, and they go. We are very proud of our team and our velocity of execution, not only in building products and taking to market, but also scaling the business operationally, creating a disciplined execution engine that takes this suite of products across the whole platform to the customers and scale the business. We are just getting started. In my mind, company's age is in decades, not in years. We just finished our first decade. We are on to our second decade.
We have the product, we have the technology, we have the team, and most importantly, we have the entrepreneurial hunger to go build a lasting company. There is Rubrik, the security and AI operations company. Thank you.
Great to see all of you here in person. I'm Arvind Nithrakashyap, Co-Founder and CTO of Rubrik. I've been in the industry for a long time. I'm probably dating myself. The internet was nascent and e-commerce was nascent when I first entered the industry. I've seen the internet transformation, I've seen the cloud transformation, I've seen the mobile revolution, but I've never seen change at this pace in the tech industry before. Organizations and businesses are dealing with a lot. When you think about risk, there is a whole new dimension of threats that they have to contend with today. Cyberattacks which were previously engineered by smart hackers sitting in a remote corner of the world are now going to be engineered by AI. What this means is that not only are breaches inevitable, they're going to become more frequent.
Of course, everybody wants to deploy agents. There's a huge promise of productivity, 10x, maybe even more with agents. They also present a huge risk to the critical business data and assets. As we all know, agents can hallucinate and make mistakes. We've all seen that every time we've used a chatbot. It's also pretty easy through malicious prompts to coerce them to do something that they should not. The challenge is all of this is happening at AI speed. Things are changing so rapidly. The only real solution for organizations is to be resilient. They have to know that these things will happen, and they have to be resilient and be able to get back up. If you think about it, any solution to this problem has to span across data, identity, and AI.
Most vendors and platforms focus on just one of these areas. If you think of a cyberattack, you need to understand what are the identities that are compromised. You need to understand what critical business data could they access. Then you need to be able to recover all of this if they do destroy your critical business data. Again, agents also assume identities, operate on critical business data. To be resilient to agent mistakes, you again need to understand both data and identity. Fundamentally, you need a platform that is the intersection of data, identity, and AI, and Rubrik is the only platform out there. We ourselves started in the data management space, but because of the unique architecture and platform that we built, we were able to bring that same platform into identity and provided solutions on identity.
Today, as agents become more mainstream, we are able to now bring that same cyber resilience capability for agents as well. Let me take you back in time about when we started this company and we built this platform. When we started Rubrik 12 years ago, typically what the solution said was really bring together multiple products. You would bring a backup software, say something like Commvault or Veeam. You would bring backup storage, say a Data Domain. You would buy media servers from Dell and HP or other such server vendors, put it all together and build a solution. The challenge with this is each of these components only have a partial view of the enterprise data. Nobody has a holistic picture. As a new startup, we had a clean sheet of paper. We said, "Why don't we reimagine this whole thing?
What if we built a single stack of software which captured data and metadata, and we created this concept called self-describing data?" This is an extremely powerful concept, and I just want to dwell a little bit here. Let's take an analogy. Let's assume you have a bunch of rooms. They're all locked. You have a bunch of keys. One of them has a safe with the crown jewels in it. If all you have is a bunch of keys, you don't know which key opens which door, and you don't know where the safe is. You have to go try each key on each door and figure out where the crown jewels are. Let's say the keys came with a little tag.
It said, "Hey, this is the door this key opens." It also tells you, "Oh, by the way, this is a room that has a safe." One, you know that is an important key you need to safeguard, and you know that if you need to get to the crown jewels, you can get to it very quickly. This is the power of self-describing data. What it does is it gives you a map of the entire enterprise. It shows you where the critical assets are. It shows you the identities or users who operate on the critical data, and it tells you what are those critical data and identities you need to monitor. Today you hear a lot about enterprise context, how that's the most important thing for AI. If you think about it, this is the crux of what enterprise context is.
Your critical applications, your critical data, and the users who operate on them. What we do is, we don't just have a point in time of the data and metadata or the self-describing data. We take snapshots of this over time. We actually have the full history. We have a full time series of how enterprise data and identities are evolving. That is what allowed us to capture the full context of applications, identities, and build a data threat engine natively into the platform. That is what allowed us to say, "Where does your sensitive data lie? Who are the users who have access to it? Who are the users who may be accessing it? Are there threats in your system? Could there be an attack that's happening right now?" All of this information is because we have this enterprise context built into the platform.
Obviously, we've been delivering cyber resilience, I call it for the software era. Now as we enter the AI era, everything is happening at AI speed. You need to be able to have AI help you recover your business as well. This is where we were able to build this agent orchestration layer. Again, you cannot do this unless you have that enterprise context. Because the agents themselves know how to access applications, how to secure them, how to set the right SLAs. That's how they're able to completely automate your Minimum Viable Business recovery. It's the same enterprise context that allows us to understand what are the applications agents that are now being deployed access, what kind of identities they assume.
When an agent takes an action, we have the understanding of what the intent behind that action is, then we are able to determine whether it's in violation of an organization's policies or not. Then if the agent does still make mistakes, we can surgically understand the actual operations that were erroneous and surgically rewind them. The challenge with cyber recovery, it's not just about recovering the data. You just can't go to yesterday's backups and recover the data. The challenge is that you need to answer a lot of critical questions before you can even start on a recovery action. Many industries, regulated industries, they need to know, was there any customer data that was compromised? Healthcare, was there any patient information, HIPAA data that was compromised? They have an obligation to disclose this to their customers. You need to understand where the sensitive data lies.
You need to understand before you can recover, you need to know what the source of the attack was. Was that malware sitting in a bunch of applications within the organization? This is the reason why cyber recovery takes so long. Gathering all these answers requires you to scan large amounts of data. If you want to recover at AI speed, you have to pre-compute all this information beforehand so that you're ready to go when an attack happens. This is fundamentally what the preemptive recovery engine we have built into the platform does. As we ingest the data, we are able to capture the data and metadata for both data applications as well as IT systems. We create a uniform representation, which builds a time series over time. Then we index the metadata, and we also hash all the content.
This is what allows us to, for example, if a new threat is published today that nobody knew of until now, within a few seconds we can go and say, "Oh, by the way, yes, this threat is present in your organization. It originated in this application 10 days ago, but now it's present in 15 different app systems within your organization." This is the power of doing all of this pre-computing all of this, so that you can quickly answer these questions. When you are hit by a cyberattack, you don't need to go scan and find this information. The information is at your fingertips.
This is also what allows us to actually figure out the right clean recovery point that we can then say, "This is the point you have to recover at so that you know the malware is no longer present in the system." This is the key behind the preemptive recovery engine that pre-computes all the information peacetime so that you can quickly act when there's an attack. Identity has become the biggest attack vector. Almost 80% of attacks, the hackers aren't hacking into your systems, they're logging in. The interesting thing about such identity compromise is that these often act as backdoors for attackers to do future attacks. This is why it's extremely important to understand the relationship between identities and data.
If you think about a hospital which has critical patient information, or a bank that has critical financial information, you need to know where this data sits, which applications this data sits in. You need to understand who are the users who have access to this, so that you know when certain users compromise, you know to react immediately. Without this knowledge, you'll be inundated with a lot of alerts, and you won't know which ones are important, which ones are not. Our platform applies the same methodology that we apply to applications where we capture data and metadata so we understand user hierarchies, we understand users, groups, roles, and how these privileges change over time. Obviously, we also understand how applications and data have evolved over time, where the sensitive data sits. And we are also able to understand the relationship between identities and this critical business data.
Obviously attacks are happening at AI speed, so you can't have humans trying to do the recovery. Think of what it takes if your cloud account or your secure billing application gets hacked. These are all the steps you need to take to get your business up and running. You don't want to be doing this on the day of the attack. It's going to take weeks to get this done. Because it takes so long and it's so complex, most organizations would love to do this on a regular basis before an attack happens, but they don't do it because it takes a lot of effort and time to be able to do this. Think of this in the agent era. The human comes and expresses the outcome they desire. They want to secure the billing application.
The agent goes and figures this all out, figures out the dependencies, builds a runbook, it says, "Hey, here's a runbook. Can you validate that I've got it right?" If a human says no, it goes back, figures it out, comes back with a new runbook. The human says, "Yes, this looks good. Go ahead and test it." It goes and tests the recovery to make sure that it works. If it doesn't work again, it'll go back, figure out what went wrong, fix it, finally say, "Yes. I have a working recovery runbook." They can say, "If you want, I can schedule it to run monthly and send you a report." Now when a CIO or CISO goes to the board, they can say, "Yes, I am prepared for a cyber attack because you know what?
I've done this 12 times over the last year. I know this works." This is the confidence that customers need to have in this era of AI-based attacks. If you look under the hood, what we have done is really built an agent orchestration layer. Again, taking this enterprise context that we have in the platform, we run multiple agents, which are all understanding their sensitive data sets, where there's malware, how to calculate dependencies, build a runbook, humans are mostly reviewing and approving. This is how we have to operate in this new era. Finally, let's talk about as organizations deploy agents are also operating at AI speed. You can't have humans stepping in and monitoring every action that the agent takes. This is why you need AI to govern agent actions.
Obviously, with the enterprise context that we have, we already understand where agents, what identities they assume, what applications they access. In terms of dynamically controlling the actions that they take, with the Predibase acquisition, we have actually built small language models that are able to efficiently just take a policy written in natural language, based on the actions that the agent is taking, they can make a decision whether it's in violation of the policy or not. If it's in violation, it will block the action, so that action doesn't even hit your production. Let's say you missed one case in the policy and something sneaks through. We can, again, surgically we can rewind the action that the agent took.
All it takes to fix that hole is to go back to the policy, add a line that covers that case, now the next time, that action will be blocked. I talked about the platform, what does it take to build such a platform at scale? We have actually been securing applications across the enterprise: data center, cloud, SaaS, identity providers, now agents. What we have essentially built is an ultimate system of record, which is a system of record of last resort. When everything else fails, customers come to us. Such a platform has to be able to restore data to production, we can't get one bit of information wrong, because if we do, the applications won't come up.
This is why this is an extremely complex platform to build, and it requires a very deep understanding of applications, identities, and the relationships between them to be able to do this across all the different kind of infrastructure that an enterprise has. Just that is not enough. These applications, each enterprise configures these applications in very different ways based on their business needs. These applications explore in the thousands of configurations. There is no manual or runbook that tells you how enterprises configure these applications. The only way to do this is to work with thousands of customers, understand the different ways in which they configure these applications, bring that knowledge back to the platform, and iterate. This is what we have been doing over 12 years with thousands of customers.
It requires a lot of time and deep relationships with enterprises to be able to get the platform to where we are today. That is why a small startup or two kids in a garage cannot whiteboard this platform. Again, there you have it. We have built a platform with a very unique architecture, with this fundamental concept of self-describing data. This allows us to deliver agentic cyber resilience to our customers. We deliver it through two product suites, the Rubrik Security Cloud and Rubrik Agent Cloud. They address the assume breach and the assume agent overreach problems. To talk in more detail about these product suites, I would now like to welcome Anneka Gupta, Chief Product Officer. Thank you.
Good morning, everybody. I am Anneka Gupta. I am the Chief Product Officer at Rubrik. I first want to say this is such an exciting time to be in product. I do not think that there is a moment in history more exciting than right now, given the speed of transformation that AI is creating for all businesses. You heard from Bipul why agentic cyber resilience is the most important problem that enterprises need to be solving today. You just heard from Arvind about the ingredients that make up this platform, the enterprise context, the preemptive recovery engine, the agentic orchestration, and guardrails that make that up. Now I want to take you a few clicks deeper into our actual product suites and product portfolio and show you the power of our multi-product platform.
To begin with, I really want to start with autonomous Minimum Viable Business, because this is a place where we are pouring in hours and tons of people and time to actually innovate because this is a place where organizations have incredibly important and complex business outcomes, and technology has not delivered on the promise of being able to secure these applications. I want to take you into the day of an organization getting hit with a cyberattack. Imagine you are an organization, you get hit with a cyberattack. The very first call the CEO is going to make is to the CISO, and he is going to ask, "Is our business up and running?" The CISO is going to go to his team, his or her team, and say, "Okay, what are our critical business applications? Are those still up and running?
Can we recover them?" Those applications are what make up your Minimum Viable Business. Why is this really hard? If you think about an example application, applications are made up of many different components. You have data, you have identity, you have configurations, you have infrastructure. All of these pieces come together to build your application. If you go to recover this application and you get the order, the sequencing, the dependencies incorrect, then your application will not run, and your business will be down. Layer on top of that a cybersecurity incident and what you have to do. Before an attack, you have to take these incredibly complex, dynamic applications and be able to map all the dependencies, map the full topology.
You actually have to create your recovery plans ahead of time so that you can make sure that that application will be up and running. On the day you're hit with a cyber attack, you then have to determine the scope of the attack. What identities, what data was actually compromised during the attack? You have to figure out the point of infection. When did the attackers first get into the system? What are all the backdoors that they created? You need to look at where the actual malware or other indicators of compromise first landed in your data so that you know where the clean point of recovery is. You need to assess the sensitive data impact. For most organizations, especially highly regulated organizations, they're going to have to report this back to their auditors, to their regulators. Then you have to quarantine the malware.
You have to make sure that the bad actors don't re-enter your platform when you do your recovery. After all of that, you have to orchestrate the recovery. This is an immense amount of steps, and when you think about the complex applications that make up an organization's Minimum Viable Business and the multiple steps that they have to take for each application in a cyber recovery incident, and you multiply that and compound that with the fact that attacks are happening now at AI speed. The reality is human operators just can't keep up. That is why we built Rubrik AI. With Rubrik AI, all of these steps that I told you about, both before the attack and during the attack, are completely absorbed. All that complexity is absorbed and replaced by agents.
