Snowflake Inc. (NYSE:SNOW) Q4 2026 Earnings Call Transcript

Snowflake Inc. (NYSE:SNOW) Q4 2026 Earnings Call Transcript February 25, 2026

Snowflake Inc. beats earnings expectations. Reported EPS is $0.32, expectations were $0.2725.

Operator: Good day, ladies and gentlemen. Thank you for joining today’s Snowflake Q4 FY ’26 Earnings Call. My name is Tia, and I will be your moderator for today’s call. I would now like to pass the call over to your host, Katherine McCracken, Head of Investor Relations. Please proceed.

Katherine McCracken: Good afternoon, and thank you for joining us on Snowflake’s Fourth Quarter Fiscal 2026 Earnings Call. Joining me on the call today are Sridhar Ramaswamy, our Chief Executive Officer; Brian Robins, our Chief Financial Officer; and Christian Kleinerman, our Executive Vice President of Product, who will participate in the Q&A session. During today’s call, we will review our financial results for the fourth quarter fiscal 2026 and discuss our guidance for the first quarter and full year fiscal 2027. During today’s call, we will make forward-looking statements, including statements related to our business operations and financial performance. These statements are subject to risks and uncertainties, which could cause them to differ materially from our actual results.

Information concerning these risks and uncertainties is available in our earnings press release, our most recent Forms 10-K and 10-Q and our other SEC reports. All our statements are made as of today based on information currently available to us. Except as required by law, we assume no obligation to update any such statements. During today’s call, we will also discuss certain non-GAAP financial measures. See our investor presentation for the definitions of the non-GAAP financial measures and a reconciliation of GAAP to non-GAAP measures and business metric definitions, including adoption. The earnings press release and investor presentation are available on our website at investors.snowflake.com. A replay of today’s call will also be posted on the website.

With that, I would now like to turn the call over to Sridhar.

Sridhar Ramaswamy: Thank you, Catherine, and thank you all for joining us today. This past year has been transformative for every business. A year ago, we were talking about the promise of AI. Today, the promise is real, and Snowflake sits at the center of the enterprise AI revolution. Across the market, AI is reshaping the software landscape, redefining categories and competitive dynamics. In our view, this is creating a clear separation between systems that demonstrate intelligence and platforms that can deploy it safely and at scale. The winners will be the platforms that combine trusted enterprise data, govern business metrics, secure execution and broad model choice and make all of it easily. That’s exactly what Snowflake was built to do.

We deliver the data foundation enterprises rely on across clouds and across data types with the performance, reliability and operational simplicity required for mission-critical workloads. As AI agents become central to how work gets done, those same capabilities become even more valuable because agents are only as powerful as the data they can access and the governance and security that surround it. You can see that leadership in what we shipped this year. With Snowflake Intelligence, we brought enterprise-grade agent capabilities directly to business teams. With the general availability of Cortex Code, we extended that to builders, accelerating the entire data life cycle and helping customers move faster from development to production. Most recently, we expanded Cortex Code CLI to encompass data systems as we work towards simplifying how all of them are used in practice.

The general purpose agency capabilities of Cortex Code CLI, combined with our AI-ready data on Snowflake are already driving meaningful operational impact just weeks after launch. Snowflake Intelligence and Cortex Code are meaningful steps in Snowflake’s evolution on the platform where enterprises govern and analyze their data to the platform where they build and run AI-native applications and workflows. Turning to our results. Product revenue in Q4 grew 30% year-over-year to reach $1.23 billion. Remaining performance obligations totaled $9.77 billion with year-over-year growth accelerating to 42%. Our net revenue retention was at a healthy 125%. Thanks to AI, we are both scaling revenue and becoming operationally more efficient. Fiscal ’26 non-GAAP operating margin reached 10.5%, expanding more than 400 basis points year-over-year, reflecting our continued focus on operational rigor.

Stock-based compensation declined meaningfully from 41% of revenue in fiscal ’25 to 34% in fiscal ’26, and we expect it to further decrease to 27% of revenue in fiscal ’27. This year’s results are a testament that the AI Data Cloud continues to deliver tremendous value to our more than 13,300 customers across every stage of the data life cycle. Built with deep product cohesion, Snowflake is easy to use, seamlessly connected for collaboration, grounded in the security and governance enterprises trust. As we innovate, we remain maniacally focused on driving great business outcomes for our customers. That focus is why leading organizations continue to choose Snowflake as the foundation for their data and AI strategies. We added 2,332 net new customers this year, and we are seeing more and more businesses move over to Snowflake.

