All right. Good morning. I'm Sanjit Singh. I cover the infrastructure software space on the Morgan Stanley Research Team. We are super thrilled to have the management team from MongoDB, CEO CJ Desai, and Chief Financial Officer Mike Berry. CJ and Mike, thank you for joining us. You guys reported earnings this week, so thank you for coming down and joining us at the TMT conference.

Thanks for having us.

So before we get into the conversation, for important disclosures, please see the Morgan Stanley Research Disclosure website at www.morganstanley.com forward slash research disclosures. Today is like a special day, like a great lineup, and I think you guys have a great story. And so I want to dive in and tackle the debates on the story around AI. We'll talk about the quarter, so a lot to get through. But maybe to start off, CJ, you know, investors are sort of getting back to basics in terms of thinking about these software companies, what the value they create for their customers. Maybe you can walk us through that. What problems does MongoDB solve for customers today, and how will MongoDB create value for customers going forward?

Sanjit, I have been in the software industry for a long time. And one of the things that I would say is I constantly speak to customers. On this question, why are you using us? If you are not using us, why not? And what value do you get out of it? Okay, very simple three questions, and you'll be amazed on the great answers you get from customers as they think it through. We are software, but we are infrastructure software. And I know you know that, but we are infrastructure software. and infrastructure software is extremely hard to create, as in to create, and then innovate on. And as you add R&D to it over a number of years, it becomes a massive mode that you cannot just cloud code your way to it. So when I am having conversations with customers specific to MongoDB, in my first 100 days, what is super encouraging is mission-critical applications at financial services, healthcare companies, insurance agencies, or government, whether it's an AI-native company or a SaaS company that is now pivoting with AI agents and using MongoDB. So MongoDB becomes this layer of infrastructure software where the data needs to reside. Then they need to be able to search on it. And all of that in a single platform with even vector search because there is a lot of unstructured data in the world. That is what customers want. And when they say value, when they look at other potential choices, because this market has existed for 60-plus years, they look at potential choices. We are fairly unique compared to like a first-party service from hyperscalers, which only works in that particular hyperscaler. And you saw recently another hyperscaler outage in the Middle East the day before yesterday. So resiliency, scale-out architecture, and having that peace of mind that my mission-critical data resides in it. And last thing I'll say is, one of the banks I spoke to in Europe, they told me one-third of payments for my 45 million citizens that happens every week comes out of MongoDB. So we are that data layer that is super important to store, performant as you run your business.

Yeah, I think that point around the global availability and the resiliency is an important one. You guys are in some sense the superset of all of the data centers across the major hyperscale. So if you want to run a distributed application anywhere in the world, Atlas is a great place to do it. I want to...

Can I just touch one last thing? You know, I was speaking to one of the AI native companies that is crushing it durably, that is crushing it durably, and they said, CJ, we want to expand even more workloads on Mongo. MongoDB. And I said, okay, everybody, every hyperscaler is trying to sell you. Everybody is trying to sell you. You are being very successful. Why? And the answer from the CTO is very simple, is that you are the only database that I can run in, he picked out two clouds, GCP and AWS side by side, without me having to worry about it. And that's the biggest advantage you have, that I can have a logical MongoDB across two physical clouds, and that is the reason we'll expand even more.

Let's talk about Q4 results fiscal year 27 guidance, especially given the stock reaction over the past couple of days. And so when you look at the numbers, a lot of great things to like here, 27% revenue growth. Atlas grew 29%. That was at one point down from 30% in Q3. Meanwhile, the non-Atlas business grew 20% with the ARR from that side of the equation, accelerating to 13% from 8%. Mike, one of the factors that you laid out that weighed on Atlas growth in Q4 was a large deal in which the customer bundled both Atlas and Enterprise Advanced. And so given that you called out that bundle deals are not a new phenomenon, Can you speak to how you get to that normalized growth rate of 30% on Atlas?