Now I want to show you what Rubrik AI really looks like in action. Meet Malika. She's an IT admin, and she just woke up to her worst nightmare. A security incident has occurred within her Microsoft 365 application. By the time Malika sees the alerts through our integration with CrowdStrike, Rubrik AI has already done the investigation. It knows what was compromised and built a recovery plan. Malika didn't have to figure it out, any of that on her own. All she has to do is review and approve. In the old world, this moment kicks off days of chaos. Security teams scrambling to understand what happened, recovery teams manually piecing together what to restore, executives asking for timelines no one can answer. It's expensive, it's slow, and it's very manual. Here's what's different with Rubrik AI. She sees that the finance group was compromised.
Eight users were deleted, the attacker escalated privileges to get unrestricted access. Typically, putting together these insights would have required collaborating with multiple teams, jumping between multiple tools. Now she can get a clear summary of how her identity systems were compromised all in one place. She also sees the attacker deleted and downloaded a bunch of data. This is a full-blown ransomware and data exfiltration attack. Rubrik recommends a previously created recovery plan. It's not asking the customer to figure out what to restore and in what order. This is the moment that used to take days. Building a recovery runbook, coordinating across teams, deciding what's critical enough to restore first. Rubrik AI surfaces the details of the identity recovery plan. You can see exactly what's going to fix. The AI has scoped this precisely. It's not just restoring everything blindly.
It knows which users were affected and exactly what needs to be rolled back. One prompt and it's done. With the recovery of identity initiated so users can go back and log in, Rubrik AI immediately follows up on recovering the data. It's staying in the loop, not just firing and forgetting. Rubrik AI pulls up your data recovery plan, which shows the data that will be recovered based on the scope of the attack. Rubrik is orchestrating the entire M365 recovery autonomously. Instead of security and IT teams having to piece together what actually happened during the attack and what to restore to get back to a clean state, Rubrik AI is giving transparency and a clear recommendation. Done. The minimum viable business recovery is in progress. What previously took admins hours and days to investigate is now down to a couple of prompts, minimizing business impact.
You can see here the power of Rubrik Security Cloud to consolidate the work of multiple teams across security and IT during a cyber attack. This is the reason why having all of these capabilities available in one platform helps organizations accelerate their cyber recovery, especially in the face of AI-speed attacks. Rubrik Security Cloud is a platform, within this platform, there are multiple products we sell. We sell Oracle protection, Nutanix protection, RDS in AWS, M365, Salesforce, DevOps protection, unstructured data, identity. With all of this, our Rubrik Security Cloud is integrating with leading cybersecurity vendors who focus on detection and prevention so that we can bring together detection and prevention with cyber resilience to help teams fight AI with AI.
You layer on top of that Rubrik Agent Cloud, which is helping monitoring for agents, providing the agent guardrails, providing rewinds in the inevitable case where an AI agent goes awry. Now together, we have an incredibly powerful set of product suites. With Rubrik Security Cloud and Rubrik Agent Cloud, we have a platform that delivers true agentic cyber resilience, cyber resilience for the agentic enterprise and agents for cyber resilience. Rubrik is a platform, we have numerous products and solutions we sell to different buyers. Agentic cyber resilience is truly an enterprise-wide problem. It matters to the CIO, it matters to the CISO, it matters to the CTO. What does the CIO care about? The CIO cares about business transformation.
When they come to Rubrik, they buy Microsoft 365 protection, cloud protection, data center protection, identity protection, and they're looking at all of this and consolidating multiple solutions into one platform. They don't have to buy point solutions, they just buy the Rubrik platform. When you look at the security teams and the CISOs, they're in charge of managing risk, which is only getting more and more complicated in an AI-powered world. They come to Rubrik, and instead of buying these niche solutions for protecting identity, whether it's Active Directory, Microsoft Entra ID, or Okta, and piecing together both the risk and posture piece and the remediation and not having any context of the data, they buy the Rubrik platform. One consolidated place for managing risk for data and identity. If you look at the CTO, what does a CTO care about? They care about accelerating agent deployment.
Instead of buying all these different point solutions and trying to layer together piecemeal a strategy for how you secure agents with monitoring, with real-time runtime protection, and with rewind, they buy the Rubrik platform. All of these organizations and teams are coming together because they care about agentic cyber resilience. How do we actually capture the value across all of these different buying centers and all of these different personas? We've developed a multi-vector monetization model that really meets our customers where they are. Whether a customer wants to start with data on-prem, cloud, SaaS data, whether they want to start with identity, whether they want to start with AI agents, they're able to pick and choose and expand with Rubrik, across all of the surface areas that matter in their enterprise.
If you look at these surface areas and if you look at the way that we monetize these surface areas, all of this is accelerated with AI. AI creates more data, AI creates more non-human identities, and of course, as organizations adopt AI faster and faster and deploy more agents, that's only going to grow exponentially the amount of agent activity. The second vector of our monetization is really looking at the cyber resilience maturity of an organization. We've developed editions that allow an organization to land with some of the more basic capabilities, but then as their resilience matures with the organization, they can expand with us. What we find is that many customers start with one, two, three products, and then they want to consolidate more and more and more solutions with Rubrik across our platform. That is why we launched Rubrik Flex.
Rubrik Flex is one license, one contract, one commitment for agentic cyber resilience. Similar to other large platform companies, we've developed the solution so that no matter where an organization wants to start, maybe they want to start in the data center and they want to go to cloud. Maybe they want to start with SaaS and then go to identity. Maybe they want to start with their applications and then go to agents. We're able to solve all of that on their timescale with Rubrik Flex. This gives customers one commitment and ability to access all of the products and flexibly allocate their spend across those products. Rubrik Flex is going to enable Rubrik to sign larger ACV deals, massively differentiate from competition, and enable frictionless sales for our sales teams. As an example, we recently signed a Global 2000 insurance company on Rubrik Flex.
It was an over $6 million ACV deal, over $30 million TCV deal over five years. They ripped and replaced legacy backup solution for their on-prem environment as well as their cloud-native backup solution. They chose Rubrik for agentic cyber resilience. They're starting with three products to begin with, and they have a plan to scale across the entire portfolio. Our belief is that Rubrik Flex is going to become the default motion for selling to tier 1 and tier 2 customers, who really buy into our vision of agentic cyber resilience and who buy into these two product suites of Rubrik Security Cloud and Rubrik Agent Cloud. In the almost five years that I've been with Rubrik, it still feels like day one because of the pace and velocity of our innovation.
We have a deep culture of being able to stack S-curves, to continuously innovate on the edge so that we're constantly differentiating from our competition. We're doing this while building all of it on one platform, even when we make acquisitions. All of this is super exciting because as organizations adopt more and more of our products, they get compounding value from those products as well. If you look at our innovation history, we started with enterprise data protection in the data center. We then layered on cloud data protection to expand to cloud and SaaS protection for all of the SaaS apps that are critical for running an organization's business. Then as the world realized that ransomware attacks were prevalent and you needed cyber recovery, we layered in data security and continued to innovate on that frontier as well.
As attacks moved from just ransomware to actually identity as the primary attack vector, we built identity resilience. Now, in the past year, we've layered on agentic security on top of that. As you can see, when we say we're multi-product, we are really multi-product. Customers are sometimes consolidating up to 10 different solutions that they were using for cyber resilience that were disconnected, that didn't talk to each other into Rubrik when they buy Rubrik. One of the real keys to this success, beyond all the R&D work that we do to drive this innovation, has been our go-to-market engine.
We have our forward team with Jesse and Alok leading the way to make sure that we're continuing to sell the products that are tried and true, and we have our Rubrik X team, led by Mike Tornincasa, who helps us incubate and take new products to market as our lateral motion. Now I would like to invite Alok Agrawal to the stage, and Jesse Green, who's going to be with us on Zoom, to talk about our forward go-to-market motion.
Thank you. Hello, everyone. Hopefully you can hear me in the room. I'm Jesse Green, Chief Revenue Officer at Rubrik, and I'm joined with Alok Agrawal, our Chief Solutions Officer. I just want to start by saying I wish I could be there in the room with all of you. I know we haven't met yet. I was in Vegas at Rubrik Forward for the past 2 days, but I caught a red-eye last night to get home for my son's graduation, which is one of those moments I need to show up for. Hopefully, I'll get a chance to meet you next time we get together. Maybe just a quick background, since this is the first time we're meeting.
I joined Rubrik 3 years ago from MongoDB, and I was at Mongo for 8.5 years, joining pre-IPO when we were around $50 million ARR at the time, and then leaving post-IPO well north of $1 billion. I was running the Americas. When Bipul recruited me to Rubrik, we had discussed moving into the CRO role really at the appropriate time, and that came this past quarter. I've been in seat for a quarter in the CRO role, but I will say the transition has felt relatively seamless, given that I spent the last 3 years at Rubrik as the president of the Americas, building out a repeatable go-to-market playbook, hiring the team, and this playbook will translate really well internationally. Alok, maybe you could introduce yourself as well.
Absolutely. Thank you, Jess. My name is Alok Agrawal. I joined Rubrik about 6 years back to lead our global sales engineering team. At that time, we were just starting our journey from single product to multi-product, and I helped architect taking our platform, creating a solution package, and taking it to market. Last year, my role was expanded to also lead our global partner org. Our idea here was, how do we create and bring more of an engineering and solutions focus to our partner ecosystem so we can take our multi-product platform and combine it with our partners' expertise in consulting, services, and last-mile customizations, so we can jointly create and add more value and outcomes for our joint customers. Over to you, Jesse.
Thanks, Alok. Look, over the next 15-20 minutes, we'll take you inside our go-to-market, and our hope is you walk away with a better understanding of our strategic priorities. We'll hit on a couple of our independent growth levers, and then hopefully you will walk away with the confidence that our go-to-market is built for sustained, durable growth. Why don't we get started? Alok, if you could flip to the next slide, please. We'll start off with our 3 strategic priorities, and we'll drill into each of these further in the subsequent slides. The first one that you heard a lot from this morning already is our multi-product platform, and it's specifically the go-to-market motion we have across these. Just as a reminder, this is our agentic cyber resilience platform, which now spans data, identity, and AI.
It provides us a critical growth engine so that we'll drive our land and expand and explode motions, which we'll talk further about in a few. This is driving for us bigger deal sizes, higher attach rates, better expansions across more product categories, and industry-leading net retention. If I shift over to the middle, our second is really our evolution to an AI-native selling motion. This, for us, is really about the efficiency of our sellers and the productivity where we embed AI across the entire sales process, everywhere from research to meeting prep, all the way through post-sales. We're measuring the outcomes, which we'll hit upon. Then third is really our distribution leverage and reach, which Alok is going to spend more time on later on in the presentation.
This, for us, is really about how do we scale through our global partners, and how do we leverage our internal transformation team, which is really unique to Rubrik, really as a growth multiplier, extending our reach both up as well as across org structures as we now sell across multiple buying personas, as Anneka just mentioned. Next slide, Alok. Let's first dive into our segmentation strategy so you know how we're organized. We'll start with our serviceable market, which to us is about 100,000 organizations, which we've defined in our ICP. Within each one, as I mentioned, we will now sell in multiple different areas across that organization. It's no longer just selling into the CIO organization for data center protection. We now have the ability within these 100,000 accounts to really target the CIO org, the CISO org, and the CTO org.
Then when we looked at segmentation, we really looked at it horizontally, in 3 ways to slice up this serviceable market. The first is strategics. These are, think of it as your Fortune 500 organizations and a subset of our Global 2000. This is a much bigger high-touch motion, right? More complexity in the sales process. The sellers here have smaller books. They're certainly more intimate with their accounts, and they have strong relationships with the partners across these core initiatives. Then you move down to our enterprise, which we define as companies above $500 million in annual recurring revenue, and we refer to this as our core. This is a critical segment for us. Then, of course, our mid-enterprise, which is up to $500 million.
Increasingly in this, you'll see a lot of digital, a lot of cloud, a lot of AI-native companies here, and you have more of a high velocity and less complexity motion sellers earlier on in their career and working on scale plays. Then if you take our segmentation and you break it out by sort of market penetration, you can see in strategics we've now landed north of 40% of the Fortune 500, which to me proves that we have sort of the talent as well as the platform, the playbook to go win in the most demanding accounts and the most sophisticated companies in the world. Yet in enterprise and in mid-enterprise, we're still only in the low teens and in the single digits of penetration.
To me, this presents an opportunity as we've proven we can win in the most difficult, demanding use cases and organizations, and we can scale with them. Now we have the opportunity to go penetrate the serviceable market in some of these other segments, and there's a big road in front of us. Of course, across each of these segments, you can kind of think of it as different sales plays, which we've defined and trained the field on. For example, in mid-enterprise, like I mentioned, we're focused more with the value-added resellers on scale plays, focused more maybe on our SaaS solutions, our identity solutions, our cloud protection. Where in strategics, we have a very large opportunity to do significant transactions around larger initiatives. Things like legacy modernization, things around cloud migrations. We partner here very heavily with the global systems integrators.
We're super excited about the press release that came out yesterday around agentic cyber resilience for cloud and code, partnering with some of the GSIs mentioned in the press release. We're working on large transformational cyber resilience initiatives across these. Just one other point I want to emphasize here before we move on is that we're not concentrated in any single vertical within these segments, although we do have several high-growth industries, specifically healthcare and banking, financial services and insurance, and public sector, where we have the most market share, we're growing faster than some of the other segments. We still are very relevant in segments like retail and manufacturing and services and high tech and others. This gives us the ability to be very opportunistic as we look at ways to grow our segmentation strategy and potentially some vertical motions as we scale.
Really the takeaway here on the slide that I wanted to leave you with is that we're still in a very large, very under-penetrated serviceable market, and we have several opportunities where we can evolve this segmentation strategy so we can acquire more market share. Alok, if you could go to the next slide. We'll just turn to our global opportunity. Right now we've sold products across 79 countries, despite only having, I think, sellers in feet on the street, direct sellers within our organization within 26.
We obviously have some partner coverage in others, but directly we only have sellers in 26. Given that we have demand and deployments in dozens of markets where we have not yet put dedicated coverage, and for some of these markets, by the way, I won't call them out, but several of these markets are places where peer companies like, say, a Datadog or a Palo Alto or a MongoDB are doing well north of $100 million in ARR. We believe this provides us a good area to be opportunistic as we look at scaling out our capacity. When you look to the bottom right on the revenue mix, you'll see that 68% of all of our subscription ARR is coming from the United States, with 32% coming from the rest of the world.
