Confluent at Morgan Stanley Conference: Real-Time Streaming Focus

On Wednesday, 05 March 2025, Confluent Inc. (NASDAQ: CFLT) participated in the Morgan Stanley Technology, Media & Telecom Conference, where CEO Jay Krebs and CFO Rohan Srivium shared insights into the company’s strategic direction. The call highlighted Confluent’s growth in subscription and cloud revenues, its strategic partnerships, and its vision to become the central nervous system for real-time data. While the company showcased strong financial results, it also acknowledged the challenges of adapting to a tighter economic environment.

Key Takeaways

• Confluent’s Q4 subscription revenue grew by 24%, with cloud business expanding 38% to over $500 million.
• Strategic partnerships, especially with Databricks, are central to Confluent’s growth strategy.
• The company is focusing on converting Kafka users to its platform and expanding its Data Streaming Platform (DSP).
• Confluent aims to make streaming more accessible and cost-effective, emphasizing AI integration.
• Management is optimistic about future growth, projecting low 20s growth for 2025.

Financial Results

Confluent reported a strong financial performance in Q4:

• Subscription revenue increased by 24%.
• Cloud business expanded by 38%, achieving a run rate exceeding $500 million.
• Operating margin was over 5%, with a free cash flow margin of 11%.
• The company improved margins by about 40 points since its IPO.
• Guidance for 2025 indicates growth in the low 20s, driven by streaming opportunities and AI integration.

Operational Updates

Confluent is actively working on several operational fronts:

• Over 150,000 organizations use open source Kafka, but less than 5% are monetized.
• The paid customer count is just under 6,000.
• The company is testing demand elasticity by reducing storage costs and introducing new offerings like WarpStream.
• Strategic partnership with Databricks focuses on real-time data acquisition and open data formats.

Future Outlook

Looking ahead, Confluent is positioning itself for continued growth:

• The company sees a significant opportunity in streaming, with plans to increase the DSP mix in its cloud business.
• AI adoption is expected to play a significant role as it becomes mainstream.
• Confluent is committed to enhancing its go-to-market strategy, platform selling, and partner ecosystem.
• Investment will balance current products with future bets, ensuring sustainable growth.

Q&A Highlights

During the Q&A session, several key points were discussed:

• Confluent targets global 2,000 companies with significant IT spending.
• Efforts are underway to reduce streaming costs, unlocking more use cases.
• The company is positioning itself as a long-term beneficiary of GenAI.
• Confluent differentiates itself from competitors by focusing on real-time data as a central nervous system.

Confluent’s leadership expressed confidence in its strategic direction and growth potential. For a deeper dive into the conference call details, please refer to the full transcript below.

Full transcript - Morgan Stanley Technology, Media & Telecom Conference:

Sanjit Singh, Infrastructure Software Analyst, Morgan Stanley Research Team: All right. Welcome to day three at the Morgan Stanley TMT Conference. I’m Sanjit Singh.

I do the infrastructure software on the Morgan Stanley Research team. Super happy to have the management team from Confluent. We have CEO, Jay Krebs and Chief Finance Officer, Rohan Srivium. Thank you, Rohan and Jay for joining us once again this year.

Jay Krebs, CEO, Confluent: Really excited to be here.

Sanjit Singh, Infrastructure Software Analyst, Morgan Stanley Research Team: And I have to say, this is I think we should look at this conversation as kind of an appetizer for tomorrow. They’re going to have an Analyst Day. It’s been, I think, more than a year since they’ve had Analyst Day. So this is kind of our appetizer for the for a very interesting Analyst Day headed tomorrow, which I’m excited about. Before we get into the conversation, let me go through the disclosures.

For important disclosures, please see the Morgan Stanley Research Disclosure website at www.morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representative. So Jane, maybe I’ll just kick off on just maybe taking a step back. I think it’s been about three years since we’ve been in this kind of higher rate, more sober, more tighter budget environment. Can you give us the storyline of how Confluent has changed to operate in this environment?

And how do you feel like Confluent’s positioned us for success given this environment that’s quite different than the one you sort of IPO ed out of?

Jay Krebs, CEO, Confluent: Yes. Yes, it was a big change. Some of the things I think we were well positioned for, some we adjusted. The it starts with us with having a really good TCO story for our cloud offering. Ultimately, why is it that working with Confluent is going to be better than trying to do it myself with open source, any of the other alternatives, solutions from competitors.

