Confluent, Inc. (NASDAQ:CFLT) Q3 2024 Earnings Call Transcript

Confluent, Inc. (NASDAQ:CFLT) Q3 2024 Earnings Call Transcript October 30, 2024

Confluent, Inc. beats earnings expectations. Reported EPS is $0.1, expectations were $0.05.

Shane Xie: Welcome to the Confluent Third Quarter 2024 Earnings Conference Call. I’m Shane Xie from Investor Relations and I’m joined by Jay Kreps, Co-Founder and CEO; and Rohan Sivaram, CFO. During today’s call, management will make forward-looking statements regarding our business, operations, sales strategy, market and product positioning, financial performance and future prospects, including statements regarding our financial guidance for the fiscal fourth quarter of 2024 and fiscal year 2024. These forward-looking statements are subject to risks and uncertainties which could cause actual results to differ materially from those anticipated by these statements. Further information on risk factors that could cause actual results to differ is included in our most recent Form 10-Q filed with the SEC.

We assume no obligation to update these statements after today’s call, except as required by law. Unless stated otherwise, certain financial measures used on today’s call are expressed on a non-GAAP basis and all comparisons are made on a year-over-year basis. We use these non-GAAP financial measures internally to facilitate analysis of our financial and business trends and for internal planning and forecasting purposes. These non-GAAP financial measures have limitations and should not be considered in isolation from or as a substitute for financial information prepared in accordance with GAAP. A reconciliation between these GAAP and non-GAAP financial measures is included in our earnings press release and supplemental financials which can be found on our IR website at investors.confluent.io.

And finally, we will post the Confluent earnings report to our IR website after our prepared remarks. And with that, I’ll turn the call over to Jay.

Jay Kreps: Thanks, Shane. Good afternoon, everyone. Welcome to our third quarter earnings call. Subscription revenue grew 27% to $240 million. Confluent Cloud revenue grew 42% to $130 million and non-GAAP operating margin expanded approximately 12 percentage points to 6.3%. And I’m proud to report that total revenue grew 25% to $250 million, surpassing a $1 billion revenue run rate in just 10 years since Confluent was founded. In Q3, we hosted Current 2024, the only industry event fully dedicated to all things data streaming. More than 4,200 people from 1,200 companies participated, making it our biggest and best current yet. Data leaders from Mercedes-Benz R&D North America, Viacom 18 and Accenture joined me on the keynote stage to discuss how Confluent sits at the heart of their companies, allowing them to push the boundaries of what’s possible for their customers.

And some of the most popular sessions focused on how companies leverage data streaming to power transformative AI use cases like creating customer chatbots, building AI and ML pipelines to detect fraud and delivering hyper-personalized AI customer experiences. We continue to see excitement, interest and use cases around Gen AI growing across our customers and in the ecosystem of AI solutions providers. Last week, we hosted our first Confluent AI Day, a 1-day event designed to help our customers advance ideas into fully built AI applications. In partnership with AWS and MongoDB, we brought together hundreds of attendees from companies like Google, PNC Bank, Whirlpool and Rocket Mortgage, who joined expert discussions, interactive sessions and an exciting AI hackathon.

Q&A Session

At the event, we launched the Confluent for Start-ups AI Accelerator program. This exciting new program is about empowering early-stage AI companies with the tools and mentorship they need to lead in the world of generative AI. Confluent for Start-ups AI Accelerator program provides start-ups with early access to Confluent’s latest AI tools, expert mentorship and product credits from Confluent, MongoDB and Anthropic. We’re committed to helping these start-ups create new breakthroughs in real-time AI. We spoke about our relationship with OpenAI during the Q4 2023 earnings call when we discussed how OpenAI’s team uses Confluent to deliver real-time data streams. I’m happy to report that OpenAI has expanded their use of our data streaming platform to help scale with the increased usage of their platform.

The momentum and growth further validate the strategic role of data streaming in the generative AI landscape. Last quarter, we also celebrated our 10th anniversary as a company. When we started Confluent 10 years ago, data streaming was just emerging as a nascent paradigm. What started with a small group of companies like LinkedIn, Uber and Netflix, disrupting the status quo with real-time data streams has turned into a movement. Today, more than 40% of the Fortune 500 rely on Confluent to set their data in motion. We serve customers broadly across industries, including 10 of the 10 top U.S. banks, 8 of the top 8 global carmakers and 9 of the top 10 U.S. insurance companies. We see significant expansion opportunities across our customer base as we expand from individual use cases to the central nervous system for real-time data.

Our growth in capturing this opportunity has gone through 2 distinct waves and is now entering a third. The first of these waves was built directly on the back of the open source traction and was about commercializing that with our software offering, Confluent Platform. That provided the bulk of our business in the first 5 years of Confluent’s growth. However, we knew that for the long-term platform we wanted to build and to capture the bulk of the opportunity around streaming, we needed to make streaming far, far easier to consume. This spurred the early investment into what fueled our second wave of growth, Confluent Cloud. Indeed, even when we went public 3 years ago, our cloud business was a small percentage of our revenue and we were in the early stages of taking our cloud business to scale.

That being said, we strongly believe that the secular shift to cloud would present a meaningful long-term driver of growth. I’m proud that our team has successfully executed on our cloud vision and Confluent Cloud is now more than 50% of total revenue and continues to outpace our Confluent Platform business. At the same time, cloud is our most frictionless path to monetizing the thousands of organizations using open source Kafka. But with 150,000 organizations using Kafka, we’re just getting started. Already, our cloud product comprises over 90% of our customers, demonstrating its broad appeal as we continue to grow into this space of open source usage. These first 2 waves aren’t done. We continue to work to serve the broad base of Kafka users through compelling pricing and packaging optimizations for Confluent Cloud and Confluent Platform.