Currently, we're seeing especially strong growth in our international markets, we'll continue to invest, we'll scale international opportunities at the appropriate time. The way I break this map up from a go-to-market standpoint is, from a sales leader perspective, we have four key theaters with leaders that report directly to me, we have the Americas, we have EMEA, we have APAC, given the opportunity and some of the nuances in Japan, we have that broken out as well into its own theater. Next slide, Alok.
Data protection, we could start with cyber recovery, it could be identity resilience, it could be SaaS, now agentic resilience. For us in the go-to-market org, this is effectively a playground for our sellers as we no longer are dependent on a single product area or a single buying center to get in the door. These solutions now span, as I mentioned, multiple different buying centers across these orgs, we have several plays that allow us to go identify and figure out how we can enter into an account. It also means that even if a competitor's in there in the data center, we still have plenty of opportunities to be able to get into the account and grow our wallet share.
Once we've landed with an initial product in this multi-product motion, we built sales motions around each of those. You probably heard Kiran mention on the earnings call, our primary unit of measure, what we index on, is data. Once we land in an account, we'll likely see those accounts, given the growing size of data within the enterprise, we'll see those accounts organically grow within the workload that we sold to. Then we can also sell to additional workloads. Anneka mentioned that we could sell to the Oracle databases, we could bring on VMs, we could bring on RDS if we're in the cloud, as an example of data cloud services that we can protect, we can continue to expand workload coverage.
Then the third area is we can sell additional product categories. We could sell security products, identity resilience, as an example, to a data center customer. For us, this gives us a massive opportunity to not only land but also expand, continue to expand, then get to the point where we explode. I couldn't be more happy personally, our partners frankly couldn't be more happy with what we announced yesterday around flex licensing. Anneka gave you the case study, so I won't say it back again, our largest deal in the first half, or one of our largest deals, I should say, leveraged flex for agentic cyber resilience across multiple different workloads. We think there's a big opportunity here, we think this multi-product platform strategy gives us a real opportunity to continue to land, expand, and explode with our customers.
Maybe just to illustrate this with a few statistics. This past year alone, over 75% of our customers now are using multiple different products in multiple different product categories. Even our new customers now, 70% are landing with multiple product categories attached. This helps us also grow up our average deal size, which I mentioned earlier. To make this real, Alok, why don't you share two customer examples, since you're in the room, of what we mean by land and expand and explode?
Absolutely, Jesse. I have two customer examples. I'll start, of course, with a healthcare IT provider serving over 10 million citizens. Their initial problem was very classic. They essentially had multiple different tools, lot of complexity, ransomware exposure, and to top it all off, they had very high costs. We landed into this account with their on-prem data center. We essentially were protecting and securing their data for a subset of their data centers. We consolidated five different tools on the one Rubrik platform. We simplified their operations, and we were able to show significant TCO savings. Their entire ROI model completely changed. The second, very quickly thereafter, they went and added our DSPM software to secure their healthcare data.
They were extremely afraid, they were very concerned that if they have a breach, that their sensitive data, their healthcare data, their very important citizen data will get exfiltrated. I can tell you that as soon as we were deployed very quickly, we not only showed them, but they were certainly shocked with where their sensitive data lived and who had access to it, helped them remediate it, and of course, now we do it on an hourly and daily basis. I want to say the defining moment, and both of these examples I'm giving from examples where I was personally engaged along with the rest of the Rubrik team. I want to say that after this happened, the defining moment for us was when this particular company had to go through an audit based on the NIST framework.
Not only did they pass with flying colors, the auditors had zero, and I'll repeat that, the auditors had zero remediation requirements. For a public sector company, that's the difference between operational confidence and regulatory crisis. And I want to say personally for me, this was a very fun moment in the account because they completely standardized on Rubrik. They essentially made the decision, we want Rubrik everywhere. I should keep clicking this. That was the DSPM, the defining moment. What they did at this point in time was they standardized on Rubrik and displaced all of their legacy tools and put Rubrik in all of their data centers across the entire country. Big moment, huge expansion. They standardized. They started asking questions. The next question they started asking was, "What happens to our Active Directory in an attack? Who's protecting Microsoft 365 in the environment?
What happens if that gets compromised?" Both of those questions were answered by a single click. It was a single flip. I think Bipul mentioned it, a single flip of a button, and we just enabled the license. They had the capabilities. Within hours, they were up and running. I call that the platform in action. It's the platform effect. The customer has confidence in the technology, they have confidence in our account teams, they have confidence in Giri's support and post-sales teams, and they just go and standardize and keep adding more and more software footprint. Earlier this year, this particular customer also added Rubrik Agent Cloud to their environment. Their criteria was really very simple. They basically said, "We are rolling out agents.
How do we get a level of confidence that we have some control on the agentic actions and our strategy is not just hope?" The second was, they had seen all of these, the last six, nine, 12 months worth of agents going in, deleting a calendar, deleting a database, deleting someone's email, deleting code repos. I think that's more of recent news, at least from my memory. What they wanted was, how do we have some level of control, some level of guardrail for agentic actions? But if an agent does something destructive, how do I go and recover and remediate? This particular account has not done adopting Rubrik. They're already talking to us in terms of how they can add Rubrik for their PACS data and footprint.
For those of you who are familiar with PACS, that's a ton, petabytes and petabytes worth of capacity. They're not done yet. Big platform moment for us. We'll keep expanding here in this particular account. Healthcare. I think Jesse mentioned 2 of our top verticals. I would say growth from a growth perspective, healthcare, financial. The second example I'll give is, in the U.S., so in the Americas, financial institution, they came to us, and this was introduced to us by one of our top partners here in the U.S. The requirement was, we have a lot of data across multiple different data centers. This was a few years back. They essentially said, "We need better ransomware protection for our core applications." I want to say this was MVB before we coined the term MVB. They basically said they want immutability, indelibility, air gap.
I remember asking the person who is now our champion, I asked him, I said, "Why don't you just do that with your existing solution?" He said, "We did talk to our existing vendor, and they said their idea of immutability, indelibility, air gap is selling them more hardware, more software, more complexity, much higher cost, much higher cost of operation." They did a demo with us. They did a pretty extensive proof of value with us. We won that. We landed with $300,000 on their data center for a small core set of apps. Very quickly after that, they essentially said, "Microsoft 365 is a tier 2 application for us. None of our employees can work if they cannot collaborate with each other, email each other, use Teams, use SharePoint, things like that." They deployed our SaaS protection, our M365 product for 26,000 plus users.
I want to say at this point in time, again, similar motion. They have confidence in the technology. They love our sales team. They love their partner, the partner that introduced us into this account. They have a very, very long-standing trusted relationship with this particular partner. They decided at this point in time, we are going to roll out Rubrik for all of our data centers across all of the Americas. Big, massive expansion at this point in time. This is where, again, I want to say that the platform effect starts to take place. As they started bringing up new applications, they started their journey towards developing applications in the cloud, moving workloads, moving data to the cloud. They essentially said, "We have Rubrik everywhere else. Why don't we click this button and get cloud data protection, cloud data security?
How do we go and better protect our Entra ID with M365? How do we go and better protect our Okta footprint? How do we go and better protect our Active Directory on-prem footprint?" That's where they adopted cloud identity. I want to say the next purchase after this was they essentially said, "Hey, we have good immutability, indelibility, air gap. We have confidence that with Rubrik, our backups are going to survive a cyberattack and a ransomware attack. We are now afraid that we will not meet our commitments and our SLAs for cyber recovery." This was the moment where they upgraded from what we call the Rubrik Foundation Edition to Rubrik Enterprise Edition.
This is where you get all the capabilities in terms of very quickly identifying your blast radius, your sensitive data, threat hunting, threat monitoring, what we call the preemptive recovery engine. Preemptive recovery, the world cannot exist. You cannot wait now, no longer. MITRE has changed the world in the last three months. You can no longer wait for a process to start after an attack. You need to have the tools and capabilities that will help you find a clean state of data preemptively, and we launched this preemptive recovery engine, I want to say three years back, two, three years back. This particular customer upgraded their entire footprint from Foundation Edition to Enterprise Edition. I want to say with this customer, we stopped being our relationship as a customer and vendor. It became literally a three-way partnership between the customer, this particular partner, and Rubrik.
Now we have very regular calls. Arvind joins a lot of those calls. I'm very much a part of this as well, where we are strategically thinking about how do we protect this particular financial institution? How do we go and enable them to use this as a way, as a competitive advantage with their particular competition. Jesse, back over to you.
Great. Thank you, Alok. Thank you for putting us in those moments. I felt like I was sitting side by side with you back in that deal. In those deals, I should say. Let's hit on our second strategic priority. We'll move on here. This is really the evolution I mentioned around our AI-native selling motion. I recognize you guys likely in the room are hearing go-to-market teams talking about AI all the time. For Rubrik, I promise you, it's a little bit different for us. We are truly making the investment to try to embed and thread AI through every single touch point in our sales process and the buying journey.
That starts from demand gen, from meeting preparation, from call transcription and being able to gain insights from product feedback back to the product team into validation around how we're doing proof of values, proof of concepts, and winning the technical buyers over from how we create presentations to how we create proposals. Ultimately, how do we identify within our territories the best opportunities for cross-sell and up-sell to accelerate expansion opportunities? What is the right product at the right time, at the right moment to the right seller and the right buyer, I should say, to go land a new logo, given that we have a multi-product platform now to go land on. There's a massive opportunity within AI to really help simplify things for our sellers. We can measure this in multiple different ways.
When a seller at Rubrik walks in, from day one, we're making the investment where we give them a full AI stack. A couple of applications that we've built in-house, a few that we bought, but they also get access to Claude Cowork, they get access to Glean, some of them get access to Claude Code even. We pair them with eight purpose-built skills that every single seller gets from day one so that they can immediately get started analyzing their territory, conducting the research they need, building out their contact list, building out their account, sort of maps with their partners, identifying the right sales plays and the right messages to go start doing outreach immediately. The last thing I'll just highlight here is, for us, we also think about AI as engagement.
Every seller right now knows that AI is changing the way that they should go to market, and they're looking for organizations to step up and provide them with the enablement and the development and help them continue to grow. One thing I think we're doing that's pretty cool that I haven't heard any other organization doing right now is we actually have our first go-to-market hackathon in the second half of this year, which I'm sponsoring. We already have a ton of excitement, a ton of ideas being thrown around, and it's driving great engagement with the team. I'm really excited about this. We're still early on in this AI-native journey, we're making the investments in the right AI roles. We're making the investment in the tools. We're instrumenting, I believe, the right processes unique to Rubrik.
We're obviously making a pretty significant investment in the training and development for the field to make sure that we get value out of this, and we can then quantify it back to the business. Next slide, Alok. Our third strategic priority, maybe build out the first one, is, we had talked about distribution leverage and our reach. We'll start with our reach, where this is an area where we've made significant investments in that you may not be aware of. I don't know if we've talked about it publicly all that much, but we hired a top-tier transformation leader from Zscaler who's built out a dedicated team extending our reach and executive mindshare within our organizations even before deals exist. They help us foster our ongoing CXO-level relationships and the community there.
Practically, if you want to build it out, you can think of it as going through several reinforcing programs, starting off with the executive access and relationships, which is really our in-house C-level experiences. You might have seen some things on LinkedIn or some advertisements ranging from some CXO collectives that we've put on. We just did one in Rome. We just did one in New York City in Q1. We're doing things like global roundtables, women in technology, executive briefing programs, for example, and we're doing it uniquely to Rubrik. Just a quick stat here, this team alone, through these programs in Q1 alone, we were able to reach over 1,800 C-level and technical decision-makers. I believe Bipul mentioned in the earnings call that he personally met with over 140 of them just in Q1 alone.
I don't have the stats from last year, I can tell you it's a meaningful increase and a big impact. The second area that they're adding value is around the business value economics team, which is helping us build CFO-ready business cases around cyber resilience and helping us conduct cyber resilience assessments, which are really not product pitches per se. They're helping you understand and baseline where you sit from your cyber posture relative to some of the industry benchmarks out there. That's typically helping us early on in the sales cycle gain mind share higher up in the organizations. In any given quarter, we have somewhere around 40 of these active engagements running in the field with this team. The third area I'll highlight is Rubrik Zero Labs.
Rubrik Zero Labs is our own bespoke research lab, which provides independent intelligence and credibility. Our most recent research was able to reach over 14.2 million people, and the threat newsletter, which we rolled out and launched, is now serving over 67,000 subscribers. This is really connecting Rubrik with being a thought leader in the cyber resilience space and helping our customers understand what's happening broader out there in the environment. Lastly, they have a community programs which ranges from our customer advocates and practitioner programs that drive peer-to-peer credibility and foster deeper relationships with our best customers and our best executives that are truly betting on Rubrik for the future. This team, I'll double down on it. They provide tremendous influence for the field, and they provide influence over our pipeline. They help us get higher end accounts.
They help us gain mind share with the executives. It's a very unique program, and it's very difficult for our competitors to go out there and copy. This is a key area to help us extend our reach. Maybe shifting over to slide, next slide, Alok. We'll hit on the distribution leverage part. We talked about the third priority being distribution leverage and reach. We just covered reach, and I've talked a lot about our direct strategy. Now, Alok, in the room, who's on my team, could you just talk a little bit about our global partner strategy, which rolls up to you, and maybe highlight how we're going about it from an indirect go-to-market standpoint?
For sure. A big part of my indirect team is to create leverage for our direct sellers. That leverage comes from our partner ecosystem. That includes global systems integrators, our value-added resellers, our alliance partners, as well as our MSP partners as well. Now, Rubrik has always been, right from day one, has been partner first and channel friendly. Since I took over the global partner org role, a big part of my strategy has been how do we do a lot more focus? I want to say that the first part has been identifying partners that have the ability and willingness to grow and scale with us. The second has been how do we go and get top-down alignment from their executives?