We want to make sure that that’s rock solid and that kind of begins with the product offering that we have. The ability to run multi tenant to really drive efficiencies of scale that we can kind of take back to customers, that’s really been important for us and continues to be a theme. But it doesn’t stop just with product functionality. It’s also how we convey that to customers, how in our sales process we make them aware at different scales, how we can help them understand what are some of the older technologies that they’re getting rid of as they’re bringing in new things, all of that matters. And then in our own operations, making sure that we’re efficient, right?

Some of the adjustments we made were certainly on driving efficiency in the growth engine and making sure that we’re being very thoughtful about where we’re putting the incremental dollars to increase the top line, being very thoughtful and focused in our R and D investments. But at the same time, not giving up on the big picture. For Confluent, we think that there’s a move to a major data platform around streaming. And we want to be in a position to capture that. And so on the R and D side, we never step back.

That. And so on the R and D side, we never stepped back from that kind of larger vision of what we’re trying to do, the investments in stream processing and the other kind of larger data streaming platform functionality that we’ve been building. We kept that going even through tighter time going even through tighter times. Yes. And you see

Sanjit Singh, Infrastructure Software Analyst, Morgan Stanley Research Team: in the financial results like sustaining, very attractive growth. The cloud business is growing robustly. And in that time, I think you guys have improved margins about 40 points since the IPO, which speaks to that some of the things that you were talking about. Roeland, let’s maybe bring us up to speed on in terms of where Confluent adds from a financial perspective coming out of your Q4. You guided to the low 20s for 2025.

Frankly, one of the few companies in our coverage that we’re able to guide that well out of the gate. Can you shed some light on how you sort of constructed the guidance, kind of the fundamental assumptions in terms of your 2025 outlook?

Rohan Srivium, Chief Finance Officer, Confluent: Yes. We’ll be happy to. Let me start by taking a step back and talking about our Q4 performance. We had a strong Q4. Our subscription revenue grew 24%.

Our cloud business grew 38% at a run rate of over $500,000,000 And as we did that to Jay’s point earlier, we showed really good operating leverage, operating margin of 5% plus and free cash flow margins of 11%. And what underpinned that performance was essentially stabilization in consumption patterns that we saw for some of our larger customers. And we did see coupled with the momentum in DSP that we saw. So that’s a little bit of a backdrop for our Q4 performance. For our guidance for fiscal year twenty twenty five guidance is guidance that I shared at the time of our call a few weeks back.

And what I’ll tell you is from an overall philosophy perspective, it’s not changed a whole lot. It’s fairly consistent over the last couple of years. But what I can share is some of the puts and takes as I think about the next twelve months from a growth perspective. I’ll put it in three categories. The first is all around streaming.

We have this really large opportunity in streaming with over 150,000 organizations using open source Kafka. And we’re monetizing less than 5%. So big opportunity. And we are doing a bunch of things around it, around pricing, packaging, how we are going to market, etcetera. The second piece is our data streaming platform.

We mentioned we exited Q4 with roughly 13 mix for data streaming platform as a percentage of cloud business. So that’s our starting point. And we have multiple products that are in earlier stages of their growth curves. So we expect that mix to go up as we go through the year. And the third piece is around AI.

We feel that as mainstream adoption of AI happens, we have a very important part in that ecosystem. So when you summarize it all, it’s always good to have multiple drivers as you think about your guidance. And different drivers have different contributions. It’s more of a portfolio. So that’s how I think about it.

Sanjit Singh, Infrastructure Software Analyst, Morgan Stanley Research Team: Yes, that’s a great way to frame it out. To pick up on some of those themes and, Jay, when you think about the core streaming opportunity, I love the phrase that you sort of characterize the opportunity as soaking up the world’s Kafka. And as we sort of stand, we have north of 150,000 organizations using Kafka today. Your paid customer account sits at around just under 6,000. Can you describe what seems to be a multi pronged strategy to convert Kafka customers to Confluent?

Jay Krebs, CEO, Confluent: Yes. One of our goals when we think about the shape of the business, we want to make sure that we’re both building breadth, that kind of broad base of customers in-depth, able to take customers to really large scale with the offering. And I think we’ve made a lot of progress on both. When we look at the opportunity ahead of us, of course, we’re always like, well, there’s a lot more to do. And open source Kafka continues to spread, so that’s a good thing.