Our differentiated cluster types like enterprise and freight enable us to deliver data streaming offerings for all customers and workloads with low TCO and strong ROI. Our recent acquisition of WarpStream adds a third deployment mechanism of BYOC to this portfolio. WarpStream’s bring your own cloud model offers a deployment model midway between fully managed and self-managed and opens up opportunities in a set of high-volume, high-tech customers that form a good chunk of our digital native customer base. WarpStream is BYOC done right, built directly on top of object storage. WarpStream’s zero disk architecture enables zero ops auto scaling while making it 5 to 10x cheaper than other alternative systems. And unlike traditional BYOE offerings, WarpStream prioritizes security by avoiding break glass access to customer networks and systems.

Confluent is now the only company with a data streaming offering for everyone, regardless of use case, cloud environment or deployment type. Kafka is the foundational layer of our data streaming platform and could sustain our business for many years on its own but it only represents a portion of the opportunity ahead of us. We believe our third wave of growth comes from being a complete data streaming platform, a one-stop shop for all real-time data needs. To do this, we are bringing together the key capabilities to stream, connect, process and govern continuously flowing streams of data so organizations can power their next-generation real-time applications. Over the course of the past year, we have been on our most aggressive pursuit of our vision since we started the company and that is starting to yield strong traction.

Major new product and pricing innovations like Flink, TableFlow, freight clusters, AI model inference and new connectors will extend our already significant category lead. And we continue to see strong traction across our customer base. Our DSP portfolio continues to grow substantially faster than overall cloud revenue. One of the areas we’re most excited about is the opportunity around stream processing at Apache Flink. Let me share 2 examples of how customers are using Flink on Confluent Cloud and Confluent Platform. One of the largest private companies in the U.S., a Midwest grocery chain with over $20 billion in revenue is using Confluent’s fully managed Flink offering to accelerate the growth of its e-commerce business, a critical driver of the company’s revenue.

This retailer had already overhauled its e-commerce solution with Confluent Cloud and wanted to integrate stream processing for all the Kafka topics that had built inside its digital environment, including pricing, promotions and inventory details without any lag in production. So the retailer implemented Confluent Cloud for Apache Flink to combine and enrich streams of data flowing across hundreds of retail stores, its website and mobile app and third-party fulfillment partners like Instacart. This data spans more than 100,000 product SKUs and tens of millions of orders. With our Flink offering, this retailer’s real-time inventory and pricing are accurate and customized to each local market so the company can consistently deliver a trustworthy and personalized shopping experience to its customers.

Since working with Confluent, it has grown its e-commerce business by 700% and can stay a step ahead of the national grocery chains it competes with every day. A Fortune 50 telecom company in the U.S. and a Confluent Platform customer is using our offering for real-time analytics. Initially, the telecom provider used an alternative stream processing tool which struggled to meet the demands of real-time data processing. This affected how the telecom’s enterprise customers could serve consumers and led to higher churn. So the telecom provider deployed Confluent Platform for Apache Flink, shifting processing to the left and rolling out thousands of Flink instances across its infrastructure to run real-time analytics on data earlier in the data pipeline before it moves downstream.

Flink processes and analyzes data such as network performance to help its customers deliver consistent personalized experiences to consumers and network visibility for threat detection. By using Confluent Platform’s Flink offering and tapping into our team of Flink experts, the telecom provider has saved tens of millions of dollars and significantly reduced churn, boosting its overall margins. In closing, I’m pleased with our strong third quarter results and I’m incredibly excited about the opportunity ahead of us. I’m even more excited for the next 10 years. We’re in a prime position to win the $60 billion data streaming category. With that, I’ll turn things over to Rohan to walk through the financials.

Rohan Sivaram: Thanks, Jay. Good afternoon, everyone. In Q3, we drove robust top line growth, record gross margin and another positive quarter for both non-GAAP operating margin and free cash flow margin. These results demonstrate our market leadership in data streaming and our commitment to driving efficient growth over the long term. Q3 subscription revenue grew 27% to $239.9 million, exceeding the high end of our guidance and representing 96% of total revenue. Confluent Platform revenue grew 13% to $110.1 million and accounted for 46% of subscription revenue. The strength was driven by healthy demand for Confluent Platform in the financial services industry. We serve 10 of the top 10 U.S. banks with an average ARR of greater than $5 million.

The substantial majority of their ARR is attributed to Confluent Platform as these banks are still early in their move to the cloud. Confluent Cloud revenue grew 42% to $129.8 million and accounted for 54% of subscription revenue compared to 48% a year ago. We saw consumption stabilization in our digital native customer cohort during the quarter. While they remain cost conscious, we were pleased with the consumption growth trajectory of our largest cloud customers, many of whom are shifting their focus to implementing new use cases and adopting our DSP products. Q3 cloud revenue also saw a onetime low 7-figure revenue benefit. Adjusted for this benefit, we still handily exceeded consensus expectations. Revenue from DSP continued to grow substantially faster than our overall cloud revenue.

While monetization remains in its early days, we are pleased with the adoption of new products by our large cloud customers. 19 of our top 20 cloud customers have adopted at least 1 DSP product and 13 have adopted products across all 3 categories. Additionally, multiproduct customers continue to grow at a faster clip and exhibited a much higher NRR profile. Turning to geographic mix of total revenue. Revenue from the U.S. grew 28% to $152.4 million. Revenue from outside the U.S. grew 21% to $97.8 million. Moving on to rest of the income statement. I’ll be referring to non-GAAP results, unless stated otherwise. Subscription gross margin reached a new record of 82.2%, up 210 basis points, while total gross margin also reached a record high of 79%, well above our long-term target.

Our gross margin outperformance continued to be driven by strong Confluent Platform margin and the improving unit economics of our Confluent Cloud offering. Turning to profitability and cash flow. Operating margin expanded approximately 12 percentage points to a record high of 6.3%, representing our ninth consecutive quarter of 9 points or more in margin improvement. Free cash flow margin of 3.7% was also a record, expanding 10 percentage points. This marks our third positive quarter for both operating and free cash flow margins and reflects our team’s track record of driving margin expansions at scale. Net income per share was $0.10 for Q3 using 353.6 million diluted weighted average shares outstanding. Fully diluted share count under the treasury stock method was approximately 366.8 million.