This is alignment that they believe that cyber resilience, cyber recovery is a problem worth solving and a strategic bet worth making for their business. That's alignment number 1. Of course, importantly, alignment number 2 is that they believe that Rubrik is their chosen horse based on our superior tech as well as our ability to execute. Third really is, I want to see how we go from more transactional to very intentional and meaningful strategic partnerships. We are seeing this strategy play out where several of our partners are making cyber resilience, cyber recovery Rubrik part of their top 5, if not their top 3 strategic bets for their business this year. We have several GSIs that are truly committing to Rubrik and have created unique solution offerings powered by Rubrik.
Just yesterday, you may have seen this, we launched Project Hourglass, where we have 6 GSI partners that are creating solutions powered by Rubrik to provide agent resilience for Claude Code to their customers. With our strategic value-added resellers, we are creating very intentional growth plans. These growth plans include targets and not just booking targets. We're going 2 levels deep. We are creating pipe gen targets. We are creating activity targets. We are creating very specific execution plans, aligning incentives, aligning investments. I can tell you that based on our Q1 results, this cohort of partners, we are seeing accelerated growth from them and with them in a short span of the last 3 to 4 months. Talking about our MSP partners, last year, 1 of our top 10 deals happens to be a large MSP pre-buy commit that they did with us. We are seeing a trend.
We are seeing more and more customers that want to consume technology as managed services. Especially as some of the sovereign cloud requirements are coming up, we believe that this area is an important growth area for us. With our tech alliances, one of the things that I did, one of the big changes I made was I essentially said, "We need to go deeper." We've invested a lot more technical resources in this area, and we are driving a lot more deeper product integrations. I'll give you an example. We have deep product integration with CrowdStrike. We integrated the CrowdStrike ITDR product with our identity resilience product. The value that we offer to our joint customer here is they are now able to operationalize detection, correlation, along with recovery. Neither of Crowd or Rubrik can solve this end-to-end workflow on our own.
This joint product integration that we did, deep product integration that's unique to the two of us, we have existing customers that have gone and adopted this. That is the goal here. I'll say in summary, the goal here really is how do we take the strengths of the Rubrik platform, our multi-product platform, and combine it with the expertise that all of these different partner ecosystems provide, and how do we compound the value that our customers get from Rubrik and from our partners? It really is, I want to say, creating a very nice win-win-win scenario here. It's a win for our joint customers because they get a lot more value, one plus one equal to 11. It's a win for our partner ecosystem because they get another vector for growth. They can go and build a business powered by Rubrik.
It's a win for Rubrik because this particular motion is driving leverage. It's creating leverage for us as well. Jesse, back to you.
All right. Thanks, Alok. If we hit the next slide, that covers the three strategic priorities I just wanted to highlight at the beginning of the presentation. Our multi-product platform, our go-to-market native motion as-- selling motion, I should say, as well as our global partner reach, as well as distribution. In the last three minutes here, I'll just hit on a couple of additional growth levers that are really independent of each other. We talked about some of them, but scaling international growth markets, as you saw earlier, we're experiencing strong international growth. It's only 32% of our overall allocation, and we believe we're just scratching the surface in some countries, and we'll look to opportunistically invest where it makes sense. I mentioned several high-growth industry verticals that provide us an opportunity to evolve our segmentation as we continue to grow at the right time.
We believe even though we do very well in global public sector, it still provides us with a very large, durable opportunity where we're seeing very good traction, especially in state and local and in certain federal markets, yet we still have just low single-digit market share. Alok just touched upon our multi-channel distribution investment. We are just scratching the surface with some of the largest Global Systems Integrators that are out there, as well as the overall channel and our alliance partners. We'll continue to push the needle on how we go to market around this multi-platform, multi-product, multi-platform, as well as the AI-native sales motions. Last slide, and I'll just close with maybe why we win consistently and why we're winning at scale. It really comes down to these five things, right?
Cyber resilience, as you've heard, has become a board-level topic, and the mindset of assumed breach is no longer a concession. It's now a strategy. The answer is preemptive machine speed recovery. Rubrik's preemptive recovery engine was built for this moment. We also deliver this, as we've talked a lot about, through a true platform that manages your enterprise cloud, unstructured data, SaaS identity. Customers are not looking for cyber resilience to be a set of point solutions that they have to stitch together in their worst moment. The multi-product platform offering that we talked a lot about, where customers are able to get started and grow over time on the platform. Land, expand, and explode.
From a go-to-market standpoint, I joke with the team that I'm confident that we could actually switch products with our competition, and we could still beat them because of our playbooks, because of the talent we hired, and a relentless go-to-market sales culture, which starts with Bipul and permeates the entire organization. I promised Anneka that we won't actually switch products with the competition. We'll just go beat them with our products and our go-to-market, but we think we have a real advantage there. Underpinning all of this, of course, is we have an unwavering focus on serving our customers, and that shows up in our NPS, which is well above 80, as well as our industry-leading net retention, at 120.
With that, we took a little bit more than 20 minutes, but I hope it was valuable, and hopefully I get the opportunity to meet you all the next time that we get together. Thank you for the time.
Thank you.
Thank you. Okay. We are just going to take a handful of questions at this time. After that, we're going to take a very quick break for lunch. Grab your lunch, come back, and we will get started on the panel. I think to start, we'll wait for the microphones to get grabbed. You can go first. Thank you. Actually, if you don't mind, let's get the team up.
Okay. Is this thing on?
Yeah.
Yeah.
Great. Brad Zelnick, Deutsche Bank. Congrats on a great conference. I can think of no better reason to be out here in over 100 degrees in Las Vegas, Nevada. I think listening to the keynote yesterday, listening to you today, the message of the platform, appreciating how every layer builds on the others to unlock value of each other to meet this moment is very intuitive and really, for us at least, relative to others out there, resonates very well. I wanted to talk about Rubrik Annapurna. Bipul, I know you subscribe to maximal thinking and you're thinking very long term. We've been talking about Rubrik Annapurna for a couple of years now. I don't ask this to be critical that it's taking so long, but more to better understand what goes into it.
It would seem that for the company you're building decades to come, as you say, and in an era where code is becoming more and more of a commodity, help us understand why is it so difficult and why it will serve as a long-term, durable, competitive moat for Rubrik.
About two years ago, what occurred to us was we have all of the enterprise data, plus metadata, plus security and user context, and everybody for AI is looking for it. Why can't we make this data available to the enterprise users so that they can actually take advantage of the actual Rubrik data and use this as one of the data infrastructure platforms? To make this thing easier to adopt, we also built in a vector search so that we can actually supply vector embeddings directly out of Rubrik. The benefit that we realized of such a platform was we reflect the production permission because we have to restore the production permission. Every other platform that extracts data from production, whether it's data warehousing or lake, have to re-permission because they don't have the production permission natively captured.
We said that this is an amazing data, plus metadata, plus security model built into our platform. Let's make this available to our customers. When we started to have conversations with our customers, what we realized was that most of the folks who are actually doing AI-related work in the IT departments used to be SQL programmers or SQL data collectors, and they are comfortable with the SQL language, but they are not developers that are actually programming to a new set of APIs. We didn't have their mind share because they were SQL users. We kept discovering and talking to them to figure out how do we make this data available and for them to use. More and more, we realized that there is a certain set of platforms where they are standardizing on because of the SQL nature.
The other realization we had was that our unstructured data engine, which actually powers the Rubrik unstructured data platform, has a unique capability to classify and categorize data natively as bitmap in the system. We can take that classification categorization as a catalog and populate, for example, Databricks Unity Catalog, so that they don't have to load the data before they can analyze and figure out what data they want to use as part of their AI work. There is no dependency on the SQL language and programmability because the mind share is on Databricks, not directly on Rubrik. It was a great way for us to discover this particular part of the market where the needs were highest. There was no economical solution.
It solves for both the speed of AI because you are quickly getting the metadata and getting started, you can get the data. Economics as well because loading all the data and classifying data is very expensive before you figure out you want only 10% or 20% of that data, for example, X-rays or videos or what have you. This is a part of the market discovery process, and that's why every time we introduce new products, I'm always saying that we have to find product-market fit because our job is to not push the ball uphill all the time. Our job is to find out how our customers will get maximal value from our platform and how do we find that use case that unlocks it.
We are still early, but we feel like unstructured data gives that opportunity to help solve a mission-critical AI problem in a very unique way.
Okay. I'll get what he thinks of that talk timing. Okay. All right.
Fatima Boolani from Citi. Thank you so much for hosting us. Jesse, I'll direct this question to you. You were very particular in your commentary about how surgical you are and have been with respect to deploying AI and AI capabilities in the sales process and sales enablement. I was hoping you could give us some feedback or quantitative data points around the outputs you've seen from those investments, right? Anything related to rep tenure, rep ramp time, speed to expansion, anything additional you can share as it relates to how a lot of these investments that you're going to continue to make are manifesting in productivity gains and what the implications are for how you think about overall go-to-market capacity. Thank you.
Of course. Great question. Would love to talk about that. Look, as I mentioned, we're still in the beginning of the journey. One of the things we did when we first started was try to identify what are the core skills, what are the core use cases that every AE needs. Basically look at what does our best sellers do, and then how do we go replicate them and expand those skills across the field. What are the areas that if we could do that really well, we could impact. You just hit on a couple of them, right? We believe ramp times could potentially come down. I don't want to put a figure on it, but roughly let's call it somewhere in the 40%-60% range is our ramping to our ramp sellers.
If we can shorten that time to ramp, it's going to show up meaningfully in our productivity and our ability to deliver more. I think the second area that we're seeing some early traction on is AEs being able to work more deals. We call it deal throughput, but instead of maybe only being able to work on four opportunities in a given quarter, can we actually expand that? That gives us the opportunity to scale faster. The third area, and we actually rolled this out, and we're seeing early traction, is if you think about what a seller does before they even step into the first meeting. We built a product, a platform. It's not even deployed as a scale. We actually built our own platform. We call it COBI.
That's really helping do what was taking reps manually two to three days per account. We've condensed it down where they can get very surgical research and a very surgical point of view that they can then go engage with the channel or in a partner community, that they can then go take into their first meeting to show up with a perspective on where Rubrik can help that particular customer. It's basically aggregating a bunch of information and then giving them the ability to create messaging off of it and create targeted persona-based messaging, I should say, and then be able to pull together their point of view so that we're showing up and we're differentiating ourselves against other sellers that just show up with a generic pitch. Those are some of the areas.
I think the last area I'll just highlight is, people talk about it oftentimes as an efficiency or a productivity lever. It's also a revenue lever, where if we get this right, and we're in the process of building some things right now, we think it gives us a tremendous opportunity to go drive our cross-sell and our upsell motions across that platform by helping our sellers be able to land the right product, the right expansion play. We've got five or six sales motions I mentioned in sales plays, be able to land that at the right time to the right persona. If we get that right, we'll be able to start stacking those S-curves that people talked about earlier, even better. Does that answer your question?
If I may just add one piece to that. Specifically on the sales engineering side, we actually deployed a tool to help with RFPs. As you can imagine, sales engineers spend a lot of time answering RFPs, and some of these RFPs can go tens of pages. Just imagine you have hundreds of SEs with different levels of tenure, different levels of English language. Imagine being able to feed all of your previously filled RFPs into a GenAI tool, and now you feed a new RFP, and within seconds and minutes, this tool is giving you answers. It's taking into account all the information that we have. We also work with recency, because sometimes, the answer to a particular question might be no from eight years back, but we just released something a few days back. We have all of that capability built in.
Two things are happening here. One is, of course, efficiency gains. A particular SE, and now multiply this with hundreds of SEs, efficiency gains where they're able to go and respond to RFPs much faster versus going and looking at a lot of different sources of data. The second is also the quality of responses has gone up as well. Just giving you one particular use case that, I want to say one of our initial use cases on the sales engineering side, both Jesse and myself, many of us are very passionate about this particular area. Thanks for asking that question.
Okay. Well, thank you for the questions, and thank you, team. We're going to have additional Q&A time at the end of the day. At this point, we'll wrap it up here for the first half. We're going to take a quick five-minute break to grab lunch. Please come back. We'll do a working lunch, and we're going to have a great customer panel with our CMO, John Koo. We'll go into the partner panel with Alok. 10 minutes, we'll be back here. Thank you. For the webcast, we'll be offline until about 1:00 P.M. Pacific. Thank you.
Ladies and gentlemen, please take your seats. We would like to continue the program. Once you have grabbed a plate of food, please take your seats. We would like to resume the program.
Okay. Thank you, everyone. Hopefully, you got some lunch. Welcome back, everyone, take a seat. Right now, we're going to kick it off with an awesome customer panel. I will pass it over to John Koo, our Chief Marketing Officer. Thank you, John.
Thank you, Melissa. It's great to be here. Thank you all for coming to Investor Day. I've got a distinguished panel of folks that have experience with Rubrik, our customers. I'd like to allow them to introduce themselves. Saeed, why don't you kick us off?
Can you guys hear me okay? I'm Saeed. I'm the Chief Information Security Officer for a company called Symplr. They focus on healthcare technology.
I'm TJ Bondarenka. I head up the global IT infrastructure and operations team at Wesco.
I'm Nick Rose. I'm with Genesis Energy, and I am the VP of information technology.
Awesome. Thank you.
We haven't had lunch yet.
That's true.
You only get fed after you perform. Kidding. Let's actually get a little bit of background information from you guys. I think the audience would love to know how you got started with Rubrik, what made you initially start to rethink your data security strategy, and how that relationship's developed over time. Saeed, would you like to kick us off?
Sure. As I said, Symplr focuses on healthcare technology. We sell our own platform. We have 50 SaaS applications used by nine out of 10 hospitals in the U.S. Heavily regulated. We get audited by customers all the time, and if we go down, it will be Change Healthcare on steroid. For us to make sure that we are protecting our data and/or having access to our data in case of a breach is critical and was critical in our initiative to engage Rubrik. We started with Rubrik backup on-prem, and that relationship evolved to including Rubrik Cloud, Identity Resiliency, M365, DSPM, and enterprise security.