But entering this last few years, the focus has definitely been been being able to have the right offering and a low friction motion to pick up as many of these Kafka users early on. That starts with some of the work we did on consumption in our go to market, really orienting our field team around customers’ consumption. This meant that we could start with customers without necessarily a massive upfront commitment. And often, especially in tighter times, it’s just easier to get that first project done as its own project rather than trying to envision everything they’re going to do over the next few years. The next thing is having the right offerings.

And so if you looked at some of the product announcements we’ve had, it’s really how can we support the full spectrum of usage that companies have and make sure that we’re doing that in a way that’s better as a product experience, more complete in what we’re bringing, but also cheaper, right? TCO positive, something that’s better than the open source, do it yourself, competitor offerings. And that comes down to these kind of fundamental operational characteristics. Can we pool customers better in our multi tenant offerings? Can we engineer better for the ingredients in the cloud, make that more efficient?

A lot of science goes into making that true. And some of the results for customers are actually really impressive. They get an offering that’s more elastic, that just scales up and down transparently as they use it. They get something that has really good cost characteristics. They get something that’s just easier to use and plan around.

And we’ve introduced some new cluster types that open up parts of the market that were maybe stickier before. So some of the very high volume cost sensitive workloads, we offer freight clusters and we added through an acquisition WarpStream, which is something that runs in the customer’s account. Both of those really open up that segment of usage, which we think is important. We’ve done other work on the getting started, that first use case, how can we make that better. And that certainly added fuel to the fire.

So if you look at the customer adds last year, certainly a step up from the year before. We want to continue that progress. This is a really important thing for us. Now when we think about it, it’s not just adding any customer. It is about getting into the right customers and being very targeted in the high propensity accounts.

And so that’s been the other dimension of this. It’s probably a little bit less visible. But are we landing in the accounts that we think over a number of years can be really large in their usage? And I think we’ve gotten better at that kind of targeting over the course

Sanjit Singh, Infrastructure Software Analyst, Morgan Stanley Research Team: of the year as well. Maybe just to follow-up on that last point that you made, Aman. What does that target customer look like that equals higher propensity to consume?

Jay Krebs, CEO, Confluent: Yes. So you can think of it as kind of broadly global 2,000, but with a bit of a tech lens. So if ultimately if somebody spends a lot on IT is going to be a good customer for us. And so kind of taking that lens and really disproportionately focusing on those customers, that’s what has proven to be successful for us over time. And the work we’ve been doing is just trying to it’s fairly obvious, but align your marketing and sales motions to make sure that you’re disproportionately investing in capturing those.

One of

Sanjit Singh, Infrastructure Software Analyst, Morgan Stanley Research Team: the things that like over the last year, what it sort of felt to me like that you guys are sort of testing the elasticity of demand. You made some brought down some of the storage costs. You’ve brought in the portfolio with warp stream flight clusters. There seems to be considered efforts to bring the cost of streaming down on average to unlock more use cases and greater usage. How would you assess the early efforts?

And looking ahead, what are the signals or the early metrics that you’re looking to gauge progress on sort of the unlock, if you will?

Jay Krebs, CEO, Confluent: Yes. This has been really important for us. When you think about what is out there to go capture, it’s all this batch data processing, which is relatively cheap to do and kind of a standard, but is slow and just doesn’t meet the needs at the moment. And so if we want to go get that, it has to be the case that our offering is easy to use, but also cost effective. Like people will only take a huge step up in cost if they really have to, right?

And so that’s an adjustment that we’ve been making in our offering over time. And it’s really paid off. Whenever we announce any kind of price change or new offering that’s cheaper, certainly in the investor base, people are like, yes, what’s going to happen, oh my. And of course, we watch it very closely as well, right? But when you think about these things, the negative impact is, of course, immediate, right?

Your existing customers can optimize, etcetera. But the long term impact of soaking up this larger pool of usage and expanding that is the much bigger opportunity. It kind of comes back to what percentage of the market do you already have. And so we’ve looked at these relatively scientifically in terms of what we think it’s going to drive. We’ve rolled out a series of tests.

Those have paid off even on a pretty short timeframe. And we’ve been through some of those examples. Storage was one of the ones where we went through it and we said, well, okay. We cut storage prices. On day one, of course, we’re getting less money for storage.