And we ended the third quarter with $1.86 billion in cash, cash equivalents and marketable securities. During the quarter, we acquired WarpStream to further differentiate our data streaming platform to include the BYOC native form factor. WarpStream is particularly well suited for digital natives and high-scale workloads with relaxed latency requirements such as logging, observability and feeding data lakes. In fiscal year ’24, we do not expect WarpStream acquisition to have a material impact on our financials. Over time, we expect WarpStream to be a growth driver as it expands our reach into more workloads across customer segments. Turning now to other business metrics. During the third quarter, we saw a notable increase in overall win rates for new business, both year-over-year and sequentially.

Our win rates against smaller start-ups were well above 90% as we compete favorably with our cloud-native complete and ubiquitous platform. This translated to sustained momentum in new logo acquisition and customer expansions. Total customer count growth accelerated to 16% and ended Q3 at approximately 5,680, representing a sequential add of 240 customers, 3x the sequential add of the year ago quarter. New customers include a top 3 U.S. airline company, a Fortune 50 carmaker, one of the largest online meal kit providers, a leading lifestyle retailer, one of the world’s largest online furniture companies and many more. The network effects of our data streaming platform continues to take hold in our large customer base. We added 40 customers with $100,000-plus in ARR and 7 customers in $1 million plus in ARR, bringing the total to 1,346 and 184, respectively.

Our $100,000-plus ARR customers continue to represent more than 85% of our revenue. Our new $1 million-plus ARR customers include customers from a variety of industries, including health care, travel and retail, technology, financial services and more. Q3 NRR was 117%, while GRR remained above 90%. We saw many of our large digital native customers shifting their focus from cost optimization to new use case implementation and adopting DSP products. This trend has continued into October which we believe will help stabilize our NRR around current levels in Q4. Turning now to guidance. We are increasing our Q4 revenue outlook in addition to raising full year subscription revenue, non-GAAP operating margin, non-GAAP EPS and free cash flow margin.

For the fourth quarter of 2024, we expect subscription revenue to be in the range of $245 million to $246 million, representing growth of approximately 21% non-GAAP operating margin to be approximately 2% and non-GAAP net income per diluted share to be $0.05. For full year 2024, we are raising subscription revenue to be in the range of $916.5 million to $917.5 million, representing growth of approximately 26% non-GAAP operating margin to be approximately 2% non-GAAP net income per diluted share to be $0.25 and free cash flow margin to be between 0% to 1%. Looking back at the last 10 years as a company, we have established data streaming as a major category in the tech stack. As the data streaming pioneer, we have continued to extend our market leadership by delivering world-class innovation and business outcomes for our customers.

This has enabled our growth and profitability journey at scale. We exceeded $1 billion revenue run rate in just 10 years since inception, including growing Confluent Cloud revenue run rate from less than $50 million to more than $0.5 billion in just 4 years. We serve 5,680 great customers, including more than 40% of the Fortune 500 across a variety of industries. We sustained positive non-GAAP profitability metrics in Q3 with 79% total gross margin well above our long-term target threshold. 6.3% operating margin, now within the range of our midterm target and free cash flow generation at a record margin of 3.7%. And for the first time in Confluent’s history, we expect to exit 2024 with positive non-GAAP operating margin and positive free cash flow margin for the full year.

These are fantastic milestones for a 10-year-old company. I’d like to thank our employees and partners for your important contributions and our customers and investors for your continued support. Looking ahead, the intersection of cloud, data and AI reinforces our vision of companies becoming software and AI. Harnessing the power of data streaming will be more critical than ever for companies to deliver differentiated products and services, ultimately driving their success in the AI era. The secular tailwind puts us in a stronger position to drive durable growth while generating significant free cash flow over a long runway. We are more excited than ever about capturing our market opportunity ahead. Before turning to Q&A, I would like to announce that we will host Investor Day 2025 in San Francisco on Thursday, March 6.

Management will provide an update on driving profitable growth for the next few years. Please save the date. Now, Jay and I will take your questions.

Shane Xie: Thanks, Rohan. To join the Q&A, please click the raise hand button on your screen. When selected for Q&A, we ask that you limit yourself to one question and one follow-up. And today, our first question will come from Sanjit Singh with Morgan Stanley, followed by RBC.

Sanjit Singh: Congrats on a very solid Q3, particularly the revenue acceleration and the margin expansion that you’re seeing year-over-year. So Jay, I guess when we look at the past couple of quarters, there’s been sort of fits and starts with the digital native group and it looks like those stabilized. Are you at a point where you’re starting to see more confidence from your digital native customers in terms of bringing on new use cases, getting past optimization? Or is it still kind of touch and go quarter-by-quarter?

Jay Kreps: Yes. Look, I think each quarter is kind of a mixture of optimization activities and new use cases. I do feel like we felt like in conversation with the largest set of customers, they’ve kind of done the bulk of what they need to do in terms of larger changes in their environment. So we did see better growth this quarter and feel that puts us on a good trajectory going forward. So yes, I do think we feel pretty confident about that segment when we think about the year end.

Sanjit Singh: And I really appreciate the 2 customer examples on Flink. As we look at kind of this stage of where Flink is, do these consumers sort of represent the early sort of beta customers that are now exploring these use cases? And where are we in terms of getting like the broader base to start to onboard in use cases in their environments.

Jay Kreps: Yes. Yes, I think they’re very representative. So we’ve seen a ton of enthusiasm. We’ve had a bunch of product unlocks as we’ve released the private networking support across some of the clouds and getting it out to all of them. We’ve announced the programmatic support so you can write direct on Java and Python programs and that will be going through EA and GA. And so yes, we’re starting to ramp of production use cases. And these are some of the early ones. It was nice that we’ve been able to land not just Confluent Cloud customers but also Confluent Platform Link customers, even though that product is still in limited availability and just going towards GA. And so yes, we’re actually seeing a ton of enthusiasm in the customer base and we’re really excited about where that takes us in the year ahead.

Shane Xie: We’ll take our next question from Matt Hedberg with RBC followed by JPMorgan.