Sure, and thanks again for having us. So this is a little bit of housekeeping. So I apologize if I didn't do a good enough job explaining this. So what we wanted to do is we typically don't call out one-time items, but we wanted to help bridge you from Q3 to Q4 and also make sure and connect our comments, which is the business performed largely as expected. And we got a lot of feedback from several investors saying, hey, thanks for bridging it. I think there's some folks that are not following it. So at a very high level, we have a very large customer that's great. They use both Atlas and EA, but they bought them separately over their journey with Mongo. Great. In Q4, we did a new transaction with them, a multi-year transaction, and we had to combine them. It's one transaction. And the accounting rules say when you do that, now you need to go through and use all the 606 rules and allocate the revenue. When we did that, no change to the underlying terms or consumption, Atlas got less of an allocation in Q4 than it did in Q3. If you normalize for that, Q4 growth rounds up to 30%. Okay, full stop. That's that transaction. Three important things. It's discrete to Q4. We don't expect this to recur again. All this is fully baked into the 27 number, and this is not a trend, folks, so don't make more of it than it is. We wanted to bridge you from Q3 to Q4. We have a lot of great customers. We love them all. They're wonderful. This was a big enough change, Q3 to Q4, and it didn't connect with our comments, hence the additional disclosure.

Thanks for asking the question. Just a couple of follow-ups on that. So you mentioned that outside this large deal, basically what you saw basically came in line with the expectations. When we go back to Q2 and Q3 when we saw a bigger beat just to be bumped, was there anything in those quarters that drove that higher magnitude of outperformance from a one-time nature perspective? Was there anything unusual about what we saw in Q2 and Q3?

Thank you. Important note there. Q2 and Q3 came in better than expected. Atlas a little bit better than we thought, but the bigger beats in Q2 and Q3, as we talked about, were really related to EA. This is the multi-year deal impact and dynamic that we have. No other big one-time events. We have 60,000 customers. It's a $500 million business in Atlas. there's always puts and takes, but nothing to call out, and the bigger beats that we saw in Q2 and Q3 were directly related to EA.

One key year strength in Q4 was RPO. What drove the strength in RPO, and how did the growth in RPO compare to the growth in current RPO?

Yep, so thank you for that. So this is a number that we put in our Qs and Ks. Going forward, we will put it in the press release table so you get it when we do the call. At the end of Q4 last year, RPO was $748 million. These are commitments across all of our products. That was mostly EA, and about 59% of that we expected to recognize in the next year. At the end of this year, that number went to $1.47 billion, and we expect about 52% of that to be recognized in the next 12 months. So current RPO grew by about 74 percent. Very importantly, folks, commitments on consumption businesses, that's our estimate. That number will vary because they still have to consume the commits. But for us, it's a wonderful sign because it's big customers committing long-term to Mongo. And so that's a great thing. Don't assume that's all incremental. We always try to get incremental business, but it really gives us more visibility into this year.

So just on a current RPO basis, we saw about 70% growth in current arms.

74% growth, and it grew by $325 million. So that was great.

It's great to see. Let's talk a little bit about guidance, and then we'll talk, CJ and I will dive into the growth story at MongoDB. For next year, for fiscal year 27, you're targeting 17% growth at the midpoint. Atlas growth starting at 26%, ending the year at 21 to 23. The question for you, Mike, is what are the swing factors that could drive upside to the outlook? And here's the unfair part of the question because you weren't the CFO last year. But how would you compare the degree of conservatism, you know, for the current fiscal year 27 guide versus the same time last year when the company guided for fiscal year 26?

So this is more about the year? I just want to make sure the question is for the year, correct?