Yeah. Our journey, slightly different. If you rewind about six years ago, our company went through one of the largest acquisitions that we've ever done. Overnight, it effectively doubled the size of our organization. As a result, we had to do a tech stack rationalization. I'd say historically, we considered backup and recovery as more of a means to keep our operational systems running. Really, we had to start to think a little bit differently as a larger company about cyber resiliency as a strategy. When we were evaluating our backup and recovery technologies, there were some pretty vast differences in terms of the approach that was taken by the various vendors to how that data protection occurred, and then what that enabled us to be able to do, in terms of building out our cyber resilience program.
When we did that evaluation, we felt like Rubrik had the best set of capabilities, that's where we started rationalizing our tech stack and using it, in our on-prem environment across the board, we slowly
Thank you. Thank you. Thank you so much. Do you want me to introduce you or?
Yeah.
All right. Let's keep it going. My mission is to continue to increase the monetizable products on the Rubrik platform. That's my mission. It's also to mitigate the risk of the innovator's dilemma. Quick introduction. My name's Mike Tornincasa. I'm the Chief Business Officer. I joined this company, I joined Bipul and Arvind 12 years ago in the first year, six months before the first product was released. I've worn a bunch of hats since I've been here. First AE, first sales leader. I led the transition from backup and recovery to cyber resilience. The role that I have now is probably my most favorite and the most fun. I'm continually launching the new products. I'm doing perpetual zero to one or zero to maybe 100 million, it's really fun. I really enjoy it.
There's a lot of entrepreneurial energy still in the company from when I joined, to paint the picture for you. A couple of beanbags, two dozen employees, a ping pong table. That energy lives on within Rubrik, I think that's how we continue to stay on the edge of what's next. I have two missions for this next 15 minutes. I want to introduce you to the team that I lead, the lateral team, as Bipul calls it. We call it Rubrik X. I'll speak specifically to one of our current missions, which is the identity mission. Then I'll introduce Devvret, our GM of AI, to speak directly to Rubrik Agent Cloud.
We have a bunch of irons in the fire within my group, obviously those are two that are very exciting, I think that you'll be interested to learn more details about. Let's get right into it. Most companies at scale die because they're successful. They focus on the product that they have, they scale it, they optimize for it, they extract as much revenue from it, but they miss the adjacent opportunities or the next wave of innovation that ends up unraveling them. We built Rubrik X specifically to solve for that. Within Rubrik, we have a structural advantage in that we have two teams. We have a core team on the go-to-market side, led by Jesse and Alok, that drive and scale the mature products. Then we have a lateral team that stacks the S-curves. I lead that lateral team.
It's a structural solution to the innovator's dilemma. It's working. This team, by design, is small. It's a small entrepreneurial team with a ton of tenure. That ton of tenure gives us trust and ability to execute within the organization. We also have the ability to experiment and find product-market fit. The way that we do it, you may be interested to know, is that we put together a pod at the beginning with a point of view, representation from product, from engineering, and from go-to-market. We bring it to our customers. We define on the front end a power curve, a bookings trajectory. We have an idea at the beginning of what the potential of this product is. We hold this team accountable to that outcome. This is a focus team that's executing with performance orientation outside of our core.
If I had to sum it up, they define the win recipe, they show the path to the rest of the organization of the opportunity. Then we integrate it directly into the core team to scale or create a product-led sales organization that has their own quota that takes that product to scale. It's not by accident that we've been commercially successful with more than six new products over the last 10 years. It's a recognition that there is a recipe. It involves the right set of capabilities out of the gate, a vision and roadmap that aligns to where the market is going, one or two steps ahead of our customers, and the GTM tactics that allow us to win.
Every time that we've done this, we've had to face a new buyer, a new set of competitors, a new set of objections. We've had to figure it out. You can imagine when I join, nobody knows who Rubrik is. We have this whiz-bang data center backup thing. We're knocking on a bunch of doors, shaking a bunch of trees. We have success. The cloud is now where workloads are running. It's no longer test and dev. Production applications are running in the cloud. Rubrik's opportunity is to port our capability to the cloud. Our sellers don't know cloud people. We don't know what they want. We don't know what their priorities are. We had to focus on this specifically. It took us a couple of years to figure it out. We did.
We crossed $100 million in just under five years with the cloud product line. We did the same thing with SaaS. Microsoft 365 was becoming a critical workload. You heard that from some folks. We did this faster. It only took us three and a half years to cross $100 million. I'll note, these two products today are, I think, $400 million of Rubrik's total ARR. It's not just zero to 100. It's zero to 100, get the flywheel going, hand it to the core. The core scales. We did it again with the enterprise applications for data security. We had a couple of product ideas that would enhance cyber recovery. We put those into a bundle. We crossed $100 million in two and a half years. This product line in and of itself is north of $400 million for the company today, total ARR.
Maybe most exciting is identity. Most exciting because it's outside of that core data vector, it's a whole another vector. We've successfully transitioned into the CISO's budget. We have our hand in the CISO's cookie jar. In just over four quarters, we crossed $50 million in ARR. As a company, over the last five years, we've been growing at 50% CAGR. The products that come through Rubrik X have been growing at 100% CAGR. Today, the products that have come through Rubrik X are more than 40% of the company's ARR. I don't know of another company that at this scale has been able to organically stay on the leading edge of innovation, let alone build a brand-new product from scratch for a brand-new budgeted line item with identity. There's two missions that define what Rubrik X is building next, identity and AI.
I want to dive into identity with some detail. Then I'll introduce Dev to take AI. Identity is the number one attack vector. Yes, you guys have heard this a bunch of times. What does that mean? 90% of organizations have had an attack on identity or via identity in the last 12 months. More than 50% in the last 24 months have had Active Directory compromised. There's a huge opportunity here. We have $10 billion as the TAM. That's a bottoms-up number based solely on human identities. You guys are probably well aware there's an explosion of non-human identities, service accounts, agents. We didn't put that into the calculation. We think that there's a category here to define that is not dissimilar to the category that we already defined. We have data resilience, cyber resilience. Why not identity resilience? What could identity resilience mean?
Well, if it follows the same format, it's easy for our sellers to present, it's easier for our buyers to consume. For us on data, first was recovery, because if you can't recover, you're not in business. Then we build left into posture. What things can I do, as some of the panelists were just saying, to be proactive? The same concept applies to identity, so we're building the identity resilience solution. Fully complementary, network effect will be illustrated in the forthcoming slides, but also can land by itself as a lower friction door opener for our sales team, where maybe the customer's not ready to replace us in the data center or in the cloud. Let me speak to the problem statement here. The modern attacker has figured something out. Locking data causes more damage than stealing it.
Yes, theft is bad, breach is bad, if you can't run your business operations, it's not just painful, it's existential. We've been recovering data and applications for the last decade. We do it fast, we do it clean, we're trusted by our customers. What's new is the identity provider is being compromised. Even if we can quickly and cleanly restore data and applications, if the users can't authenticate to those applications, it's for nothing. This is the challenge. If you look at these elongated recovery times from these cyber attacks, the identity provider recovery is measured in days and weeks itself. It was a logical extension for us to focus on this because we were on the front lines of recovering our customers' data, and we were seeing that that was useless if they couldn't get to the applications. We built a solution for them.
This is the identity resilience solution that we built. We combined identity and data now in one platform and have complete cyber resilience messaging to our CIOs and CISO that is resonating exceedingly well. If you put that all together, this is the reason why this product is growing so quickly. It's a message that our core team can articulate, that our buyer inherently understands and feels very natural as part of the same platform. It's a familiar playbook for us. We start with recovery, we add posture, and we get results. Last year, a customer of ours that's classified as critical infrastructure suffered a cyber attack by a nation state that compromised their Active Directory environment. Rubrik was able to recover their IDP in 3 hours and 24 minutes.
Microsoft's own best practices, which is more than 100 steps, frequently requiring hand-holding from their consultants, is measured in days and weeks. This is the type of impact of this product. The problem is very painful, acute, well understood, and easily funded directly from the CISO's budget. We see an opportunity here, and we're doubling down. Transitioning from recovery to resilience is what we're doing right now. Roll forward, we announced yesterday quickly on this, very exciting to me personally and to the market. This is the idea that you might have an adversary in your environment masquerading as a trusted employee. You want to evict them, but you don't know how.
Rubrik can tell you who they are, where they are, rebuild Active Directory from before they got there, play all the sanctioned changes back so you don't lose the data that happened before they arrived to that day. Nothing in the market like this. The customers that we've messaged this to are super excited about it. It will only be available with the resilience SKU, which is an uplift SKU. CrowdStrike is super excited about it. They don't have an answer for their customers. Tons of upside in this. I want to just mention the partnership and complementary nature of the CrowdStrike relationship, the Microsoft relationship, and Okta. Okta is one of the trusted IDPs, and we built capability for them. Stacking S-curves. I think it's exciting that we've been able to do it over and over again with cloud and SaaS and security and now identity.
Even within identity, you start to see how these opportunities compound. We're stacking S-curves even within this one product line. Case in point, 2017, we launched our first Active Directory solution. Very basic. We included it with the core product, but we seeded the market with it. We got close to customers, and we learned. Last year, we launched the Active Directory Force Recovery product, which automated the restoration of complex environments, and we charged for it, and customers willingly paid. Now this year, we've launched the resilience product, which brings the posture and hardening capability and the roll forward capability to market. We see a very large opportunity. Here's what we didn't expect. Identity is making our entire platform stronger. A couple of stats. Last quarter, 52% of new M365 logos included the identity product, nearly doubling our ASP.
40% of all M365 deals included the identity uplift. Those deals had a higher win rate, 8% higher, over deals that didn't include identity. Every mission that this team has taken on had a different buyer, different set of competitors, different objections. We figured it out each time. Identities is just the latest example. AI is next on the conveyor belt. No one better to tell you why Rubrik for AI than our GM of AI, Dev Rishi. You get a welcome.
Thanks, Mike, for the introduction and setting the bar high. My name is Dev Rishi. I was one of the co-founders and CEO of Predibase, now I work as GM for AI. I started a company, in 2021, that was all about being a leading-edge AI infrastructure provider. Predibase was the core backbone that really, I think, deployed production small language models for a number of our trusted customers, ranging everywhere from smaller tech-forward digital natives to larger enterprises as well. What we really had as our core competency was an AI infrastructure platform that was optimized for organizations tuning models on their data and being able to deploy them efficiently and at high scale.
We joined forces with Rubrik via an acquisition at the end of last summer, and the key motivation was to be able to take our core platform that was literally accelerating AI, we helped organizations get faster models than they would've been able to otherwise, with Rubrik's core assets in data and identity to be able to secure and accelerate the world's AI transformation. When you join forces with a new company, as a founder, you actually sort of underwrite what your own impression is of what the parent company's mission is. As I think about Rubrik, I really like the slides that Bipul actually had a chance to be able to share earlier, which is that Rubrik is a company that's helped organizations be able to navigate, and enterprises navigate, different waves of transformation.
It started with the automation transformation, as we were making a shift from tape stacks or something a little bit more digital. The digital transformation obviously led into the rise of cloud, where we decided to launch the Rubrik Security Cloud to be able to target cyber as the number 1 use case that organizations needed to be able to be protected against from a risk standpoint. I think coming up now, we feel convicted that the next enterprise transformation width is in AI, and that AI potentially is operating at a speed that was faster than the other transformations, but also has a large amount of headroom in terms of the productivity gains that we anticipate. Each of the different transformations has brought in its own amplification of risk. Digital with cyber.
With the AI transformation, the risk that we see is with AI and AI agents, which we really define as models that get access to your internal IT infrastructure and have the potential to be able to do an order of magnitude more damage at a tenth of the time. It is no surprise that I think as organizations are actually thinking about their AI transformation, the number 1 blocker that we have seen towards their ability to deploy production AI is if they have a governance, guardrails, and then compliance solution in place. At Predibase, we had focused on a number of different solution vectors. We had focused on latency, costs. As soon as I joined forces with Rubrik, one of the things that I did was I spoke to over 200 customers in my first five months. The most consistent theme was, models are getting quite good.
The agent harnesses are actually not that bad themselves either. We have a lot of orchestration frameworks. Everybody is trying to sell me a platform to make it easier to build agents. The hard unsolved problem is actually how do I manage the risk around what my agentic deployments can do? It is unsurprising, I think, that this is the case. From our firsthand experience building and deploying agents at Rubrik, I think we have realized that the conventional security stack was not really built to be able to secure and govern agents. You heard in the panel earlier that agents have sprawled across platform. I have agents running on my endpoint, I have agents running in the cloud, managed platforms inside of SaaS applications and others. In my opinion, agents tend to be overprivileged by default, and they operate on LLMs that are non-deterministic.
Finally, they can access and do work at superhuman speed. We now have a real confounding of variables, where I am giving a model that is based on something probabilistic rather than deterministic access to my core IT production infrastructure because I need it to drive results, but it can actually operate outside of the guardrails that I might have set up, and it operates at superhuman speed. I think it is no surprise then that we see a number of these instances make headlines, whether it is coding agents that go rogue and delete production databases or email servers, or the number of 71 outages we have seen from leading hyperscalers all the way down to hallucinated refund policies. I think it becomes very difficult for an organization to feel comfortable in truly unlocking AI for their company today. What do organizations find themselves needing to be able to choose between?
I find that most enterprises we speak to are in a few different modes. The first is, well, we will do security through blocking. I know that my AI agent will never leak sensitive data from Salesforce or leak source code because I have never given it access to Salesforce or my GitHub. Ultimately, I think that this is a relatively secure mechanism, but it is going to kill the ROI of AI. Because agents need access to your internal enterprise infrastructure in order to be able to drive the productivity gains we are all looking to underwrite. The second thing that an organization is trying to do is we try and write static rules. We try and write a number of if/then conditions to be able to try and guardrail what an agent may or may not be able to do.