Matt Hedberg: Congrats all of my congrats as well. Maybe as a follow-up to Sanjit’s question, it was great to see the acceleration in cloud and the digitally native stabilized. I guess my question is on the go-to-market. You guys made some changes at the start of the year. It really looks like they’re paying off in terms of the kind of the record customers that you guys keep adding. Can you talk a little bit more about that process? And maybe comment on the level of consumption that you’re seeing. I know you’re trying to land fast and then accelerate those, that consumption trend. But maybe talk about what those customers are looking like when they land.

Jay Kreps: Yes. Yes, I’m happy to do that. So one of our goals this year, both in what we were doing with the sales team and sales compensation as well as on the product-led side of the building was to really broaden our reach into the large set of open source Kafka users and land more customers more quickly. And I think that’s an ongoing journey. We’ve made a ton of progress this year. We’re really pleased with that. I think we’ll continue to work on that in the years ahead. We feel like, look, there’s 150,000 organizations using Kafka. We want to go get them all. So yes, the sales focus on these lands has been really important. It’s both about numbers but also about targeting. I think we’re much more intentional about the customers we wanted to land with.

We have this idea of the Confluent 2000 which are the highest propensity or potential accounts and those are the ones that we’re targeting on the sales side. On the product-led side, it will fluctuate. We try out different things to try and land more customers but also track them through and making sure that we’re getting to high ROI customers, not just landing university students with their projects but really getting into the right companies. And so that total customer count will fluctuate quarter-to-quarter as it has. But we do think we’re on a better trajectory in terms of landing more customers at a faster pace. And then, of course, we follow these customers all the way through their adoption. Our goal was get into more customers earlier in the journey.

We’ve definitely done that. We’ve now been able to see a lot of these companies kind of start that ramp towards production usage and expansion out into other use cases. And so we feel pretty good about the overall trajectory of these organizations. So that’s definitely a positive factor heading into next year.

Matt Hedberg: That’s great. Yes. So I think a lot of momentum on the expansion side, too. I guess, Jay, you mentioned in your prepared remarks, sort of seeing early Gen AI demand. I think you mentioned some of those comments that [indiscernible]. Can you talk about like how that’s actually showing up in the customer conversations? Is it just translating to more consumption, maybe a little bit more specifics on how you’re kind of identifying that within your base?

Jay Kreps: Yes. Yes, absolutely. So I mean there’s 2 impacts. One is kind of growth in the set of AI providers, right, OpenAI being an example of that. The second is a new set of use cases in the wider enterprise customer base. There, it’s really about delivering data to these AI applications. And I think there’s kind of 2 things happening. The initial thing is a set of use cases really around that. I think the larger push that this is leading companies towards is more thinking and investment in data infrastructure overall. All of this, I think, takes time. I know when I talked to investors, some people felt this is going to be like an immediate pop in every infrastructure layer. I think we were very upfront that we didn’t think that would be the case.

When I talk to people now, some people think, oh, these AI things are never going to materialize. I don’t think that’s the case either. There’s something very real happening. There’s definitely a set of use cases and applications in customers. Different customers move at different bases. The tech companies are faster. The more conservative enterprises are a bit slower. But there’s definitely the rise of a new set of use cases around this. I think that’s a very positive thing for us.

Shane Xie: We’ll take our next question from Pinjalim Bora with JPMorgan follow by Barclays.

Pinjalim Bora: On the quarter from me as well. Jay, I wanted to ask you on WarpStream. Talking to your channel, we kind of picked up a large opportunity unlocked by WarpStream, seems like. But I want to ask you in general, is that broadly true? Is WarpStream kind of starting to bring you into conversation, especially as it relates to migration of open source Kafka?

Jay Kreps: Yes. Yes, absolutely. So the reason we thought that this was appealing was there’s a set of large users of open source where the wholesale migration to some fully managed cloud thing is actually a very big jump. And often hard to accomplish in one step, something like this that has a nice cost savings, keeps the team running it kind of in place but it allows you to get kind of halfway there is really beneficial. And we think that this can help us open up some of these large digital native companies that have been on the open source for a while. In some cases, since before, Confluent the company even existed and start to bring them into the fold and we’re really excited about that. We’ve started to see some progress in some of those companies. These are big accounts, so it takes time to land them but we’re pretty confident in where that’s going to take us in the year ahead.

Pinjalim Bora: Yes. Understood. And one for Rohan. Rohan, the NRR seems like downticked a little bit but your cloud kind of accelerated. So I’m trying to reconcile the 2, right? Was there a downtick and expansion from the long cloud portion or where the core organic cloud trends a little bit lagging, if you take out working.

Rohan Sivaram: Yes, Pinjalim, thanks for your question. Listen, when you really look at the results for Q3 and you look at the cloud results, we’re actually very pleased with our cloud growth. We grew the business 42% and it’s a $0.5 billion run rate business. And when you double-click into the NRR dynamics of it, it is the health of the installed base continues to be very solid. Our GRR for the overall business is greater than 90%. And there are 2 drivers of performance that we called out in the — especially on the cloud side for Q3. That was when you look at our digital native customer base, we saw stabilization in consumption. And second, for some of our larger cloud customers, we did see that these customers are taking up new use cases and adopting the DSP.

Why is that important? That’s important because these 2 trends have kind of moved into Q4. And that’s why I made the comment around stabilization of NRR around the current levels. To your specific question, listen, I mean, it’s very marginal. So not a whole lot to call out there with respect to the move from 118 to 117. But what I’ll call out is the couple of drivers that we are entering Q4 with gives us confidence around the stability of NRR around these current levels.

Jay Kreps: Yes. And I’d just add, when you think about the trajectory the kind of longer-term trajectory, I think we feel really excited about the set of product investments. Like I mentioned this in the prepared remarks that this has probably been the most aggressive period of new product development and release over the last, say, 1.5 years in Confluent, history, more or less as expounded. And I think that’s now kind of starting to come to fruition and we’re seeing really good signs in how that’s being adopted with our customers. A lot of work to do. But when we think about that longer trajectory, I think that’s a really solid driver for us of expansion in the customer base. the obvious use case for all of these is growing from just a pure Kafka usage in those customers to a broader platform that not only has more items you can spend on but actually allows you to address a set of use cases that would have been inaccessible otherwise or too difficult.