Thank you for that. So I wasn't here last year, so I'm implying a little bit. As your great question earlier, the performance in 26 we're super happy with. It was a very strong year. While we did a little bit better in Atlas, the majority of the beats were in EA. So as we look forward to this year, and we said it multiple times on the call, folks, we feel really good about the business. As we go into 27, we think there's multiple growth drivers. We always want to set up the guidance to make sure that we feel really good about achieving it. And we will always make sure there's more upside than downside, especially related to the EA business because that's tough to forecast. But we feel very good about Atlas going into 27. And to your question, multiple growth drivers. We have not baked in a lot of growth in AI, either through AI natives, but most importantly, through enterprises deploying AI at a wider scale. Not if, tough to forecast when. We've had a great move-up market, and that's driven a lot of growth in Atlas. We continue to put more resources against that. We have some of that baked in, but if that does better than expected, that should drive more growth as well. We're seeing a lot of migrations now as companies try to prepare for AI. That's another growth driver as well. And then, hey, from an economic perspective, we're assuming it's largely what it is. If rates come down, things get a little bit better, that's an upside as well. So multiple growth drivers on top of the guide.

Can I just add one thing? And I think this is a very important point. In the previous two companies I worked for, our customer base was not growing as much as it is growing in MongoDB. When you look at the new customer ads that happened in fiscal 26, the new customer ads grew by 60%, and now customer base is at $65,200. So how much does that cohort play out? But overall, it's a positive sign because to get a new customer that I've learned over my career is so hard. And for us to continue to add across the spectrum from small digital AI natives all the way to the top line, some of the AI native companies we didn't even know, and we find out, wow, they got multi-billion dollar valuation and they came via our product-led growth motion. So that specific base of 65,200 and 60% new customer growth that just happened, it will take time, but it is also how that cohort plays out is very important.

I mean, it's one of my favorite assets in the story is the new logo growth at MongoDB because it not only implies like how you guys are sort of executing on the field, but it speaks to the durability of the opportunity two, three, four, five years out. and it's something that we pay attention to. You guys are in the top tier in terms of customer-based growth, and so that's definitely fantastic. The only other context I was going to add is when the company guided to Atlas last year, I think it was in the low 20s. We ended up at 29 when fiscal year 26 was all said and done. So let's zoom back out and talk about the growth opportunity. Path to $5 billion in revenue, right? So, CJ, you bring over 25 years of experience and mostly help scale service now from a billion and a half to a $10 billion annualized revenue business. When you think about your $5 billion revenue aspiration, what roughly implies doubling the business, can you outline the strategy for doubling the business over the coming years?

Absolutely. So on first principles basis, do I see a path to $5 billion plus? The answer is 100% yes. That's why I joined the company. I did my own diligence to really understand the TAM, where we play, and why should we win? Because TAM is a TAM, but you have to earn your right to say this is the right architecture, and in this world in 2026, why should somebody put their workload on MongoDB? So one is, absolutely, I see the opportunity by speaking to hundreds of customers before joining to see that we can see a path to $5 billion plus. So that's at the highest level I absolutely believe that we can get there, $5 billion plus. Number two, I see that in two buckets. Bucket number one, of course, is Atlas. Super important. I mean, now it's at a $2 billion run rate, growing 29% when I look at the year. I couldn't do that math. And if we can sustain a commitment to the long-range model that Mike gave in September, is that 20%, he said 20% plus, he said 20% is the floor, 20% plus growing, you can do that math and you can see a path in a few years to get there from that 2 billion base to 4 plus. And then the question comes about EA, right? And that's why in my prepared remarks in speaking to certain customers, the trend that I'm speaking with some customers, not all, but regulated industries, companies, definitely public sector, where there is this almost with what's going on in AI and others, they do want to use on-prem, okay, and on-prem, on-prem, it's not just banks, but there are like other companies who are saying, hey, we need to hedge and we need the hybrid option. So EA, which is still 20% plus of our business, what should be the growth rate? How should the product strategy evolve? So we are investing in the product strategy this year, as in specific innovations like Vector Search and others in EA. And then how do I consider that as a business that still has durable growth assumption that adds to the Atlas 20% plus growth that Mike laid out? That's pretty much it. Now, in addition to these two, because people have always seen us as a single product company, that there's a database, whether it's EA or whether it's Atlas. Make no mistake, we are going to add product. And coming back to my previous employers, last two, that is what we did systematically, that expand the TAM, go to the adjacent buyer or maybe to the same buyer and have that platform play. So that is an additional thing. So we are currently investing in R&D, as Mike called out on the call, besides other investments that we are making. So that's how I think about, on a first principles basis, Yoatlas might give very clear guidance. How do I think about EA, which is still a big number, on that path to $5 billion? We are going to add additional products to log that path to $5 billion. And last but not the least, which might call doubt, we are not in Japan. We have to be in Japan. So this will be the investment year to be in Japan, investment year. So expanding the team and so on. And public sector, which has very specific custom workloads. We have to be in public sector in a meaningful way over time. So that's an investment here as well. So those are the two additional things. So you combine these together, and I can see a path. Of course, we have to execute. A lot of things have to happen correctly. But if I didn't have that opportunity, I would be coming and telling you, hey, I just don't see that, right? So that's all I think.