And so that’s negative. But very quickly, we saw that then the usage of that those storage capabilities in our platform expanded and people took advantage of it because it was cheaper. And so just even on the over the course of a few quarters, now you’re on a run rate that’s better. And I think you have to have that kind of long term view on the market you’re going after if you want to really capture it. For us, the big picture belief is that streaming is not some niche over on the side that’s for some kind of very elite sub segment of use cases, but this is actually very broadly the way companies are going to move all their data.

And we want to be set up to do that. And that’s the bigger value proposition and the more strategic platform to go capture. And I think what you can see and what we’ve done is we’ve been responsible in it. We haven’t pulled the rug out from under the business. We’ve made a set of changes that have opened up our ability to bring on customers, have taken friction out of it.

And I think that’s been a successful thing for us so far.

Sanjit Singh, Infrastructure Software Analyst, Morgan Stanley Research Team: Yes. It sounds like that thesis is playing out and that’s definitely exciting. One of the questions I used to get obviously over the last eighteen months is Confluence positioning for GenAI. And I would say, I think they’re sort of a long Confluence of long term beneficiary. But I think frankly, I was a little skeptical on these sort of wave one of Gen AI applications, what the sort of real time requirements are.

I have to say that surprised me to the upside certainly in our work. But starting in the second half of last year, we were starting to hear more and more customers talk about bringing Confluent to operationalize their Gen AI use cases. What sort of use cases or pattern matching you’re seeing from customers? And why do you think Confluent is participating this early in some of these Gen AI use cases?

Jay Krebs, CEO, Confluent: Yes. I’ll start with just big picture, what’s happening in the AI landscape and what is like the architectural need for that at the data layer. With traditional machine learning, a lot of the action was kind of building these custom bespoke models. And this was something that would happen offline as model building does. It would happen in the data lake.

There’d be some team of very specialized data scientists that would work on this. With Gen AI, it’s actually quite different. The model is built by OpenAI or Anthropic or Meta or whomever, right? And a lot of the usage is now much easier, but it is happening the data usage is now at inference time, meaning at run time as part of the application. So if you step back and you think, hey, what’s happening with data here?

We’re taking a set of uses of data and we’re moving it into the operation of the business. We’re trying to take problems that needed humans and we’re trying to do it with software. And the way humans work is continuously throughout the day and they’re part of a larger business that’s all exchanging things and doing things all day long, right? And so we’re taking the use of data into the operation of the company in real time. The action is now at the inference side where the context is brought, where you’re bringing data to bear and where you need things that are up to date with the state of the business.

And so all of that is pulling the use of data out of the batch world and into the real time world. So that is a shift companies have to make to be ready for AI period. Whether they do it with us, they do it some other way, Fundamentally, they’re going to have to have more of their data available in real time for real time use. And then the opportunity for us is we’re by far the best way to do that. And so, yes, we’ve seen this the first wave of usage is really just being the supply chain for real time data.

I think there’s a second wave of usage, which is really around these agents, which are actually taking on more sophisticated parts of what the business would do. That kind of, hey, take over some business process and run it is very much a continuous stream processing problem in many of the cases. And so both the building of these agents and the supply of data for them, I think is a huge opportunity for Confluent and a structural shift that’s incredibly positive for us.

Sanjit Singh, Infrastructure Software Analyst, Morgan Stanley Research Team: Let’s dive into the agent opportunity. Like one of your keynotes at your customer conference last year, you very powerfully signal that we’re moving beyond chatbots and Q and A style apps. Agents are definitely the next frontier. Maybe see the perspective of what an agent needs to be perspective, maybe lay that out and then how Confluent fits into those AgenTek architectures?

Jay Krebs, CEO, Confluent: Yes. So, yes, it can be a little confusing because any of these buzzwords gets heavily used. So like when you say agent, some people mean something that’s clicking buttons on my computer. Some people mean a chatbot on my phone that can do a background task to book a flight for me. But what I’m going to mean is actually something which takes over some business process you had that used to involve a human component and does it without that.

And so what would examples of that be? In our business, this is something where, okay, a lead comes in, we would gather a fair amount of data, research, put together a whole thing for SDRs, some outreach. In insurance, it would be the processing of claims. All of these are areas where there’s been a fair amount of digital technology, but there’s also like humans doing work as part of the business process and the opportunity is to substitute for that or amplify it in some way. And so, yes, for Confluent then the opportunity is, hey, bring that data to bear that this agent would need.