Shane Xie: Our next question from Raimo Lenschow with Barclays, followed by Deutsche.

Raimo Lenschow: Congrats from me as well. Two questions. One for Jay. Jay, if you think about work stream, could that be actually a bigger opportunity for you guys? Because I remember when those guys started out, they were like, okay, we want to be even more modern Kafka or like be more on serverless, et cetera. Is that kind of in theory, actually like a broader opportunity for you in terms of like using that more than where you are at the moment? And then the one for Rohan, like you called out the little bit of extra help you got this quarter in cloud. Can you kind of help us there a little bit because cloud in theory is like subscription, like how do you get that as a one-off thing? That would be helpful.

Jay Kreps: Yes. Yes. Let me try and address that. So — the way I would say it is this, open source Kafka is a very good kind of open source project. You can take it and download it and use it. What we look for is in each of these deployment models that we try and support, do we have the best possible product in that area? So if we think about our self-managed customers, there’s a set of things that they need. We’ve built an offering with Confluent Platform that’s really good at that. When we look at a fully managed cloud offering, we put a ton of investment in Cora which is the back end for that. It’s an amazing piece of software. It’s extremely sophisticated in making something that runs itself, balances data, expands elastically, all the really hard things in a full cloud service.

When we looked at this kind of bring your own cloud opportunity, we thought, well, is there something there? We felt there was a segment of the customers that would be easier to access if we had an offering there. We looked at like, hey, should we take Confluent platform and try and turn it into that? Should we try and take Kora and try and push it into the customer’s account in some way. The reality is neither of those would have been very good. Like we could do it, we could check the box but it wouldn’t really be a good product. What made us excited about WarpStream is they’ve actually done that in the right way. They actually had something that was designed from scratch for that deployment model. And that’s kind of a deep enough architectural divide that I think you need to do it that way to have a really good product.

And so now we feel really good where we have something that is kind of best-in-class across each of these deployment models. And so across customers, across customer use cases, across different parts of the company, we can now really span everything they want to do in the streaming world, kind of without hesitation or reservation. And I think that’s a really powerful position to be in. And that’s always been our goal with customers is make this something that’s a ubiquitous technology across the business.

Rohan Sivaram: I’ll take the second part of Raimo’s question. Raimo, as it pertains to the onetime revenue benefit, one of our existing customers had plans to basically expand into a new international market which did not end up materializing. As a result, we took some revenue at the end of the quarter as unused credits. As we speak, this customer is a very strategic partner of ours and we are working on multiple use cases with them. So if you take a step back and I know you commented on the Q3 performance, our Q3 performance which was very solid, the underpinnings of that performance was, I’d say, twofold, primary drivers. The first one was the stabilization of the digital native segment. And the second one was some of our larger cloud customers actually adopting DSP and starting to look at net new use cases. So this — the third benefit which you called out was onetime. But if you adjust it out, we still handily beat our consensus expectations.

Shane Xie: We’ll take our next question from Brad Zelnick with Deutsche Bank follow by William Blair.

Brad Zelnick: It’s great to hear all the excitement coming out of current and the improvements you’re seeing in win rates. And I appreciate it takes time for wins to turn into revenue. But is it too early to be characterizing what you’re seeing out there as perhaps green shoots? And how does it inform your early thinking about sales capacity as you gear up for next year?

Jay Kreps: Yes. I can address some of those and Rohan, you may want to address some as well. So yes, we’ve definitely seen positive signs in the customer base. We don’t typically draw that line too far forward. So we don’t try and make some kind of big pronouncement about IT spending next year. But yes, we’ve seen, I would say, stabilization and maybe some acceleration in investment in the digital native segment which is, I think, very positive. We’ve seen continued expansion across the broader base of cloud customers which is great. So yes, all of that gives us confidence. When we look at the sales capacity that we have in place, we feel very good about that. So this is a quarter where we saw kind of good ramping of new sales reps, lower than target attrition. Overall, that puts us in a good position as we think about what we’re entering next year with in terms of both our sales motion and kind of systems as well as just ramped sales reps to go execute.

Brad Zelnick: Rohan, maybe just a follow-up. As we think about the range of outcomes for next year, I appreciate that you’re not giving us guidance just yet. But is there anything you can tell us about perhaps the visibility that you have coming out of this Q3 versus the last couple of years where you did give us an early look. And I mean, it would seem to me that the bias would be towards acceleration, at least from the exit rate that you’re guiding here for Q4 for 21% subscription growth. How are you thinking about that? And how should we as we begin to lock down models and project next year, think about that?

Rohan Sivaram: Yes, Brad, we’re not early guiding for fiscal year ’25. I’ll say which is very consistent with most other software companies out there. But what I’ll tell you is the momentum of the business in Q3 was solid. We entered Q4 with a similar momentum which shows up in our guidance. So — and we also spoke about stabilization in our NRR. So as you kind of look at the second half of the year, in general, we feel good with where we are. So to your question around looking ahead, I’ll kind of reiterate a couple of points that Jay made. I mean, in general, you’ve heard us talk about it. DSP has been a big focus for us this year. With respect to the number of product innovations we’ve had this year has been, I would say, the highest in the history of the company.

And we’re seeing good adoption with respect to our DSP products. 2025 is where we’re expecting monetization. And when you look at the different DSP products, all 3 of our DSP products, be it Connect, be it governance or be it stream processing, they’re in their earlier stages of their S curves. So that’s one area of growth driver as we look ahead. The couple of others that Jay briefly touched on was Gen AI, although timing — exact timing is still a TBD but we feel that in general, streaming as a category will benefit from Gen AI. And then we have a couple of other ones like Table Flow and FedRAMP which are also lower in the list but again, growth drivers. So overall, a decent amount of growth drivers as we look at next year. But again, we’ll be providing a more formal guide for 2025 in our Q4 call.

Shane Xie: We’ll take our next question from Jason Ader with William Blair, followed by Wells Fargo.