I spoke with one large leader of a financial services company in N.Y. who mentioned that their current solution towards agentic security and governance was a 1,000-person offshore team writing rules around every type of scenario that may go wrong. In our experience, you cannot secure and govern dynamic AI agents with static rules. At best, they capture a very small % of scenarios, and agents are notoriously good at circumventing these rules. The final thing that I think people oftentimes find some level of comfort in is going human-in-the-loop in the process. Human-in-the-loop allows us to be able to exercise some semantic judgment. I can understand the intent of what the AI agent is doing, and I can try and make a judgment on whether or not to allow or deny that action. The issue with human-in-the-loop is it's not 24 by 7 or scalable.
If agents are operating at machine speed, the way that we actually secure and govern those agents have to operate at machine speed as well. These sets of challenges, I think, have led to the fact that most enterprises that I speak to are currently and actively looking for solutions today. Very few feel like their existing stack has them covered as they start to roll out more dynamic agents, whether that's something like Copilot Studio or Claude Code or Codex. What we see is that organizations tell us that they're interested in a comprehensive agent platform that can provide them the key tenants that they need in order to feel comfortable that they're unleashing AI and not an undue amount of risk. Earlier this year, we launched the Rubrik Agent Cloud.
The Rubrik Agent Cloud is our platform to help organizations securely deploy AI agents at scale and feel comfortable to be able to integrate into the existing IT infrastructure they need to get the ROI that their boards are demanding. There's three key tenants towards the Rubrik Agent Cloud. The first thing that every organization needs is continuous monitoring and observability. You need to plug in to the agent runtimes where they exist and be able to scan that environment to show you what agents are out there, what kind of access do those agents have, and what's the risk profile for those agents. What I often describe monitoring and observability as this base layer of the cake. It's something that is decent to get started. It's a prerequisite, but it's not necessarily sufficient for what you want when it comes towards controls.
The next thing that we think about is dynamic runtime security with our own AI to secure and govern AI approach that we call SAGE. I'm going to dig into this in just a minute, but really the core intuition that we found was that in order to be able to secure and govern AI, we needed to use models and small language models, specifically ourselves, to be able to introspect every prompt, response, and tool call an agent might be making and ensure that it adheres with the organization's policies. Monitoring will give you the observability. SAGE will give you the runtime controls you need in order to make sure agents are staying on their guardrails.
Finally, Rubrik has always had a mentality of assume breach, which is to say that eventually something will go wrong, and we need to make sure that the business is resilient when this happens. In the AI world, we think about agent rewind as the core feature that enables AI resilience. The ability to undo a destructive action an agent can take by tying our observability with our core understanding of data protection and backup snapshots, and say, if an agent did take a destructive action, we can take the observed action and find the exact healthy snapshot that exists and allow you to do an easy orchestrated recovery from that source.
These core set of capabilities around monitoring, security with SAGE, and the ability to rewind those actions have been really instrumental in seeing Rubrik Agent Cloud's initial release and the success that we've been able to see internally. I think there's two things that really set Rubrik Agent Cloud apart from the other solutions that have started to make their way into the market. One of them is the ability to do agent rewind, and the second is our own agent for agentic security that we call SAGE. I mentioned SAGE earlier, and we think about SAGE as a Semantic AI Governance Engine. SAGE really does is it acts almost like the air traffic controller over all of your agentic traffic.
It hooks into your existing agent runtimes, whether that's a Claude Code session or a Microsoft Copilot Studio agent, and starts to introspect every prompt, every response, every tool call that those agents and AI systems are making. We run our own proprietary small language model judge over those instances to be able to adjudicate whether or not to allow or deny a given action. SAGE comes with its own AI security research built out of the box based on our experience doing cyber at Rubrik, and it's enriched with a data and identity context that Rubrik has for each of our enterprise customers. When I talk about data and identity context, what I'm thinking about is when it comes to data, where does sensitive data live inside of my organization?
When it comes towards identity, who should have access towards these different silos of data, and what are the indicators of compromise we might be thinking about? Ultimately, our view is that the organization that does the best job of securing and governing AI will use AI themselves. The companies that build the best AI will have the best infrastructure to be able to deploy those models and the best context for being able to feed into those models. The context that we feed in is the existing data and identity context inside of the organization, as well as the organization's own policies that they can define and customize directly from their own policy documents or through a natural language interface to customize SAGE to what they need inside of their company.
SAGE is built and deployed on top of the Predibase platform, which is instrumental because we need to be able to do this type of judgment at a very low latency and high accuracy way. What we have found with SAGE is that we can actually deploy our own custom runtime guardrails at a faster, lower latency, and maintain higher accuracy than a number of the leading frontier models that we've tested and evaluated. That small language model infrastructure, coupled with our context that we understand for the data and identity inside of an organization, helps reinforce this as a one-time security solution. Our customers have already seen this become valuable first party.
The Rubrik Agent Cloud is our newest product and offering, we've started to see some great success with a number of the customers we've deployed to, including, for example, a European healthcare IT provider that needed to be able to start to enable their AI initiatives, which involves building agents inside of Microsoft Copilot and also Copilot Studio, but had to have the protections for sensitive data and enforce compliance guardrails that they had as a regulated industry inside of EMEA. They viewed and acquired RAC as a solution because it was essentially their AI enabler. They could not go to production without a solution like the Rubrik Agent Cloud in place to make sure all their agent interactions were secured and governed. We're actually fortunate to have a lot of first-party data on this ourselves, where we've become a relatively large Claude Code shop.
You saw the Anthropic C sale on stage yesterday as well. We have seen Rubrik Agent Cloud be instrumental in securing and governing our own rollout of this technology. As a first-party user of Claude Code myself, I've got to tell you, the technology is incredible and quite scary in terms of what it can do. What we've been able to do with the Rubrik Agent Cloud is deploy and protect over 1,000 developers, multiple trillions of tokens processed in just a handful of months, and enforce that runtime to ensure everything from destructive actions to data exfiltration to circumventing of guardrails can be protected and prevented through the Rubrik Agent Cloud.
This has given us confidence that every organization that is starting to actually adopt the higher agency agents like these are going to need a solution like ours in order to be able to mitigate the downside risk. We've been pleased with the initial momentum that we've seen since we went GA at the end of February, most of our customers think about us as a solution that helps them get from a proof of concept that's heavily sandboxed in one environment and low risk, to an actual production deployment of AI. Because I'm a product person, what I'd love to be able to do is actually to spend a few minutes showing you what the product can look like in action, and how it looks to be able to secure and govern deployments like Claude Code for our customers.
This is the Rubrik Agent Cloud interface. As soon as I go in, I'll log in. The first thing that we see inside of the dashboard is we automatically scan the different environments where agents might be deployed, whether that's endpoints in the cloud or in managed platforms, and start to populate an agent inventory that shows you where are the alerts, where are the violations, and where are the highest risk agents that I might see inside of my organization. One of the things our customers told us is that they don't always know all the instances that they need to investigate, so we also prioritize insights about what kind of vulnerabilities they might have. Like this instance where Claude Code deleted a repository on GitHub. I can click in to investigate this insight further.
Once I do so, the Rubrik Agent Cloud actually populates what we call an agent map. This is a list of Claude Code, but also all of the different applications and tools that it has access to, like GitHub, my file system, and npm. It can even highlight to me where the specific problematic action happened. In this specific example, I have a developer who's running Claude Code, and we saw that what ended up happening was there was a destructive tool and a shell command. I can dig in and get a full audit trail of what this agent was actually doing up to the point that this action was taken. What I saw in this instance was Claude Code was trying to commit a credentials file, leak some credentials internally, but that was stopped by one of our runtime guardrails.
It moved on to a cleanup step and decided to do a gh repo delete, which essentially actually deletes the entire repository inside of my code base. These are the kind of elevated permissions that are actually not uncommon for coding agents to have because they use the delegated identities of the underlying user and developer. I can intersect and understand what was the command that was actually written here. I can see the traceability of the steps that led to this. When I see a destructive command like this, my immediate intuition is I want to do two things. The first is I want the ability to hit Ctrl + Z. I want to undo this action as quickly as possible and bring back the production system.
The second thing I want to be able to do is I want to make sure something like this never happens again. The Rubrik Agent Cloud can help with both. The first thing I did was I hit rewind. With rewind, we automatically come up with a recommended rewind plan, which will show you where to restore these actual assets from using your previous healthy snapshot from backup and our AI agent observability. You can always edit the rewind plan if you want, but in many cases, we actually find that the rewind plan automatically generated comes with the right level of intelligence, and you can execute it then directly in one click. My rewind was successfully completed, and that means my repository is now back online. The full commit history is there. I want to make sure this doesn't happen again.
Rubrik can also recommend a runtime policy to block all coding agents from running destructive shell commands on GitHub. This is an AI-generated policy that I can actually click and review, and we'll show you what this policy directly does. We'll show you its scope, what the goals and instructions are, key definitions. If I want, I can even see and ask for it to show me reference examples where this policy would trigger or not. I can edit this behavior if I'd like. I can see things like if the developer is running something like git reset --hard, which is more a destructive force command, then that would be a violation. If I was just checking the status, that's totally fine and safe. I can review key definitions, and this is something customers love.
They can actually edit the way that our runtime policy enforcements are going to work directly in natural language. They can provide their own context to the system about what they want to be able to enforce. When I click Save, I can publish my policy in either monitor mode, which will give me an alert or a violation anytime that it's actually triggered, or I can run it also directly in a blocking mode, where we give policies actual teeth to be able to stop the type of destructive actions at runtime. This is my instance of Claude Code. When I use Claude Code, for example, I might start typing in something like archive the old experimental repos because I want to be able to do a cleanup step. This is the type of thing where I was a little ambiguous in my intent.
Claude might start to go in and decide what it needs to do is delete a repository. Because I just configured that runtime active guardrail, it's actually blocked here by the Rubrik Agent Cloud. We have the runtime hooks to be able to secure and govern agents like Claude Code directly inside of their interfaces. We not only tell you what was the reason that we blocked it, we can tell you the policy, the severity, and give you the ability to edit these more in the future as well. That's the Rubrik Agent Cloud really in a nutshell. The way we think about it is a lack of governance, guardrails, compliance, and risk is the number one blocker towards enterprise AI adoption today.
In order to be able to actually securely and govern AI agents, you need to be able to use AI and small language models at runtime yourself, and that the organizations that are going to build the best version of these models will have a combination of performance small language model infrastructure and the best-in-class context to be able to drive it. I've been really happy at the progress that I think our team and organization has made in the deployment of Rubrik Agent Cloud, and I'm very excited to be able to see where this goes next as we help organizations secure their AI transformation. Thank you very much, everyone.
Okay, great. Well, thank you, Dev. Thank you, everyone. We're going to take a five-minute break, and then we'll go into Kiran's section, and then we'll do Q&A. Let's return at 1:45 P.M. Pacific, and we will finish up. Thank you, guys.
Ladies and gentlemen, please take your seats. Our session is about to resume. Please take your seats. Our session is about to resume. Thank you.
You want the solution that has the best context.
We'll get started here. Just give it a few seconds while people take their seats. Good afternoon, everybody. Nice job making it this far. This is the last session, and then we'll take a few minutes for Q&A. Thank you for your patience. I know we're running a few minutes late, but we should be done in the next 15-20 minutes with this session, and then maybe a little bit of Q&A to wrap up. Some of us will stick around to take further questions offline as well. Thank you for the investment and time and making the trip here for those in person, and hello, everybody online as well. I think I know most folks in the room and several people online, but by way of introduction, Kiran Choudary, CFO at Rubrik, been with the company for almost eight years.
Hopefully, the sessions you had in the morning for the last 3 to 3.5 hours have been very helpful in understanding our opportunity, thinking about through our go-to-market and product strategy, our overall TAM. I'll spend a few minutes here giving a financial update, but really my goal is to connect what you heard in the morning thus far to the financial model related as drivers for both durable growth as well as improving profitability towards our long-term model. I'll cover four areas. The first, we'll look back over the past few years, including our time as a public company, to talk through our track record of execution. Second, as you heard, we operate in three large markets: data, identity, and AI, and we are innovating relentlessly in it.
The end result of this is multiple product categories at scale, and I'll share some data and context around that. Next, we are a true multi-product platform, and when you combine it with our unique GTM approach-You heard about the forward motion with the core team as well as from Mike about the lateral motion with Rubrik X, and together they create a flywheel which really results in rapid customer acquisition and growth. I'll give you some data around that as well to give you some context. Finally, I'll wrap it up here with a little deeper look at the P&L. We are coming towards the end of a very successful cloud transformation and wanted to understand and explain to you a few trends around that, and then really talk a little bit of margin drivers and long-term model.
With that, we reported our ninth quarter as a public company last week, Q1 fiscal 2027, which finished end of April. These are the metrics we shared. Just as a reminder, we crossed $1.5 billion in subscription ARR, which is our core top-line metric, growing well at scale, over 30%. We continue to benefit from what we believe is best-in-class net new ARR of 120%, almost 3,000 customers at scale, which we define as 100K+, and strong gross margins as well as cash flow. If you look back in time or a few years, I've known many of you all from before we went public and including the time we were public, we have grown well at scale. We got to $1 billion at strong growth rate, now $1.5 billion at over 30%, and our guidance will imply strong growth ahead.
At the same time, we have improved operating leverage in the business as well. It's not growth at all costs. If you look at this chart here, which shows subscription ARR contribution margin, which is a good metric to understand our operating leverage as we go through the cloud transformation. We have made significant improvements over the years, turning positive in the year of the IPO and then continue to improve. Our guidance, the last column, will indicate that we have further improvement this year. Last but not the least, in terms of cash flow, again, if you look back at the history from where we started reporting these numbers, significant progress over the years, turning cash flow positive the year we went public and then guiding this year to almost $300 million in cash flow.
Hopefully, our nine quarters as a public company, and looking at our results, you've seen that we have grown consistently at scale, improved profitability, and raised guidance in each of the nine quarters. Hopefully that's built trust and you see a management team which can deliver. Moving on to the next topic I have is, you heard about earlier today the three large markets we operate in. It's not just sufficient being in large markets. You have to do more than that to build successful businesses in these. You have to allocate capital thoughtfully across these markets, have a rigorous prioritization process, and then you have to relentlessly execute both in terms of R&D as well as go-to-market to build meaningful businesses there. That's just what we have done. Going back in time, just to give you a little product history for Rubrik.