These models have context about the basic world as of six months ago, nine months ago, a year ago, whenever their training cutoff was. So they kind of know Wikipedia stuff and they have increasingly powerful reasoning capabilities. But to do anything useful with that, you need to know about the business you’re in, the context of that customer, that product, their interactions, what happened before. Those are going to be the things that actually allow you to make a decision, produce an output, take an action. And so bringing that together and then orchestrating that, that’s kind of the opportunity for us.

And I think that’s very much, as I was saying, pulling out of the batch world and into this kind of continuous streaming world.

Sanjit Singh, Infrastructure Software Analyst, Morgan Stanley Research Team: And so it’s sort of like you’re sort of enabling that sort of real time decision making for the agent and we’ll definitely just see how that’s going to play out going forward. Maybe let’s talk a little bit about a classic question that we can probably get every conference around competition, this sort of have the sort of new take on like on what the competitive environment looks like. So we’ve always had the hyperscalers that’s been around. Increasingly, you’re seeing the data platform companies, the third party data platform companies, looking to build streaming within their platforms. So maybe just kind of go back to first principles and sort of give us your perspective on why there is a need for Confluent as a sort of cross platform vendor in a world where a lot of platform companies are building out their own streaming capabilities?

Jay Krebs, CEO, Confluent: Yes. I’m happy to address that. So as you said, kind of the way we would think about competition is much as you did. There’s kind of direct competitors coming out of the cloud providers. Now the cloud providers themselves are not competitors.

Those are actually pretty strong partners. But within that cloud provider, they might have two eighty products that we cooperate with and integrate with and they might have five to seven that we compete with. And overall, it’s a relatively positive relationship, although we would, of course, bake off with some of their offerings. There’s always been startups and open source technology in this space. We’ve been, I think, very dominant in the streaming world relative to this.

And then there’s a couple other data companies, right? And I think we’ve had a good partnership with them over time. You will see throughout the whole ecosystem, everybody has to adapt to streaming in some way. Like there will not be a data company. It doesn’t have something that says streaming on it somewhere because ultimately, if you are a database, you have to produce real time streams of data into this larger ecosystem.

If you are a data warehouse, you have to have some ability to ingest data more in real time. The set of use cases for companies are very much forming around that. So that you’re going to see streaming everywhere. There is overall a paradigm shift in the data world towards this out of the big batch of data that comes at the end of the day to this continuous incremental model that’s part of how you would run a data business on data. So that’s happening.

It’s not necessarily competitive. Despite that, we have great partnerships. We just announced really significant partnership with Databricks. And the role we would play is kind of as the central nervous system is getting real time data around. We’ve built really deep integration into these open data formats that they build around.

So these are things called Delta or Iceberg. They act as an interface for any analytics tool to sit on top of. And we’ve done the work now so that any data that streams into Confluent is now opened up in those formats. We call this table flow. And it’s a huge thing for customers.

They were building in their operational apps all this integration around Kafka. Now that same data can just open up to the analytics world. Databricks is a phenomenal partner in that way. They’ve built around open data effectively from the beginning, but really kind of standardized it around Delta. And that partnership allows us to have a really great integration with them.

It’s something customers on both sides are extremely passionate about. One of the hardest things in the analytics world is getting data, getting data on a timescale that’s actually useful, making sure that that data flow is in fact reliable and doesn’t break all the time. And so this is something I think is very enabling for their customers and very positive for us. When you think about the use of AI, it’s happening everywhere. It’s happening in a lot of the operational applications throughout the data center.

It’s happening in these big analytics platforms. All of that requires this real time data. We want to connect it. When it comes back to what the role that we’re aiming at though, the role we’re trying to be is very different from what I think Databricks or Snowflake or BigQuery or any of those are trying to be. They’re ultimately trying to be an analytics and intelligence platform, right?

They want to serve data scientists, set up reporting use cases, set of uses around AI. That’s not everything in the business. And what we’re trying to be is the central nervous system that connects all the parts of the business, including that, right? And it allows you to act on data continuously in real time. So ultimately, we’re kind of selling to different people.

We’re serving different areas and the technology is very complementary. And I think that’s been a basis for really good partnership.