Jason Ader: My question is on the Q4 guide. It implies a sequential growth rate on the revenue that is well below kind of typical seasonal patterns, if I look over the last few years. I know you’re a bigger company now. But is there something specific to call out in Q4 that might cause a lower sequential growth rate than normal. I just wanted to unpack that a little bit.

Jay Kreps: Yes, you want to take that, Rohan?

Rohan Sivaram: Yes, happy to. Jason, good to see you as well. Yes. So for our Q3, I’m assuming your question is specifically around the cloud revenue.

Jason Ader: Total subscription revenue of like a 2.4% sequential versus the year ago and the year before that was 12% in Q3 to Q4.

Rohan Sivaram: All right. Yes. So there are a couple of puts and takes as you look at. I mean, obviously, when you look at our Q3 performance, we feel very solid and Q4 guide is also solid. I called out 2 drivers in last call. One was, in general, Confluent Platform business tends to be lumpy and purely because about 20% of revenue is recognized upfront and the timing of some of these larger deals can have an impact. So that’s something that I called out in our last call. And the second, I would say, a smaller driver is the onetime benefit that we spoke about in Q3 makes it a little bit of a tough compare for our cloud business. So when you isolate the cloud business and you take out that onetime benefit, it very much falls within historical trends and more normalized patterns.

So that’s probably the second driver. But in general, if I kind of take a step back, we’re seeing this momentum with respect to some of our larger digital native customers adopting new use cases and adopting DSP. We just don’t want to get ahead of ourselves and we want to be prudent with our outlook as we look ahead.

Jason Ader: All right. And then one quick follow-up maybe for you, Jay. Just talk about the federal business. We’ve heard kind of through the grapevine and some other companies have talked about it but it seems like federal spending was kind of weaker than expected for a lot of companies in Q3. What did you guys see in the federal vertical? I know that’s an important area for you guys. Did you see some of that kind of budgetary pressure that others saw?

Jay Kreps: Yes. Yes. Federal was reasonable and is a decent sized chunk of the business. It’s limited today because it is only Confluent platform. So we’re still kind of waiting to open up the Confluent Cloud side of that, at which point, I think we would see a bigger overall impact, both positive and negative, with the kind of trends you’re describing. So yes, I would say it was not particularly noteworthy. We saw reasonable performance but nothing to write home about.

Shane Xie: We’ll take our next question from Michael Turrin with Wells Fargo, followed by Oppenheimer.

Michael Turrin: Jay, I appreciate the video at the start commentary throughout the call; it’s helpful. I was hoping we could go back to AI and specifically use cases you see there for streaming. Where do Agentic solutions which we’re getting whole host of announcements around fit within that discussion. It would be great to just get your view on how this AI focus we’re getting everywhere impacts competitive positioning for Confluent streaming of the overall DSP landscape.

Jay Kreps: Yes, that’s great. Yes. So there’s 2 primary use cases that we’re seeing around AI. One is really gathering all of this context data for the AI. So all the enterprise data that wants to be used in decision-making. The second is really a Flink use case which is actually taking some of the processing that you want to run as a background task and actually operationalizing that, turning it into something that kind of runs continuously. Every time something happens to the business, it reacts or processes that. That’s very much the kind of agentic background work that you’re talking about. That’s earlier in terms of what we’ve seen. If I look at kind of customer adoption patterns, I would say that, that whole category of use case is a little earlier.

I think it’s very dependent on the quality of the models. We think that’s going to be a big thing over time and we think we’re very well positioned to do it. I think ultimately, the goal for AI isn’t just to build these chatbots. It’s actually to take some of the background work in the company and turn it into something that just happens. And maybe there’s some amount of human oversight or maybe there’s not but it’s something that’s just kind of working in the background. If you think about what does that translate to in terms of the infrastructure or computational model, it’s very much stream processing. If you think about the kind of more work that humans do, they’re sitting there and they’re answering e-mails that they’re reacting to new customer orders that they’re doing whatever it is that kind of background processing.

And so I think having something which takes that directly integrates the LLM models with some of the AI model inference work that we announced at current allows you to run that in a way that’s parallelizable, that’s fault tolerant, that’s scalable, that’s understood and that integrates this context data that you can gather, I think it’s very compelling. So, I would say that — I would think of this as a couple of directions of growth around AI. One is selling to the AI companies, that’s kind of off and running. The next lump is these enterprise use cases around RAG. I think that’s going well and will be durable. The last is this kind of agent use cases which is the most nascent but I think actually may be the biggest opportunity. And probably the thing that streaming is the most well suited to where it’s kind of the natural model for running that thing.

That’s — I would describe that as unproven but we’re very excited about it.

Michael Turrin: That’s great. Rohan, just given that ’25 shaping up is a fairly significant product cycle for Confluent as you layer on work stream other capabilities. I’m wondering how you think about or how we should think about margin progression as these newer efforts layer on? Can you keep the efficient growth motion going forward, alongside the innovation? Or just how are we thinking about the balance between those 2?

Rohan Sivaram: Yes. Michael, when you really look at the last 8 or 10 quarters for us, we’ve improved our operating margins by more than 40 percentage points. So our philosophy around growth and profitability or efficient growth is really part of the DNA of the company. So every key decision that we make is based on ROI-based thinking. And that will continue not only into ’25 but ’25 and beyond. So that’s not going to change. So as we think about next year, we’ll continue to have the same philosophy that is growth and profitability, how can you drive efficient growth and we’re not going to basically take our eye off the ball on that front.

Jay Kreps: Yes. I think that’s exactly right. And one point I would make that’s specific to that. I think what Rohan said that this is kind of a discipline in terms of looking at efficiency throughout the company. I think it’s a very important point. But I think you asked specifically around as we’re adding these additional product capabilities, how does that impact things. I think it’s important to understand that the investment around the new products, including the training of salespeople, getting people to be able to sell these things, that’s all happening now, right? So you kind of — you put in all the effort, you build a cloud platform that can operate many different product areas. You make the investment in engineers to build out the capabilities.