It's a perfect segue to my next question. It's a question kind of focused on, Tam. And if you sort of zoom out and look at what's happening in the broader ecosystem, you see some of your peers coming from the analytical data state, really trying to build, I would call, an agent stack. So they're trying to build, acquire OLTP databases. They're getting into observability. They want to get into orchestration, right, to sell customers an application stack, right? And that's kind of where MongoDB is played. When I look at what you guys have been, I feel like you guys have been more focused on just being that generational data platform. And so why does MongoDB seem to be following suit in terms of building your OEG agents, expanding to other layers of stack, and why is the focus on being a foundational data platform the right strategy for MongoDB?

One, fundamentally, I would say We always want to keep maintaining the main thing. This is architecture, even though it was created in 2007 by the founders, it is the right architecture for modern workloads, and AI being a platform shift, you do not want to take focus away from it. So we have to innovate on the core, is how I would simplify it. And then, make no mistake, as I work with our CTO and the product teams, we are figuring it out where else do we have a right to play and what else should we be building which we are building leveraging AI so that we can come at the right point throughout this year at our dot local conferences and say this is an additional thing that you should expect outside of the core database layer X, Y, and Z where MongoDB is going to play because customers want us to play there. So those are the announcements you will see, real concrete product announcement that we are playing in addition to the core database.

Yeah, so it looks like we've got lots to look forward on the product innovation side. CJ, you've been a pioneer when it comes to building and delivering AI products to market. And so when we think about embedding models, re-ranking models, vector search, hybrid search, ability to do transactional processing, where do you see the most enduring value being created when it comes to AI apps and applications?

So on AI apps, what we are already seeing with some of the AI-native companies, so I spoke to the founder of this wide coding platform, and people are coming in there. I mean, literally I said, can you give me the top five prompts or top ten prompts? What kind of apps people are building on your wide coding platform? and it varied from simple things like I'm trying to create an observability dashboard based on these data sources, something very simple that an internal team is using. Two people are putting things like I don't like my CRM vendor, give me a lead management software that I can build on my Vibe Coding Platform. So you have like a, he showed me like the actual prompts that get put in this Vibe Coding Platform. This Vibe Coding Platform, he told me, CJ, that I have very simple, one is fast APIs. He still doesn't believe in MCPs, right? So outside of LLM, fast APIs, and he thinks MCPs are not scaling for his business. Number two is React on the front end. And then M was MongoDB. So farm is basically said. And I said, why MongoDB? You built it. He came over our website, picked it. The company is doing really, really well globally. and he said because this can scale, whether it's images, PDFs, documents, search, vector search, and so on. These kind of examples are a frontier model company using us for some of the use cases, just show the power of our platform as in our database, and that's how I look at it when you think about it, that if these kind of leading-age companies that were created like a year ago or two years ago, they have not been around for that long, And if they choose us as an architecture, because I asked them the same questions, Sanjeev, you had a year ago, did they look at Postgres? Yes, they did, but Postgres didn't have multi-cloud resiliency. If you are a wide coding platform, that is your business. You want multi-cloud resiliency. Then number two, can it scale for unstructured data? And JSON native. And then he uses multi-tenant and things like that. So these are the kind of examples that gives me a lot of optimism that this is a great platform in the age of AI and multi-cloud, because multi-cloud is a tailwind for us. And every year, all hyperscalers had outages in 2025, and it's already 26. And you saw that what is currently happening in the Middle East with one of the hyperscalers and how it is impacting so many customers.