Sanjit Singh, Infrastructure Software Analyst, Morgan Stanley Research Team: You called out the data which partnership. I was going to ask about that and sort of describe kind of the technical logic chain of why that partnership is exciting. Anything else about that partnership maybe on the go to market side or any sort of joint selling efforts?

Jay Krebs, CEO, Confluent: Yes, absolutely. So it is end to end. It’s kind of deep integration between the two products and then going out to all of our joint customers and even those who are not customers yet on one side or the other and showing them this solution. Because it’s ultimately about the acquisition of real time data, it plays very well with things customers care about, which is opening up data in these open formats, so I can have a whole ecosystem around it and operationalizing some of these AI apps. If you were to ask customers in the analytics world their priorities, I think those would be right at the top.

And this is a mechanism to get both of those. And so yes, we’re taking that out jointly with our sales teams, ton of cooperation around that, and we’re very excited about it.

Sanjit Singh, Infrastructure Software Analyst, Morgan Stanley Research Team: Awesome. Let’s talk a little bit about DSP and maybe you can define for us what DSP means for Confluent. One component of DSP has been Flink and stream processing. Can you give us sort of an update? You’ve had the Flink offering available for several months now.

What has early adoption looked like? And maybe fundamentally, why is stream processing an important piece of the DSP puzzle?

Jay Krebs, CEO, Confluent: Yes. Happy to touch on that. So first of all, what DSP, every company has the acronyms, right? The first thing we started with was Kafka, which is you can think of it as just the core streams of data that would flow around a company. And that was the basis of our offering and the bulk of the business and revenue.

What we understood really from the beginning and had aimed at was to provide a more complete platform around real time data. And so what are the needs that companies have for data? Well, it’s not that different from other areas of the company. You want to process the data, you have to actually acquire the data, you want to be able to govern that flow so that you can actually open this up. And we saw our customers needing this stuff.

So we’ve made a set of long term investments around that. And so the processing capabilities are based on a technology that’s open source and it’s kind

Sanjit Singh, Infrastructure Software Analyst, Morgan Stanley Research Team: of an open source standard called Flink.

Jay Krebs, CEO, Confluent: The connectors is a large ecosystem of things that plugs into all the systems our customers have and the governance capabilities ensure that the data is correct, that it doesn’t break, that there’s a strong contract between parts of the organization that you can trace and secure that flow. And you can imagine in both the mission critical applications, but also in the AI world, that’s really important. Like what is this data? Where did it come from? It’s bad enough if you’re breaking some downstream report, if you’re breaking some action that’s being taken inside of business, it’s even worse.

So those are the capabilities that we call DSP or data streaming platform, like going from just Kafka to a larger platform around real time data. And yes, it’s been quite successful. There’s a very nice ramp for that that’s in our cloud outgrowing the kind of core cloud offering quite significantly. For Flink, that’s been one of the biggest investments, like really getting into the processing of data. And we think that opportunity over time is comparable to the Kafka opportunity in size and scale.

A lot of work. We’ve built, I think, a phenomenal product, something that’s really serverless that expands elastically. It elasticly. It kind of brings a lot of the capabilities you would expect in a data warehouse or offline DACH system that brings them into real time. And it’s really we’ve seen a lot of enthusiasm with customers, early set of production use cases that have ramped very quickly off that, spanning all kinds of customers.

We thought this might be faster and just kind of pure tech, but we’ve seen grocery stores, car companies, everything you can imagine. And that allows companies to transform the data as it flows, aggregate it, filter it, work with it or build applications to act on it can be the basis for these AI agents where you can actually take the model and apply it directly to the stream. So those processing capabilities are quite powerful and quite strategic for us.

Sanjit Singh, Infrastructure Software Analyst, Morgan Stanley Research Team: Yes. And so as Rhoda mentioned, DSP is now 13% of the business. If you back into the growth rate, it’s 3x of what the overall business is growing. So definitely seems to be like an important driver of growth going forward. I want to take the like what you sort of laid out in DSP and the platform strategy and sort of map that to the go to market strategy.

And then you guys made some changes moving sales to more consumption oriented model last year. As we execute on this sort of platform strategy, how is the sales process having to change and evolve versus when you were primarily focused on just the core streaming opportunity?