We started the business, really became operational in 2014. We started in the enterprise products area. That is where the market was at that time. We quickly progressed to cloud and SaaS products and then built a whole suite of data security applications around it, which really became our data pillar. Next, we realized the importance of identity information and the importance of connecting data with identity, and that really resulted in creating our identity business, which was the second pillar. More recently, we have made a foray into the AI market, with agentic cyber resilience. Together, these three markets you heard today are over $125 billion combined and growing nicely. Our strong execution capital allocation in these markets has resulted in really the creation of multiple product businesses at scale.
Out here in this chart, you can see how we have split up the $1.5 billion roughly in scale we delivered in the last reported quarter. Obviously, our history, we started an enterprise product, that's our largest business, $650 million plus. We launched cloud SaaS products. That's $250 million and $150 million respectively. Our data security applications, which sits on top of our enterprise cloud and SaaS products, is $400 million plus. Identity, our business which we started about a year or so back, is $50 million plus as we reported in Q1. AI is obviously in a much earlier stage. When you look at it, you see a very diversified business. These are just product categories. We have multiple products within these each. In enterprise, for example, you can think of VMware, SQL Server, Oracle, and others.
In cloud, it's the big three clouds, more recently Oracle as well. SaaS, M365, Jira, and Salesforce. Each of these product categories have a lot of depth within them as well, and then each of these product categories are now at scale. Now, this is looking back at the business we have built over the last 12 years, really 10 years of selling. If you look at the product mix as of the most recently reported quarter, and you break it up by net new ARR, 80% of our net new ARR approximately is outside the enterprise products area, which is really the data security applications, cloud, SaaS, and identity. Just to give you a sense of the most recent mix. Moving on now. Talked about markets, talked about innovation and our product businesses, but let's touch upon customer adoption.
What you have seen over the day, hopefully, is a true product platform. Which when you combine with the GTM machine we have, which is very unique in terms of how do you preserve innovation with Rubrik X and the core sales team to scale, you come up with really a very unique land and expand flywheel model in this market. Bipul touched upon this, I won't go into detail, it starts with one of the different product areas you can start out with, and data naturally grows. You add more workloads, you add more applications, security, AI, identity, this flywheel continues. What this really drives is a very attractive best-in-class 120% net NRR rate, which is really strong for a company which really sells subscription software and not really a consumption model. What has that resulted in? It leads to rapid customer acquisition at scale.
You can see our 100K+ ARR customer count, which we report quarterly. It's grown about a 40% CAGR over the years, and it's almost 3,000 as of Q1. We are still early. We are only in 40% or so of the Fortune 500, a little over a quarter of the Global 2000. There's a lot of opportunity to sell into these accounts with the product portfolio we have, plus gain new accounts. We are increasingly working with our partners as well as leveraging our direct sales team to open these accounts. Next, Jesse talked quite a bit about the diversity of our business. What we do truly matters, but also it's very broadly applicable, our platform. Any business or government entity or institution which has data of any significance can be a Rubrik customer. Hence, you see this mix of diversity by industry.
We talked about a few of them earlier. Financials is big for us, so is energy, healthcare, and a few others. We have verticalized in some of these areas because they require specific product capabilities, compliance, as well as go-to-market motions, and we'll continue to verticalize where necessary to drive further growth in these areas. Next, international has done really well for us, but still early. If you look at some of the bigger scale cybersecurity companies, close to half their business is international, a lot of room to grow for us there. Let me talk a little bit about product adoption. You heard earlier, I shared with you a few product stats, the different product categories we have at scale. Let's talk about how customers are adopting these products.
If you go back in time, fiscal 2021, about five years back in this data set, about two-thirds of our customers start with Rubrik, at that time, in 1 product category. Many of them were in enterprise, probably some cloud, as well as SaaS. The rest, the smaller one-third was multiple products. Let's progress forward to our last fiscal year, which we finished about five months back. You can see based on the bar chart on the right, over three-quarters of our business is multi-product in terms of adoption. Multi-product category, just to make that clear, because each of these categories are multiple applications and products. Close to half have 3 or more product categories, and over 20% has 4 or more product categories.
There's quite a bit of evolution over the past five years in terms of the depth of product, the breadth of the platform, which has been acquired by our customers successfully. This is for existing customers, but even in new customers, if you look at this stat, 70%+ of our new customers are landing in 2 or more product categories. What this is doing is increasing the ASPs at land. I think we had shared this stat with you at the time of the IPO, but right now, if you look at it as of this most recent quarter, the land ASP has more than doubled in about eight quarters or so. We are landing bigger. We're also acquiring more customers. Ultimately, all of these trends have resulted in growing our customers successfully with us.
We're literally raising the bar for customers to be a top Rubrik, top X customer. If you look at the bar chart on the left, that's the top 50 cohort. five years back, to be in that cohort, you had to be at $340,000 in spend. Today, it has $2.5 million+ in ARR, almost 8x growth. Similarly, you see growth in the 250 and 500 as well, eight to nine times bigger. This is the result of building a true product platform and having a very differentiated go-to-market motion. Hopefully, the last two sections have provided you with a sense of our overall market opportunity, product strategy, and the innovation we're doing on the customer journey front as well.
We are a true multi-product platform in three large markets, and our GTM approach with a combination of the core sales team, the forward motion, as well as Rubrik X, which is a lateral motion, is very unique in supporting this growth for years to come. With that background, let me now look to wrap up here. We'll go into the last topic I wanted to cover, which is a deeper look at the P&L. I'll start out to give you an update on where we are in the cloud transformation, which we started a few years back. If you'll recall, especially folks who have known us for some time, about three years back, we made a very strategic decision. At that time, it was five years in the making, but we launched Rubrik Security Cloud, which really provided our customers with true data portability.
It basically moved the brains of Rubrik into the cloud in what we call Rubrik Security Cloud, and it allows you to secure data wherever it sits. Within enterprises, in your private data centers, could be in SaaS applications, or could be in native cloud applications. All could be managed centrally from the cloud. At that time, with the launch of the SaaS product, we felt the right metric to start running the business by was subscription ARR, because it didn't have the impact which revenue would have from the transformation to cloud. As you know, ARR has grown nicely at scale, but so has cloud ARR. If you look at the second and third group of bars, you see the cloud ARR scale growing, and as a percentage of the subscription ARR mix, we've grown from about 60%+ to almost 90% as the most recent quarter.
If you'd asked us at the time of the IPO, I think we had shared a good long-term goal for us was to get to 80% cloud, because we felt there was a meaningful business. There'd be customers which are more regulated, would be more resistant to move to the cloud. We've been surprised by the uptake in this product. I believe we have a decent shot to get to 90% or so in the next 4-6 quarters, which is where it might taper off. We'll have a nice size business which is non-cloud focused on regulated industries and folks who want to kind of self-host. One of the questions which I got on the call last week was around cloud net new ARR. If you just calculate the cloud net new ARR as reported by our metrics, you'll see a deceleration trend.
This is what you see on the left. That's largely because the migrations are completing, because a year ago period had more migrations. We went ahead and we shared some data on the call as well, but this bar chart should clarify things for you, is we looked at cloud net new ARR without the impact of migrations, which is a much cleaner way to look at it. You can see here the cloud net new ARR growth rates have continued to be very strong. Another important P&L item which has impact on the cloud transformation is revenue. Revenue has grown at scale, but the growth rates have bounced around depending on the time of the cloud transformation. Early in the cloud transformation, revenue growth lagged subscription ARR, and towards the end of last year, it was ahead of the subscription ARR growth rates.
As we wrap up this year, and the last part is the guidance we provided last week, the revenue growth will show about 25%, but if you normalize for the cloud transformation, particularly material rights, that's about 30%, and the guided ARR was 27%. These are all converging, and going into next year, you'll see a greater alignment between subscription ARR and total revenue as reported, because we are really wrapping up the cloud transformation. Moving on to margins. We have improved our gross margins pretty significantly over the years. We exited the commodity hardware business. We have relentlessly managed the support costs as well as cloud hosting costs. We are actually right now above our long-term margin target of 75%-80%. I will remind you that we got some benefit from material rights through last year, which has also contributed to our gross margins.
Even if you remove that, we are above 80% today. One thing going forward to think through, and I'll touch upon this in the long-term model as well, we are still early in our AI journey, in terms of how our products are going to get adopted, optimization around tokens. We are still figuring out our pricing model as adoption of our AI products grow. I think that's one area where product mix is going to drive the margins over time. How much cloud, how much AI in the years to come. Let me move to operating margins. Because we went through a cloud transformation in the past three years, we have focused on a non-traditional metric, subscription ARR contribution margin, as a proxy to operating margin because revenue was still going through a transformation in our P&L.
This is a good proxy or leading indicator for operating margin over time. You can see here we have grown subscription ARR margin pretty significantly over the years. Operating margin, despite the revenue transformation, has also improved. If you look at the guidance for this year for subscription ARR margin and make some assumptions, it will imply around break-even for this year in terms of operating margin, even if you remove the impact of material rights. That's a milestone for us. It's really an inflection point. One thing we'll do as we wrap up this year and going to next year, we'll likely retire subscription contribution margin as a metric and start to guide an operating margin, as that'll be a more relevant indicator, more traditional in what you have seen as well.
There are a number of drivers here to talk through about what could improve and grow operating margin over time. You heard some of that earlier today, but I'll mention and reiterate them again. We continue to source talent with a high bar globally. Close to half our engineering talent is India, and they're best in class, as good or better in some cases than talent we had in the U.S., depending on the area. We similarly love our operation folks as well. We hire globally and try to optimize both talent as well as the cost leverage around it. We've talked about quite a bit earlier in earnings calls as well, but renewals is growing as a percentage of business, and it's a much more efficient cost structure, and that leads to as well more natural efficiency, if you will, in our P&L.
Our international business, we had to make significant investments over the years, we'll continue to do so, but having set up the foundation, it'll scale more efficiently going forward, that should give us leverage as well. We have been early adopters of AI in different parts of the company. We obviously build AI products, but we also are using AI aggressively in different areas of the company, starting with engineering and technical teams, but also in support and other operational teams. As you heard Jesse talk a little bit about what they are trying in the go-to-market motions, not just in efficiency, but also how to drive growth. That should give us leverage. Obviously, that requires an investment in the earlier period, but we are optimistic that over time, it'll provide us more efficiency.
Last but not the least, we made investments to get public as a company. We've been public for a little over two years now, those investments don't need to be at the same level going forward, G&A will become more efficient as well. All of these combined should give us more leverage on operating margin over time. Next, let me touch upon cash flow. Three things drive free cash flow in our business: scale, operating leverage, and billings duration. As you know, we are growing fast. Business is becoming bigger every quarter, so scale is increasing. Just talked to you about operating leverage improving year-over-year. Those should help drive cash flow greater. The one area I think I've talked about in the past as well, which creates some headwinds to us in terms of cash flow, is the billings mix.
As we sell more cloud and SaaS products and AI products, they tend to be of shorter billing durations. That is something we'll have to overcome. Despite that, we are growing cash flow dollars year-over-year, and you can see that our guidance implied $300 million for this year. We take profitability very seriously. We are obviously very focused on growth and will not compromise, but we do that in a very sensible fashion in terms of investments. Give you a little bit of history over the past few years. That was how our P&L looked the year before we went public. Obviously, margins were pretty negative. This was one question which was raised as we went public as well. People liked our growth, but it was a little unclear in terms of the path to profitability and how quickly we'll get there.
I'll just build out the slide here. You look at the next 2 columns. FY 2021, we went public. We turned profitable on the subscription ARR contribution margin, as well as free cash flow margins, the year we went public. One year or more ahead of schedule. This past year, FY 2026, we exceeded our gross margin targets from a long-term perspective. As I said earlier, our guidance for subscription contribution margin will imply break-even, break-even plus for operating margin this year. We've worked hard to build your trust here, but we are focused on making investments in a very disciplined manner without compromising on growth. I'll wrap up here with a look at our long-term model. The last update for this we had given at the time we went public a couple of years back.
I'll start out with gross margin at the top. Our previous outlook was 75%-80% couple of years back. Given where we are, if you look at the left 2 columns, we have progressed past that. We do want to give ourselves some room there. That's why the range is a little broader, 77%-82%. Really, the goal will be to be 80% and above, but it's still early in terms of the AI product adoption. We're going to learn more, both in terms of the cost structure, the pace of adoption, and our pricing models as well in the next couple of years. That's why we wanted to give ourselves more room. We're going to continue to lean into R&D pretty significantly. You saw the market opportunity. You saw our product footprint, the platform.
We're going to do it as efficiently as possible. We will invest and lean into R&D. Our go-to-market motion is very human driven, and obviously, we will use AI to optimize that in the coming years. The ASPs are higher. The product is complex. It gives a lot of benefits to our customers over time. It's very sticky, but it does require human interactions to sell the value, and that'll be in the mid-30s area. G&A, as I said earlier, a lot of room to optimize there now that we have been public for a couple of years. The initial investment is done, and that'll be in the 7%-8% area. All of this leads to 20%+ operating margins and cash flow 25%+.
As I talked about, our billings duration will continue to compress a bit, and that'll create some headwinds to cash flow, but we're confident we'll get past the 25% cash flow margin. Last but not the least, now that we are on the cusp of operating profitability on a non-GAAP basis, I think the next question will be how are we looking towards GAAP profitability? The key difference obviously is stock-based compensation. That's the big portion which bridges GAAP to non-GAAP operating margin. We are focused on bringing dilution down to the low single digits. Obviously, we're going to balance hiring and retaining a great talent, given what we are building.
At the same time, we'll be good stewards of investor capital, and we will work on bringing our dilution down to the low single digits, and that'll help us bring SBC as a % of revenue down, ultimately leading to GAAP profitability. I'll wrap up here. Just a reminder of the topics we covered. Hopefully, we have a proven track record of execution as the time we have known you, obviously since a public company as well, and we appreciate your trust in us. We have a lot of levers for growth, as you've heard today. In three large growing markets, we have a true multi-product flywheel, along with our unique go-to-market approach. We'll continue to be disciplined and allocate capital appropriately and thoughtfully across our different product pillars and go to market to drive durable growth and improving profitability over the years. Thank you.