You put in all the tools to be able to track and drive revenue targets around these different things. You train up the sales force, you spend a lot of time on these things that are still nascent businesses, right? And then as those come to scale, that’s obviously a very positive thing. But to some extent, we’re kind of bearing that cost now. We feel good about that trajectory. And obviously, as these are — as DSP is a larger contributor to the overall revenue numbers, then those investments are less kind of an added weight and more a natural part of the business. So I don’t know if that’s clear but that’s the way I would think about it. It’s not like we’re thinking about, oh, we’re going to start investing next year to make that true. We have been investing for some time on the R&D side for several years, on the go-to-market side, certainly heavily this year.

Shane Xie: We’ll take our next question from Ittai Kidron with Oppenheimer, followed by Moskowitz.

Ittai Kidron: Jay, I’m going to start with you, more of a go-to-market question. There’s 2 parts for it. First of all, do you still see a lot of low-hanging fruit in improving your go-to-market motion? And maybe tied to that, clearly, with DSP becoming a bigger focus for you, how do you think about the changes you need to implement in comp next year to better perhaps fine-tune the effort around this?

Jay Kreps: Yes. Yes. Look, we feel really good about the set of changes we made. So this year was a very aggressive adjustment to orient the cloud business around consumption to drive broader reach to enable us to actually incentivize some of these individual DSP components at a higher rate. That is — I would call that a radical shift, right? Heading into next year, yes, I don’t think we need any kind of radical shifts. We’ve got some — we do some tuning each year where we look at can we tweak this thing? Can we tweak that thing? Well, there’s plenty of that, that we’ll do. But I think it will certainly be a smaller thing. I do think that as some of these DSP components come to maturity and scale, it does open up a new motion around directly landing use cases.

If you think about Confluent, maybe the motion has been primarily kind of open source upsell where you take people who are interested in Kafka, there’s better Kafka, there’s other components around Kafka. I think as we have this full set of capabilities to capture, stream, process, connect, transform and govern real-time data, there’s suddenly a whole set of business problems you can go after much more directly. And so I think that opens up another vector for the team to kind of attack and expand within a lot of these customers. I think we’ll still land in areas where there’s interest around Kafka but I think this kind of gives you another way of going to market. And so I think that’s something we’ll build over time, not comp related but just makes it — gives us another path into use cases.

Ittai Kidron: Excellent. Second question is on the win rates. I think you mentioned that win rates have actually increased this quarter, if I’m capturing the comment correctly but you also commented that the win was against small vendors is more than 90%. Is that the specific area they improved? Or are they also improve the general.

Jay Kreps: Yes, I think we’re seeing both that we saw very strong win rates overall. And then there have been questions on this point. I think we specifically called out, hey, like really strong performance well above 90% against start-up competition.

Shane Xie: We’ll take our next question from Gregg Moskowitz with Mizuho, followed by D.A. Davidson.

Gregg Moskowitz: The net new Confluent cloud ARR is higher than any other Confluent quarter that we’ve ever seen and that includes your seasonally strong Q4 periods. And it’s an impressive quarterly performance regardless. But Rohan, can you say whether or not would be a record net new ARR quarter for Confluent if we were to exclude that onetime revenue benefit that you called out?

Rohan Sivaram: Yes, Gregg, the onetime revenue benefit that we called out, what I said was if you take that out, we still handily beat our expectations. And when you really double-click into the performance which is the vast majority of the driver of the performance, is stabilization in the digital native segment from a consumption standpoint and the net new use case of DSP adoption for larger cloud customers. So we’ve not broken down that onetime exactly. But what I can tell you is if you adjust for that, we still very handily beat our expectations and the primary drivers were the first one that I called out. I hope that helps.

Gregg Moskowitz: Okay. And then you spoke earlier, Jay, about work stream and potential open source conversions. But I’m wondering if you also foresee many content platform customers adopting Work stream potentially over the medium term, whereas perhaps they otherwise wouldn’t have moved away from their on-prem deployments anytime soon.

Jay Kreps: Yes. Yes, there’s certainly an opportunity for conflict platform customers that are kind of self-managing the cloud to have kind of a progression towards a more fully managed cloud type offering. And we see that as a positive thing as well. It’s not the initial target set of customers but yes, over time, that may be appealing.

Shane Xie: We’ll take our next question from Rudy Kessinger with D.A. Davidson, followed by Needham.

Rudy Kessinger: Rohan, first with you, very strong gross margins. It looks like if I exclude low figures from that onetime revenue, it’s only 20 to 30 bps impact. So is 82% kind of a good run rate going forward? And just what do incremental gross margins look like on the DSP products as revenue from those start to come in more meaningfully next year?

Rohan Sivaram: Yes, sure. Happy to take it. Rudy, the gross margin profile for the business, obviously, record gross margins when you look at subscription gross margins. So there are obviously 2 components, Confluent Platform, software-type gross margins, fairly consistent over a period of time. So the variable is in the cloud side and we’ve consistently improved our cloud margins over time. And I mean, I put it in maybe 3 categories, right? One category is, in general, with volume and with more scale, you get efficient. So that’s one. The second area is there’s been this focus around multi-tenant and as more of our business becomes multi-tenant which includes the DSP side as well, a lot of the DSP products, right, that’s going to be a tailwind to gross margins, right?

So those are probably 2 drivers. And when you kind of take a step back and look at where we are operating, we’re actually operating well above our thresholds that we put out there. So feel good about our gross margin profile and we will obviously keep an eye on it as we look ahead.

Rudy Kessinger: Okay. And then as a follow-up, I appreciate the commentary on win rates. I guess does that comment on win rates, I guess, do you include. Do you factor in renewals where a current customer considers a small start-up or going back to open source as well when you calculate those win rates in. And if not, could you just maybe comment on gross churn or gross retention trends over the last few quarter.

Jay Kreps: We do include renewals in that. And I do include renewals but it would be equally high, not including renewals. So it’s not like we’re patting it with renewals.

Shane Xie: We’ll go to Needham Mike Cikos for our next question, followed by Cowen for a last question.