Seagate, for example, is modernizing its data foundation to better support its mission of powering data-driven innovation at global scale. By consolidating a massive data environment on Snowflake, the company is moving away from legacy infrastructure onto a platform built for scalability, reliability and predictable cost, enabling teams across the business to access high-performance AI-ready analytics and make faster, more informed decisions. Our core business remains strong, and AI is expanding workloads across our platforms. Capital One is a great example of how we are deepening our relationships with key customers. As Capital One scales its AI initiatives, they are leveraging Snowflake to unify proprietary data, optimize engineering workloads and deliver AI-driven analytics across the enterprise.

Key to our growth is the strength and momentum around our AI products. This quarter, we delivered the largest sequential increase in accounts using AI, bringing the total to more than 9,100 accounts. And in just 3 months, Snowflake Intelligence has scaled from a nascent offering to an essential capability for over 2,500 accounts, almost doubling quarter-over-quarter. For example, Toyota Motor Europe, a global automotive leader, is leveraging Snowflake Intelligence to revolutionize its operations. By enhancing enterprise search with easy-to-use knowledge chatbots and streamlining contract management through Document AI, Toyota has fundamentally shifted its development time lines, reducing AI agent deployment from months to weeks, creating a significant competitive advantage.

And United Rentals, the global leader in equipment rentals, is using Snowflake Intelligence to power a new business intelligence agent that helps teams across more than 1,600 branches get real-time answers from their financial and operational data using natural language. The agent enables faster, more consistent decision-making for frontline managers. United Rentals is also using Snowflake’s Cortex Code to accelerate the development and testing of additional AI agents, scaling trusted intelligence across the business. And that’s just the start of what Cortex Code can do. It’s a truly transformational coding agent that’s already helping over 4,400 customers build and scale AI-powered applications and massively accelerating their ability to deploy production-grade AI.

The Chief Technology Officer of one of our partners, Evolv Consulting, described Cortex Code’s impact on their business saying, “20 days, 21,000 operations, over 600 hours of work delivered. That is 16 work weeks compressed into less than a month. Development cycles that used to require extensive research, trial and error and debugging now flow naturally through AI-assisted iteration. We’re using this capability to accelerate how we bring new workloads on to Snowflake for our customers”. Cortex Code meaningfully expands the surface of AI development on our platform and reinforces Snowflake as the enterprise AI foundation. As we look forward, we continue to see immense opportunity to support enterprises across their data life cycle, and we’re innovating rapidly opportunity.

This year, we launched over 430 product capabilities, underscoring the strength of our product velocity. We are broadening how data enters and flows through Snowflake. Snowflake OpenFlow now generally available, makes it easier than ever to bring in structured, unstructured, batch or streaming data into the platform. We have also deepened how applications are built on Snowflake, now generally available. Snowflake Postgres is a world-class operational database built directly onto the Snowflake platform, enabling developers to build and run production-grade transactional applications with the performance, reliability and ecosystem of Postgres fully managed and governed within Snowflake. This transforms Snowflake from a system you analyze with into a platform that you build on.

And our recent acquisition of Observe, a market-leading observability platform, extends the value that Snowflake can deliver. By integrating observability directly with data and AI products, we reduce complexity and enable faster, more reliable operations at scale. This expands our opportunity into the $50 billion IT operations market and positions Snowflake to lead in next-generation AI-powered observability. At the same time, we are strengthening the ecosystem around the platform. Our landmark partnership with SAP is delivering incredible value, helping customers like Expand Energy unite mission-critical business data across their core systems within our AI data cloud. Our deepened partnership with Anthropic is already helping customers like Intercom see significant impact.

Snowflake provides the secure governed data foundation that Intercom’s AI is built on. By applying direct AI capabilities to this data, including their use of Anthropic’s cloud model, Intercom automates customer support at scale. This allows it to handle significantly higher support volumes with greater consistency and lower operational burden, especially for large complex customers. We also recently announced a $200 million expanded partnership with OpenAI. It brings OpenAI’s models natively into Snowflake to help our customers innovate faster while keeping their data secure and governed. And through our partnership with Google Cloud, customers now have access to the latest Gemini models natively within Snowflake, further expanding model choice and availability.

As we innovate, we are scaling efficiently. Work is fundamentally changing, and we are leading this transformation, both within Snowflake and across the industry. In many cases, we are creating entirely new AI native systems built directly on Snowflake. Across our business, Snowflake Intelligence and Cortex Code are already delivering measurable results. Our service delivery team can complete customer projects up to 5x faster, improving response accuracy by more than 25% and compress implementation cycles from days to hours to drive 40% to 50% higher project margins and enabling customers to go live more than 40% faster. We have seen our site reliability engineering investigations that once required hours across multiple engineers now resolved in minutes, dramatically reducing resolution times and further strengthening Snowflake’s reliability.