Jay Krebs, CEO, Confluent: Yes, it is a substantial change. We’ve made a number of adjustments in that area, the move to a consumption model last year and then really driving effectively multi product sales. The kind of motions that you have, the first motion is very obvious, which is just kind of land with our Kafka offering expand to these other capabilities. And inherently, of course, we’re doing that with all the customers that have adopted Kafka, but haven’t yet adopted the other components. Increasingly though, we are seeing a set of customers that really land with all of it.

If you’re thinking about AI applications, the flow of data into the analytics realm, some of these other use cases, what are your needs? What your needs are capture data from a bunch of different sources using our connectors. It’s all about the real time flow and transport. It’s going to require some processing on the fly of that data. It’s going to require governance of that data.

It’s going to require in the analytics case, this table flow component that lands it there. So it really is people starting now with that full platform, which is very much what we would like to see happen. And so I think we are going to have both motions, both go soak up all the world Kafka motion and expand out to DSP as well as a land for specific use cases with the whole platform all

Sanjit Singh, Infrastructure Software Analyst, Morgan Stanley Research Team: at once. Yes, it makes no sense. Rohan, as sort of Jay lays out this sort of multi product strategy, how are you thinking about the investment allocation piece of that puzzle? A lot of things going on in products. You guys have been getting more efficient.

What do you sort of look for to get confidence that a certain area of the business or certain product is worth putting more investment behind and scaling? And when do you like make the call of like being able to pull investment back?

Rohan Srivium, Chief Finance Officer, Confluent: Yes. When I think about investments, the resource allocation philosophy is important. What’s your true north? And that’s more around making sure we are building a company that’s driving durable growth over a long period of time and we are doing it efficiently. And when you think about resource allocation, every year we as a management team, we sit together.

And ultimately, you have a the biggest constraint is the amount of dollars you’re going to invest. And you try to allocate that into R and D where returns are between eighteen to twenty four months. You try to put your money in sales and marketing where you get returns between nine to twelve months. And you’re investing in G and A, which kind of helps you get more efficient, build the right platform over the long term. And for R and D, a big part of our resource allocation strategy over the last few years has been around future bets.

And that’s critical, that’s key to make sure that you’re really taking a step back and identifying those future growth drivers for us. And obviously, having Jay right there, he’s obviously thinking through what the next five years could look like and strategically allocating resources on that front is a key part of what we do. So specifically, the process we run, obviously, there is this technology evaluation that happens. There’s a commercial evaluation that happens. And ultimately, you test it out if you have product market fit, you’re going you’re obviously going pushing harder.

I’ll give you example of Flink. If you go back to our early decks or early series C, Series D funding decks, stream processing was a key part of our overall strategy. We attacked that problem with a different product, ksqlDB. Fast forward a few years, we realized that this open source ecosystem was gaining traction. We pivoted.

We acquired this company. We have a product that Gaid and Jay just spoke about some of the momentum that we are seeing. So in general, that’s part of our overall resource allocation strategy. That’s nimble. We have to make changes.

We have to adapt and that’s how we think about it.

Sanjit Singh, Infrastructure Software Analyst, Morgan Stanley Research Team: Awesome. Ron, when you think about 2025 in particular, how would you sort of stack rank the key investment priorities going into next year?

Rohan Srivium, Chief Finance Officer, Confluent: Yes. I mean, obviously, when you look at a twelve month timeframe, you have to start with go to market. And on the go to market side, the way I think about investment priorities is number one is core execution. You need the right capacity on the field to execute on the plans. You need to build an incentive plan, comp plan structure that gets us to the outcomes that we want.

So that’s number one. Number two, we touched on the fact we’re making this transition from a single product company to a platform. So platform selling, thinking about specialists and that area is a focus area. And the third part on go to market is around partners and partner ecosystem. We’re getting to a scale where we are also equally attractive to our partners and making sure we are leaning in, be it the GSIs, be it DSIs, be it strategic partnerships, that’s an area of investment.

So that’s on the go to market side. On G and A, we talk about AI, we talk about efficiency. So investing in efficiency, investing in AI technology for the company internally is a focus area for us. That’s something that we’ll continue to spend time on. And from an R and D perspective, you’re again trying to make sure you’re balancing out what you need for your current products, what are the future bets and that’s the balance you’re looking at.

Sanjit Singh, Infrastructure Software Analyst, Morgan Stanley Research Team: Awesome. Jay, maybe the last minute, I wanted to hit on two topics. So maybe bundle them in. We’ve talked a little bit about Tableflow. I find Tableflow really exciting just given all of the sort of evolution of the modern data infrastructure.