Jay Kreps: Just to comment really briefly on what Rudy said. To underline a little bit of what Rohan was saying. Yes, we don’t anticipate the DSP growth to be a significant headwind on margins. the products are largely multi-tenant and we see that as a positive factor. We’ve taken cost into consideration when designing those things. So I know often new cloud products kind of come in upside down and then eventually write themselves. We don’t think there’ll be a big aspect of that with what we’re doing.

Mike Cikos: I just wanted to cycle back to the strength that we’re calling out this quarter and into October now for those digital native customers. When did you actually start to see that behavior shift just because I know if we cycle back a quarter ago, that increased focus on optimization was bleeding from June into July, right? So when did that start to show up as far as that cohort that you’re speaking to the digital natives.

Jay Kreps: Yes. Yes, do you want to speak to that, Rohan?

Rohan Sivaram: Yes. I mean it was — typically, Mike, when you look at consumption, it is also a business that’s driven by momentum you did not see this like 1 particular date or where things shifted. But in general, we did see a couple of things. The first is, just Jay called out earlier in the call, some of our larger customers we feel that every customer kind of looks at optimizations for some of our larger ones, that’s actually behind us. That’s something that we saw and that had that progressively through the quarter. And the second piece is around the net use cases and some of our larger cloud customers adopting DSP. That’s been a huge internal go-to-market focus for us. And again, early days but that’s also starting to show up a little by little in the numbers. So that’s how I’d call it. And as I said, it’s more around momentum. And as we exited Q4, some of these trends actually bled into Q4 as well.

Mike Cikos: That’s terrific. I guess, for the follow-up there and it sounds like you’re partially answering this already but trying to get a sense for that recovery in demand for new use cases or DSP adoption coming from digital natives. Is that explained by maybe customers who have been rolling over or pushing out projects that are coming online now, or is it more a function of that go-to-market that you guys have and the transformation you’ve instilled and maybe that’s starting to make more headway into your existing customers?

Jay Kreps: Yes, it’s a combination of both. Customers — what you would see in any of these digital native customers, even going over a long period of time is kind of a sawtooth up and to the right, right? It’s not a pure graph. They tend to make investments and then they tune and they make investments and then they tune. We did see a little bit more tuning in the recent quarter but that — it’s not like that pattern hasn’t occurred. It just — we just sought in more customers all at the same time. Even in that time period, we did see an intention in those customers of making further investments, new projects. So I would say, on the whole, that’s in keeping with what we’ve seen. On the DSP side, there definitely is a comp as well as just intentional focus.

Like we are putting a lot of effort into driving these new products. There’s very specific sales plays. We’re tracking pipeline in a very distinct way. There’s targets and goals around that. It’s very much a new muscle to build. But the consumption comp makes it possible to do this because in the past, we would have goals that were purely on your committed spend. And really, of course, there’s no distinction of what the spend is on. Now we can have multipliers on specific components of revenue that’s Flink or Connect or whatever it is and drive those in a disproportionate way. And so I think that does help a bit in terms of making sure that the smaller products that are kind of newer in the journey still get some focus out of the team.

Shane Xie: We’ll take our last question from Derrick with Cowen.

Derrick Wood: Jay, I’ll start with you. You guys called out strength in financial services. We know this is obviously a big vertical for you but I don’t know if you’ve highlighted it in recent quarters? Anything to call out what drove this? Are you starting to see larger deal activity? Or is there some broader industry trend developing that could be more beneficial in the medium term?

Jay Kreps: Yes. I wouldn’t say it’s a sudden shift or even a 1-quarter thing. I mean I think there’s been a broad-based build in financial services going back some years. And I think we just thought it was worth highlighting kind of the point that we’ve got to, where this is now a very substantial data platform and the largest financial services institutions in the world. And there’s really exciting stuff happening. And this is — it’s not just that it’s big in terms of volume. I mean this is powering some of the most critical systems that they have. in many of these organizations, they’re actually starting to think about how streaming not just enables them to share data internally but how it enables them to connect into many of the other institutions that they share data with on a regular basis.

There’s kind of that larger economy within that industry. And we think that’s a really positive thing. So yes, it’s going very well. We called out the kind of ARR and penetration in the largest customers. We expect that to continue. No sign of slowdown. There’s not a particular catalyst this quarter. I would say it’s more just continued strength. And it makes sense when you think about the nature of those businesses that there’s just a huge amount — they’re the largest spenders on IT and a huge amount of kind of real-time event-driven machinery behind the scenes and as well as a few pushes on the regulatory side towards real-time payments, real-time reporting. There’s definitely some nudges as well as pressure on having a modern customer experience, all of which kind of nudges it forward.

Derrick Wood: Understood. Rohan, maybe one for you. The strength of new customer is obviously very impressive. I think I assume that most of these are open source Copco conversions. But just wondering how you’re thinking about the follow-on expansion. Is this something that could come in a quarter or 2 because of the consumption model or we think of a land and maybe an expand like 9 to 12 months from now?

Rohan Sivaram: Yes. The way I think about the new customer is, of course, the total customers, I think that’s top of funnel. And then how successful we are progressing these customers to $100,000-plus ARR, $1 million plus ARR is something that we internally focus quite a bit as well. So on the total customer numbers that you mentioned, it’s a combination of two things, of course, like as part of our consumption transformation, there’s been a big focus around landing new customers but not just new customers landing very high-quality logos. And I called out some examples there. right? So that, coupled with our product led, the P&G motion which is driving some of this customer growth on top. But that can obviously vary quarter-over-quarter.

So the best way to look at that is over a 12-month period. And when you look at the first 9 months, you’re right, I mean, we are very happy of the momentum that we’ve seen. When you look at $100,000-plus ARR customers and $1 million-dollar plus ARR customers, both those cohorts get see good consistent growth. And as a reminder, $100,000-plus ARR customers account for greater than 85% of our revenue. So overall, the focus is across all 3 but we are pleased with the progress that we are making.

Shane Xie: This concludes our earnings call. We know it’s easy to earnings. We appreciate so many of you joining our call today. Have a nice evening.

Jay Kreps: Thanks, everyone.

Rohan Sivaram: Thank you.