And we have built agency capabilities that help our sellers prioritize accounts, automate research and generate personalized outreach projected to recoup the equivalent of 90 full-time engineers of productivity this year. Our finance team is working on automating travel and expenses analysis, proactively curbing out-of-policy behavior, an initiative that is expected to drive millions in annual savings. And we’re seeing this transformation within our customers as well. They are leveraging agents not just to analyze information, but to automate complex workflows and in some cases, retiring entire categories of previously used software systems. Take Sanofi, for example, AI-powered workflows built on Snowflake with partners like Elementum are replacing their traditional software systems used for processes like software license and invoice management.

By running these workflows directly in Snowflake, Sanofi is streamlining operations while keeping its data securely within the platform. This is where the enterprise is heading. And we believe Snowflake is uniquely positioned to become the control plane for the agentic era. We’ve built the conditions that make agents safe, scalable and enterprise ready, covering a single enterprise-wide source of truth, governed metrics and shared business definitions, cross-cloud and cross-domain interoperability, built-in security, auditability and governance. Our continued rapid innovation, tight go-to-market alignment and operational discipline are all in high gear to capture this opportunity, and we see a long runway of durable high growth and continued margin expansion ahead.

Now I’ll turn it over to Brian to take us through the financial details.

Brian Robins: Thank you, Sridhar. Q4 was a strong quarter across revenues, bookings and margin results. Product revenue grew 30% year-over-year. Our results were driven by stable growth in our core business and a step-up in growth contribution from AI workloads. We saw no decline in our net revenue retention rate, which remains at 125%. Q4 sales execution was outstanding. Remaining performance obligations accelerated for the second consecutive quarter. We signed the largest deal in Snowflake’s history, greater than $400 million in total contract value and signed 7 9-figure contracts compared to 2 in the same period last year. These strong commitments represent Snowflake’s strategic role in our customers’ long-term data and AI strategies.

And we’ve consistently emphasized durable growth depends on 2 fundamentals: landing new customers and expanding existing ones. We’ve delivered on both. We delivered another strong quarter of new customer wins, adding 740 net new customers, up 40% year-over-year, including 15 Global 2000 organizations. At the same time, we’re proving that we can drive meaningful customer expansion. We now have 733 customers spending more than $1 million on a trailing 12-month basis, growing 27% year-over-year and a record number of customers crossed $10 million in trailing 12-month spend, bringing a total of 56 customers above this $10 million threshold, growing 56% year-over-year. Turning to our margin results. FY ’26 non-GAAP product gross margin was 75.8%.

We are demonstrating that we can scale while driving efficiency. FY ’26 non-GAAP operating margin was 10.5% and FY ’26 non-GAAP adjusted free cash flow margin was 25.5%. Earlier this month, we closed the acquisition of Observe, which we acquired for approximately $600 million in a combination of cash and stock. With Observe’s offering, we’re unlocking new expansion opportunities within our customer base. The impact of the acquisition is reflected in our outlook. In Q4, we used $150 million to repurchase approximately 668,000 shares at a weighted average share price of approximately $225. We have $1.1 billion remaining on our repurchase authorization and ended the quarter with $4.8 billion in cash, cash equivalents, short-term and long-term investments.

Before moving to our outlook, I’d like to share my priorities for FY ’27. First, I see a clear opportunity to drive both growth and operating margin expansion. We are investing in our key growth drivers. As Sridhar relayed, we deployed more than 430 product capabilities to market this year. We’ll continue to expand operating margins as we drive greater efficiency across the business. Second, it’s clear that our go-to-market motion is working. My focus for this next year is on ensuring stability and ongoing excellence. We’ve established a financial framework to support continued product velocity and sales execution. Now let’s look to our outlook for FY ’27. In Q1, we expect product revenue between $1.262 billion and $1.267 billion, representing 27% year-over-year growth.

For FY ’27, we expect product revenue of approximately $5.66 billion, representing 27% year-over-year growth. We expect Observe to contribute approximately 1 percentage point of product revenue growth in FY ’27. As always, our forecast is built on using existing patterns of consumption. There are no changes to our forecast process or our guidance philosophy. Our outlook is supported by continued strength in our core business and further growth in AI workloads. We expect FY ’27 non-GAAP product gross margin of 75%. We’re guiding Q1 non-GAAP operating margin of 9% and FY ’27 non-GAAP operating margin of 12.5%. Our hiring this year will be weighted to the first quarter, reflecting the addition of 178 employees from Observe. We expect non-GAAP adjusted free cash flow margin of 23%.