I had a question on Postgres a little bit later, but I might as well hit it now. You mentioned the global availability and the resiliency of MongoDB as an advantage. When it comes to other critiques on Postgres, and you guys have been clear that MongoDB doesn't have to lose. There's plenty of opportunity. But what other things would you call out? Is it the rigidity of the data schema? Is it scale-out versus scale-up architecture? What do you think that developers are going to look to MongoDB to be the destination for next-generation apps?

So I'll tell you the positive feedback, and I'll tell you what they expect from us. So on the positive side, and you've done Sanjit-covered infrastructure software for a long time, architecture that was built when Elliot and the team created this database, scale-out 100% matters, because they do not know how much they are going to grow, where they are going to grow, which region they are going to grow, and our availability in multiple regions and clouds and so on definitely helps. So scale out from an architecture perspective, native JSON, and how friendly. This one AI company I met, and I said, and they're a very successful AI company, why did you choose MongoDB? He said, my last two startups, I was very successful. I sold them. Now this one, I have scaled already. I'm not going to sell it. All three I have built on MongoDB. So there is also this developer-friendly thing that you hear about MongoDB, how easy it is, flexibility that as the business or use cases changes. Relational, as you know, is very rigid. And then the third thing I would say, it's all in a single data plane, including now embeddings. That is a huge advantage, because otherwise you need either bolt-on or you ETL for search performance when it comes to queries. And then the last thing I'll touch on, write performance, because you have so many writes happening. In addition to read performance, we are world-class.

When you look, historically you guys have disclosed approximately 30% of AR now comes from customers with at least one AI use case, and there's thousands of AI apps running on Atlas today. As those trend lines continue to improve, What needs to happen for MongoDB's growth to start to benefit from AI applications in a more material way?

I, you know, AI natives are helping us grow. AI natives, you know, you're fully aware that we did dot local in San Francisco less than six weeks ago on January 15th. And we are going to do it again in August just to keep that drumbeat on to do that. So we are top of mind for these companies. So that cohort, Sanjit, that cohort is important to us because that is contributing to our growth, but like Mike has said, not in a meaningful way when you look at the big base of 2 billion plus. Now you look at enterprises, and I speak to them all the time, including large banks or healthcare companies, and they have started creating agents, but these agents are not like at the scale of active users that you have in the consumer side of AI frontier model companies or on the AI-native companies. Then monthly active users are not that high, or a weekly active user. When that actually happens, we will benefit. When does that happen? I do not know, and that's why Mike said that, hey, that will be a tailwind, but we are not baking on it right now because speaking to customers, it is still early. They tell me, yeah, we have this productivity agent here. Of course, we are using the coding agents here, but that's pretty much it. Every retailer I speak to, including the larger, they talk about agent e-commerce, and I tell them, show me, and they're like, yeah, we have ways to go, but that is our strategy.

Let's bring Mike back into the conversation. Mike, yesterday last year, he spoke to the ability to keep investing while driving 100 to 200 basis points of annual margin expansion going forward. How are you driving efficiency across business operations, and what role is AI playing in driving those efficiencies?

Sure. So thanks for the question. As we look at 27, and we talked about it, Mongo has such a great business model that we are able to invest in the things we've talked about, largely driven by the growth in the business. So revenue will drive profit, but it allows us then to drive the innovation and also expand into new markets. So one of the nice things about Mongo is that, hey, we have a great customer base, but we have also not done as good a job, I'll say, expanding internationally. So a lot of the new headcount that you'll see at Mongo will be in lower-cost locations, especially across R&D and G&A. We will continue to invest in selected headcount for sales. On the AI side, that is another opportunity for us. While we've looked at it, we're like most enterprises, we've not deployed it at scale. So a little bit of coding, certainly in marketing, there are some good areas in G&A as well, but that's also an opportunity. So while we will add headcount in fiscal 27, the goal is, hey, going into fiscal 28, that headcount should be relatively consistent, and then we'll see if we can drive further efficiencies after that.