Okay.
This is the real Q&A.
Thank you. Okay. We'll start here in the front.
Thanks. Saket Kalia at Barclays. Thanks to the team for hosting this session. Super educational all around. Since we just reported earnings last week, maybe we could just start with a financial question here for you, Kiran. I think the identity business has been super successful, right? We talked about that a little bit earlier. Clearly the fastest-growing S-curve. On the other side of that, we've had a couple questions just about net new ARR, subscription net new ARR, not just cloud, total subscription net new ARR. If you exclude identity, let's call that kind of core cyber resilience net new. What it felt like in Q1 was that maybe that core net new ARR growth slowed a little bit compared to prior quarters.
I was wondering if, Kiran, maybe if you can just touch on that, and whether that's the way that we should be thinking about this going forward. Does that make sense?
Yep. I can go. Thanks, Saket, and good to see you here in person. Thanks for making the trip. Yes, if you do some rough calculations and estimates on what identity was this quarter, we didn't break it out specifically, you will see that without that, the growth rate would be slower. As a reminder, we don't run the business with a quarterly net new ARR number. We run it on an annual basis. That's what we plan towards, that's what we invest towards, and quarters can be lumpy because it's an ARR-based model, and you could have transactions which could influence it. We also had a very strong Q1 last year, and as well the compares could be tougher. The product mix growth will vary quarter to quarter.
We have a lot of different products to sell, depending on the timing of transactions, how much the pipeline has developed. Certain transactions could close in a quarter, and certain product businesses could be bigger in terms of net new ARR. We should really look at it on an annual basis.
Go ahead, yeah.
Thank you. Eric Heath with KeyBanc. Maybe one for Kiran, one for Bipul and the broader team here. Kiran, I saw that the TAM that you guys had up on one of the earlier slides in the day had a 20% CAGR on the TAM opportunity. Curious to hear how you as a team are thinking about the long-term growth opportunity for the business relative to that TAM. Bipul and team, one thing that I thought was really interesting and a significant value add for customers is the idea of the agentic cyber resilience opportunity. As these AI attacks become faster, you need to recover faster, and the agentic opportunity just seems like it's a home run in terms of the value proposition. Curious how you think about the monetization of that agentic capabilities.
Thanks, Eric. I'll start out with how we are investing towards the TAM. I think capital allocation is one of the most important parts of my job, along with the rest of the management team. It's great to be in a position where we have multiple investment opportunities as a question of prioritizing and then executing towards it, rather than not having those opportunities. Every year when we start our annual planning process, we look at a three-year view. Obviously, the world is changing pretty significantly, but we are primarily investing really in the product portfolio and then the go-to-market. We've invested across our core products, but more importantly on our emerging products, and then in terms of capacity in the field, both in the core team as well as the Rubrik X team.
And one is making those investments and we have to execute towards it, but how can we deliver things faster? That's our big focus being over the past year and this year as well. Can we get products out in the market faster with the use of AI? Can we enable reps faster, make them more successful, reduce the ramp time, you heard that from Jesse as well, faster? Because that is ultimately what's going to drive faster growth and more durable growth.
Regarding agentic cyber resilience opportunities, we definitely see both Rubrik as an agent to respond to AI attacks, as well as you heard about Rubrik Agent Cloud solutions around Guardrails and Rewind. Definitely, we feel that there's a tremendous opportunity in front of us. In terms of monetization and how this pricing packaging will work, we are still working through it. One of the things that Mike and Dev have done as the Rubrik Agent Cloud, we are actually packaging as something that is indexed under token consumption. That's what we are kind of working through. Our idea is that we want to always match how customer pay for our solution based on what they do on their core application or core AI, so that it makes it easier for them to understand.
If you're paying in one way for their actual agent, how do we actually agentic resilience is priced that is indexed on that. There is a lot of discovery that is still in front of us. Obviously, we plan to continue to innovate, invest in innovation, and we need to monetize it. Go.
Questions.
Thanks. Joseph Gallo from Jefferies. Thank you for this. This has been a great day. I want to talk through the agentic sales motion. The product pitch and why data should win is very logical, but you're also selling to a team at the customer who has historically leaned on other cybersecurity vendors historically. How do you break through that? Does that add complexity to the sales process? What does that competitive landscape even look like? I imagine it's the Wild West today.
I'll start and then I'll let Dev embellish. It's definitely an exciting time. It does feel like the Wild West.
Over the last decade or so, Rubrik has developed a brand that is highly trusted, first within the IT organization as we escalated up to relevancy with the CIO, and lately on the back of cyber resilience and identities with the CISO. The bulk of the conversations now, specifically around coding agents, the prolific agent of the day, we find ourselves interacting with VPs of engineering and CTO. That's a very different persona as you've observed. I think the fact that we have a reputation that is strong with IT and security, and we're bringing a message to market that is both around acceleration and risk mitigation, we're having a very warm reception. Dev, anything that you would add?
The only thing that I would add is on the competitive landscape, we did try and understand what the other solutions look like in the market because we were rolling out agents ourselves last year in a heavier way. There were, I think while things in the monitoring and observability space do have more presence, I think the two places that we stand apart are, number 1, in the way that we do runtime security and governance. Where we think about that being differentiated is just taking an AI-driven approach for which you need SLM infrastructure through the Predibase acquisition. That we also think that the company that builds the best model-driven approach will have the best context. We think we have the richest data and identity context for that.
That's one core capability that I think we see ourselves kind of stand more unique on, and then the second is the ability to do agent rewind and provide a resilience message after an agent destructive action as well.
Out in the front. You there.
This is Shrenik Kothari from Baird. You have framed the platform as really complementary kind of network effects, driven which each additional product is adding to the value of the platform. The team talked about S-curves ramping and now you can stack them on top of each other. Just in light of all the initiatives around, of course, the agentic cloud, the Flex program, some of the product RGLAs. How should we think about the upside durability of the NRR, right? Which of course, you are best in class, but you have this sort of S-curve stacking and compounding. Then just broadly new versus expansion mix, how to think about that going forward? Because it looks like there's much broader penetration that's available for both new and penetration.
I'll let Kiran answer some of the more quantitative way. I'll say qualitatively, at our scale and as we are growing, NRR may not be the right representation because we are landing bigger and bigger deals. Because four or five years ago, as you saw the customer adoption, customer would give us a very small part of their footprint because we were still newer in their environment and will grow quickly into the other workloads. Now we are landing larger and larger deals, which impacts the expansion opportunities in those accounts. What we are focused on is how do we transform our customers with multiple products to a start rate? That's what is driving higher average deal size.
If you look at 70% of our customers landing more than two products, these are all indicators of how Rubrik is maturing and as any larger scale mature companies, dynamics would evolve over time. Anything to add?
Yeah, I'll just add that if you look at the components of the NRR, we continue to have a very strong gross retention rate, and that's very stable, has been for many years. That's really the foundation for NRR. On top of that, if you look at all the three vectors we've talked about in the past, whether it's data growth, workload growth, organic data growth, workload growth, or the additional applications we sell on that, security, identity, and now AI as well, they're all healthy levers for growth. We're also not a consumption-based pricing model per se. We have a little bit of it. It's more subscription oriented. If you look at that in that context, close to 120 is pretty high there. Unlike consumption models where it can spike up and down, we have much more stability there.
At the same time, we have a lot of room to grow and new. You heard about it today. We got more than 50% of the Fortune 500 left, 75% of the Global 2000, plus a lot of more product to sell at land time as well. The lands are becoming bigger, and that'll have a larger tension on the NRR. We obviously want to do both, and there'll be a tension with each other, but ultimately the goal is to keep growing.
Good afternoon. Matthew Martino from Goldman Sachs. Thanks a lot for doing this, guys. Bipul, I wanted to zoom out for a moment. If you look at Mythos, the model basically proved that the window between finding a vulnerability and exploiting it has collapsed dramatically, and it doesn't seem like any patching cycle can really keep up with that. What I'm trying to understand is what does that do to Rubrik's role? Does resilience start moving up the security budget, less the insurance policy you buy last, more of the first platform decisions a CISO makes? Does that increasingly put you in the conversation alongside some of these traditional infrastructure security platforms as opposed to more downstream of them?
If you look at Gartner's most recent advisory to the customers, they say that cyber resilience is a must. Customers would explore preemptive technology for patching and recovery. If you think about it, obviously, if you have a house, you have to close the windows and doors and set an alarm. But that's basic. But that's not your strategy. That's basic. My contention is that in the age of Mythos and AI frontier models, cyber resilience is the only cybersecurity strategy. Everything else is basic. What does cyber resiliency mean in a broader context? Ability to patch without human in the loop if it is a known issue, and ability to recover at AI speed. You cannot recover at AI speed if you don't have preemptive processes.
That's what Rubrik's preemptive recovery engine comes in, because we are doing the work in peacetime to be able to recover at AI speed in wartime. This is where the world is going. At the end of the day, the security used to be incidental things outside of operations, and recovery used to be the same thing. If attacks are going to be continuous because there is no detection and there is no prevention, then the recovery has to be continuous. At any time in your process, you should be able to recover at machine speed, AI speed, and keep going. Our belief is that cyber resilience is going to be not incidental but fundamental to your operations. This is the only cybersecurity strategy that has any meaning in the world in the age of Mythos.
Thank you, guys. Question for Mike and Kiran. First of all, great first Analyst Day. Congratulations. I really learned a lot, so thank you for that.
Thank you.
The question for you guys is, you talked about, obviously, AI transforming the pace at which you develop products. You've got a product suite you're selling between AI, data, and identity. You also have new ways to go to market, accelerated go to market. When you put those three things together, and there is effectively infinite white space ahead of you, make the case for why your growth should be in the twenties. Because it seems to me if you do some basic math on this, there's no reason for your overall business revenue scale to accelerate. That's the first question I have. Why do we think there's deceleration in the model? That's the first question. Second question is, Kiran, I saw your last slide on long-term margins. I'm going to challenge you on this because that seems to be the large generation software company, right?
80% margins, pop, pop, 25% free cash flow margins. If you, as Bipul just said a few minutes ago, you drink your own champagne in the company, your agent define your entire business model, paint the case for why Rubrik is not the case study for what the long-term unit economics of an agentic software company looks like, which probably results in north of 35% free cash flow margins. Those are the two questions for you guys.
Yeah. Let me start with the second one, and then Mike and I can do the first one. The way we look at long-term margins is it's not forever, right? I think about it as maybe 3 to 5 years out. We are definitely very much in investment mode towards driving that durable growth. We have multiple levels for investments, multiple areas we want to invest in, right? It's still very early in the AI age around gross margins. A lot of companies are figuring it out, right? Model cost, token cost. I think right now it's a badge of honor to use tokens, but quickly it is like, what are you producing with the tokens, right? Our pricing models also evolve. We are trying to be prudent there, give ourselves more room there. That's why the range you see in the gross margin.
Definitely over the next few years, we want to invest pretty significantly in R&D. The product we sell is complex, as you see, right? The depth of the product. The value has to be sold, today very much in humans. It's not a consumer application. You put it out there, and it'll just start scaling, right? You don't have that kind of transaction sizes that reach in the Fortune 500 customers spending $10 million plus unless you build the product depth and the breadth we have built, and that requires human-based selling. That's why it will require the sales and marketing expense, which we have outlined as well. We have two teams. Though the smaller team is the Rubrik X team, it still requires some overlay there. We're trying to be prudent there, giving ourselves some room.
Yes, I think there could be upside, obviously, in the years beyond that. What was the first question on this? Do you want to start it? I missed the key piece of it. I think it was talking about growth re-acceleration before.
Why would the same business model product be gross in the 20s when I can model out acceleration?
Oh, it should be growing faster. Yeah. I mean, our goal is, if you hear Bipul and Mike talk about it, right? It also requires the art of stacking curves, which we have done in the past, and we'll have to do it now at scale. That's obviously the challenge and opportunity for us. Again, I'll remind you that we are not a traditional consumption-based model that some quarter things just spike up. We're much more of a subscription SaaS business. It's more steady, but you won't see growth accelerating or days accelerating. You'll see it more stable. Yes, that's the opportunity and challenge for us is could we accelerate growth given the product footprint we have?
Okay, we're going to take one last question. Okay. Well, thank you. Does Giri want to go? I'm happy to donate my question. Okay.
Thank you for doing this. Fatima Boolani from Citi. Kiran, the question is for you. It's a little bit of a riff on the prior question. The Flex model, very interesting. I think many of us in the room have a lot of experience with companies going down this path. Just so you can make yourselves that much easier to do business with, that's great. How should we think about what you contemplated in Flex adoption in your guidance? How does that change the complexion of revenue, the complexion of billings? ARR probably is accepted, but does RPO become a more important metric? Is there an element of ramped accounting mechanics that we should consider? Again, just so we're well-positioned to think about how this might create some volatility in numbers, also the way you're thinking about forecasting and guidance.
Just as an adjunct to that question, is this motion going to exclusively be for new customers, or is this something you're going to start to talk to with some of your most enthusiastic existing customers who would naturally be predisposed to spending more with you anyway?
No, great question. We're very excited about it. I think this is the model which other cybersecurity companies and software companies have successfully launched. It is still early days. I did want to caveat that. We've done a couple of transactions, including the one which Jesse mentioned earlier, which is a large intellectual transaction we had in Q1. We have some assumptions for it in the guidance that's already assumed. We believe it'll be a ramp. It'll start out being pretty measured. It's almost the second half now through the year, and we can provide you a little bit more update as we head into next year. We do think it can become more significant over time. At least for this year, we think it'll be more selective in terms of the types of transactions, types of customers.
We have a minimum size as well for customers to avail that. We'll grow into that. I think the way you should think about from an ARR perspective, it'll be seen as TCV by the term length, and that's how both ARR and revenue will work.
Well, with that, I'd like to wrap everything up. Thank you so much for spending the day with us, and we'll be around for a little bit for additional questions. Thank you.
Thank you, everybody.