This includes an approximate 150 basis point headwind related to our acquisition. As in prior years, we expect our bookings will continue to be weighted to the fourth quarter, and we expect next year’s non-GAAP adjusted free cash flow seasonality to mirror FY ’26. Finally, we’ll host an Investor Day in conjunction with our Summit Conference the week of June 1 in San Francisco. If you’re interested in attending, please e-mail ir@snowflake.com. With that, I’ll pass the call to the operator for Q&A.

Q&A Session

Operator: [Operator Instructions] The first question comes from the line of Sanjit Singh with Morgan Stanley.

Sanjit Singh: Congrats on reasserting 30% product revenue growth in Q4. I have 2 questions, starting with Brian and then hopefully for you, Sridhar. Brian, on the guide for fiscal year ’27, it basically implies sustained growth around 27% throughout the year. And just sort of just want to get your perspective on the durability of that 27%, given that it’s a consumption model, sort of a sustained growth off of a really good year this year onto just sort of the confidence in that. And then for Sridhar, as we go into the first full year of Snowflake Intelligence, and an expanded product portfolio. I was wondering if you can give us a sense of where we are in terms of momentum with the areas of the business outside of the core. I think we got an update on the data engineering revenue run rate or growth rate several quarters ago. So I wanted to get an update on that and where we sort of stand with the AI portfolio exiting this year and going to fiscal year ’27.

Brian Robins: Thanks, Sanjit. I’ll go first. From a guidance perspective, we guide based on the Observe customer behavior up until really the point of earnings. And the guidance, if you sort of double-click into it this year, it’s really based on the high stable growth that we see in our core business. It’s also the growing contribution from AI workloads. And finally, we called out in the prepared remarks, there’s 1 percentage point of growth from our Observe acquisition. I’ll turn it over to Sridhar for the second part.

Sridhar Ramaswamy: And to just reiterate on top of that, our overall guidance philosophy hasn’t really changed. We continue to be very stable with respect to that. I see products like Snowflake Intelligence now with 2,500 customers as a major driver of growth across all aspects of the data life cycle. I think what products like Snowflake Intelligence, and I never tire of showing every single CXO and CEO that I meet Snowflake Intelligence on my phone, the ready access that it offers is truly magical to critical business information. And that reinforces the need for enterprises to adopt Snowflake to get their data estates in gear so that they can bring the transformative power of things like Snowflake Intelligence to that data.

The really important thing also to remember about Snowflake Intelligence is that it works fine on all open data. You can build Snowflake amazing agent with using Snowflake Intelligence on data that is sitting in S3, managed by Glue or sitting in other places. Any open data ecosystem is supported by Snowflake Intelligence, and that’s really very powerful. But Cortex Code is the real game changer for us because it is a massive accelerant for every part of the data life cycle. What I mean by that is we can build open flow pipelines to bring in data from complex systems into Snowflake at a fraction of the time that it used to take before. Similarly, building DVD pipelines to run data engineering on that data or to build dynamic tables or debug performance issues with either of these now is again 10x plus faster.

And what’s magical about Toco is also the ability to actually go Snowflake intelligence agents faster. I think that’s the unlock of AI using AI to make things go faster. And we see this, as I said, of having transformative effects on our business. I’ll give you folks an anecdote. One of our partners wrote to us after using Cortex Code CLI and said that all this time, they had been using shovels to dig, and we just gave them bull dozers. Let’s go to the next question, Mark Murphy.

Operator: The next question comes from the line of Mark Murphy with JPMorgan.

Mark Murphy: So the bookings and RPO figures look very robust once again. It looks like the biggest bookings figure in the history of the company actually by a pretty wide margin. I just want to ask first, can you describe the $400 million deal in terms of the customer type because I don’t — it’s a gigantic contract. I just don’t think we’ve heard anything like that. And second, I’m curious if you see some sustainable new drivers kicking in there for bookings, like maybe thinking back on achieving a faster product GA cadence is something you’ve done? Or is this a little more temporary onetime? You had the hiring surge several quarters ago, and I think you’ve been incentivizing reps a little more heavily on bookings this year. So I’m just wondering if you can comment on this.