You touched on an important point that I think the market's sort of debating, given some of the announcements in the recent weeks. To what extent is MongoDB's ability to grow in the future at an attractive rate, how much of that is tied to headcount growth, and do you see these two forces decoupling as we progress into the AI era?

Yeah, it's a great question, and you read about all the other reductions, and that's certainly not what we're doing. So for us, there is certainly reliance on sales headcount. Quota matters. Sales capacity matters. And then in R&D, probably more importantly, all the innovation that we need, we do need headcount to do that. Now, over time, that's going to get more and more efficient, and obviously with the coding tools that folks have. So you'll see that increase in 27 because we have a little bit of catch-up to do. But going into 28, all those things should allow us to be a lot more productive. I'll give just a huge shout-out to the sales team. They had a really good year, and they drove their productivity up in 26. The rest of us need to do the same thing as well. AI tools will help that. So we don't need to continue to hire at the rate we are at this point, and we will become more efficient as we move forward.

My next question for you, CJ, was around partnering with the ecosystem, just partnering with Anthropic. The company announced that partnership with Anthropic, and you also announced partnerships with other ecosystem partners, including Langchain and Vercel. What will the Anthropic partnership mean for MongoDB's customers, and how will that accelerate your traction with AI workloads?

I think of this in two buckets. So one, when frontier model companies use us, it is a validation of our architecture because they could have said totally themselves because they are in that business, coded this ourselves, but they said, no, we are going to use MongoDB. So on the first principles, that is a huge validation that when they are building version 4.x or whatever it is, and they are doing that on MongoDB, it's a huge validation of our theory because they could have easily said, we don't need a data layer and we are going to code it ourselves. Okay, so that's one. Second is, so them being our great customers, frontier model companies, it's always a good thing. And as they grow, we want to grow with them based on the use cases that they use us for. Second thing is, I had a conversation with one of the frontier model companies, and I said just recently that now the need for modernization, I would argue, is even higher in a Fortune 500 Global 2000 because of AI. So I said, hey, we want them modernization. We are the most modern database. Is there something we can partner with that if we provide you all the things we have learned as customers have modernized on MongoDB new workloads. Can we work together so that you just in the model help people get to MongoDB as a destination that we can go to the bank and so on? So we are in very Sanjit early stages for those kind of things. But that is the second area of partnership that I think that where can they help us if the destination is MongoDB, which will be Atlas and will drive Atlas consumption over time. all this technical debt that large companies have, what role they can play to partner with us, because we, from a future perspective, want those workloads on Atlas or EA, because depending on the company and where they want to run it. And so if they are going to help us and partner with us, what can we tell them about what we have learned so that they can do that while we add our own IP, which we are in the process of, to get to the destination of something running with the right architecture and right performance on MongoDB. So these are the kind of extremely strategic conversation. And third thing that I underscored in the earnings call, which is really, really important, when I understand, even after speaking to the founders and speaking to customers who love MongoDB, there's a huge fan base that I'm really proud of, is the reason the leaf exists is that, hey, this is a very natural with the database to create a database for that human developer. And I want machines as an agent who are creating apps, whether it's a wipe coding app, this is your question around some of the partners that you talked about. I want that the same thing, that how easy for the machine to say, oh, this use case, it is perfect for MongoDB, and I'm going to spin up an atlas, and we are scale-out, so don't worry, we are not going to over-provision it, and it will autonomously scale for you, perform for you. That's the hallmark in the next few years from an innovation perspective. Human developers love MongoDB. Machine developers should also love MongoDB, which requires a bunch of work to be done on the platform side for us, and then right partnerships to be had. That's why I hired Erica.

That's a very compelling vision. CJ and Mike, thank you for giving us an update on the MongoDB story, and best of luck if it's clear to 27.

Thank you.