But from a governance and table flow, how do we think about those capabilities driving better monetization, better growth on this PCN?

Jay Krebs, CEO, Confluent: Yes. Well, maybe I’ll talk a little bit more about table flow, what it is the opportunity and then I’ll start with your second question, which is like how will it drive better monetization? Yes, I mean, they’re both paid offerings that drive consumption. And so we will see we internally, of course, have a line item that would show up on customers bill for governance or for table flow usage. But beyond what they directly drive, they actually incentivize more usage of everything.

And one of the things that we found with governance was if you want to really open up sharing of data across the company, if you can provide a channel that is safe and reliable, they will put a lot more through that. And so the first investments in that area were actually just to motivate more usage of the platform, unlock more across our customers. The direct monetization was actually sucking. That direct monetization has gone extremely well. And so that’s turned into a really successful product area for us.

Similar thing with Tableflow where of course we’re directly charging for it, but every time you’re using it, you’re also using connectors and Kafka and Flink. And so we expect to see some amplification there beyond just the direct monetization. So that’s the that’s the second question. Back to what it is, I touched on this briefly, but it’s worth maybe recapping. What’s happening in the cloud is ultimately the storage layer, this object storage S3, it’s become very cheap and very reliable.

And in infrastructure, that kind of cheap, reliable storage, it has a sort of gravitational pull that forces the rest of the infrastructure world to orient around that ingredient. And for Confluent, we’ve done a ton of work to orient our offering to work off of that. That’s part of what’s driven you mentioned the fact that as we’ve grown a larger percentage of cloud, we’ve actually improved gross margins, right? And part of that comes from multi tenancy, part of it comes from underlying efficiency driven by modern architectures on the cloud. But it also is a huge impact for data storage and sharing in the large.

And the other big implication of this is in the analytics world. And so in the analytics world, it’s traditionally been the case, hey, you buy some data warehouse, that thing owns all your data. It’s a box that has all the data in it. Anything that uses data in an analytics context, you got to pay the toll of the data warehouse and drive some query processing loads just to get to the stuff. And so effectively for a data warehouse, you would compete to win to be the data warehouse and then you would kind of earn money on every single usage of analytical data from then on.

That is changing, right? What this open data means is in this cheap object storage in every customer’s account, They have a copy of all their tables of data and they can use it with any tool they want. So if there’s some new AI product they want to plug on top of it, they can do it. If they want Snowflake, they can do it. If they want Databricks, they can do it.

If they want something from the cloud providers, they can do it. And the thing that held that back before was of course, you would have to recreate all the flow and processing and updated to get it into a usable form. That’s gone. It’s now the same data for everybody. So it’s a massive shift in that part of the world.

One of the reasons I’m excited about the partnership with Databricks is they’re a company that built around open data in S3 from the beginning. So they’re actually quite well positioned for it. The so what does that have to do with table flow? Well, as the standard for these tables is emerging, what we’ve done is we’ve taken the stream of real time data that populates the table and we’ve merged the two. So in Confluent, you would have both the stream of the real time events as well as that representation of the long term storage.

And that long term storage can now be exposed, opened up to these other offerings. And so, yes, it’s a huge strategic bet for us. It’s not just a better way of getting data into the analytics world. It also now opens up all this historical data for stream processing. So when you think about real time processing, of course, partially that’s about the new stuff coming in right now, but you also want all the historical reference data of what has happened, all the information about your products, your customers, etcetera, that you can join on, that you can take into account.

And one of the things that’s held back real time stream processing has been well, you don’t have that stuff in that context. You have that offline in the batch systems. So now we get access to that. The many analytics systems get access to that. We all get access to it.

And we’re driving growth in our business around that. So that’s why we’re so excited about it.

Sanjit Singh, Infrastructure Software Analyst, Morgan Stanley Research Team: Hopefully that new day. Yes, quite the paradigm shift that we’re out of time. Thank you for giving us a tasty opportunity. Yes,

Rohan Srivium, Chief Finance Officer, Confluent: it’s very much a pleasure.

Jay Krebs, CEO, Confluent: Anybody who wants more, we’re doing an Investor Day tomorrow, so you can join us for that. It’s right down the street. Thank you, Jason. Awesome.

This article was generated with the support of AI and reviewed by an editor. For more information see our T&C.