Sridhar Ramaswamy: I can start. Brian can add on. Bookings and multiyear contracts are a clear indication of the trust that our partners have in their future with Snowflake. And yes, the product acceleration and velocity goes a lot towards convincing customers that we are a platform for the future. We didn’t do anything particularly special in the quarters. Yes, we did adjust the compensation plan to also take bookings into account last year. But in many ways, that represents a reversion back to how things were 2 years ago. And we plan to continue that this year. So it’s very much business as usual. I do think that the $400 million, $400-plus million deal that we signed is an indication of the importance that we deliver to that large financial services customers. We have previously talked about deals in the $250 million range. I think it represents a maturity of Snowflake as a durable provider, not just today of data services, but also into the future. Brian?

Brian Robins: Well said, Sridhar. I would say one of the big contract, over $400 million, it was an existing customer. So it’s already built into the run rate. We did sign 7 9-figure deals as well. And so just to reecho what Sridhar says, it’s…

Mark Murphy: It’s just Q4?

Brian Robins: Yes, just Q4. And just to re-echo what Sridhar says, it’s really a buy-in from our customers on our product road map and AI strategy and the positive business outcomes that we’re delivering for their business.

Operator: The next question comes from the line of Brad Zelnick with Deutsche Bank.

Brad Zelnick: Great. And I’ll echo my congrats. Sridhar, I guess this one is for you. Just coming away from sales kickoff and now the first full year with go-to-market under Mike Gannon’s command, what are you going to do differently in the field to win and drive upside in fiscal ’27?

Sridhar Ramaswamy: Mike’s had a year. He has had a very positive influence on the sales team. But I think what drives momentum for the whole company and absolutely, the sales team are great products that let our sellers, our solution engineers deliver value for our customers. And I have never seen more excitement from our sales force about the products that we create. We have had multiple people. I’ll let Christian chime in because he gets a lot of these accolades. We have had multiple people come and tell us how Cortex Code is absolutely transformational in what people can do with Snowflake. Many folks come and tell us that they have never felt as much excitement about a product that we have created since when the original product was created.

And Christian had a section of Cortex Code Heroes that highlighted their experience. I’ll let him say it since he was the one that ran that project. Yes. Super quickly, like partners, customers and our internal field are all incredibly excited about the results we’re seeing with Cortex Code. The original value prop of Snowflake, which is change what’s possible in terms of ease of use, it’s just gone like 10x with Cortex Code. We showcased a number of instances where people are building pipelines faster, transformation faster, insights faster. And I think we’re only at the beginning of what is possible. One of our sales leaders, who I assure you, would be the last person to declare himself to be a software engineer, built a streamlit application, deployed it on Snowflake and had his team use it.

That’s how easy Cortex Code makes it to use data from Snowflake.

Operator: The next question comes from the line of Kirk Matter with Evercore ISI.

Unknown Analyst: This is Chirag on for Kirk. Sridhar, observability is a big market, right? How does Observe fit into that topography? And what were you seeing in the market and in the company that it made sense to bring them in-house?

Sridhar Ramaswamy: Observability, especially in the world of AI is a big deal. As you point out, it’s a very large market, a $50 billion-plus market, which means that it has many different angles of expertise that go into it. And AI observability, in particular, with agents is a big, big deal. I’m sure many of you use agents, and no one is ever going to accuse a coding agent of not being chatty. There’s just volumes upon volumes of text that then need to be distilled into things like skills, into things like what went right and what went wrong. And so we see this as a critical data problem. And we also see it as a natural extension of our overall role as a data platform. Observe was built on top of Snowflake. So it inherits the excellent data and compute foundation that Snowflake has.

And for a lot of our customers, especially ones with very large volumes of data, observability as traditionally done has become a little bit of a sore point with respect to just the sheer cost of it. And this is where Observe is able to offer a value prop that is factors away, not like 10%, 20%, factors more efficient. And I think those are the kinds of customers that are going to benefit enormously. There is a huge overlap between potential customers of Observe and customers of Snowflake. And it’s really that one-two punch of Observe is built on Snowflake, our job of integrating it is very simple. Observe has an excellent value prop for a large set of customers that also happen to be Snowflake’s customers. That was the — ultimately, the thing that made both Jeremy and the Observe team want to be part of Snowflake.

We are very excited for what’s ahead. Christian, anything to add? That’s great. Let’s move on to the next question.

Operator: Next question comes from the line of Raimo Lenschow with Barclays.

Sheldon McMeans: This is Sheldon McMeans on for Raimo. As you keep making the Snowflake platform more accessible to users and your solutions, you certainly have an exciting opportunity to expand users and consumption, but there is also a risk of maybe sticker shock as AI agents proliferate or new users create more applications and workloads on your platform. So how are you working with customers to help reduce the risk of cycles of strong growth and optimization? And just a little bit on do you feel like customers truly understand kind of the potential consumption uplift they can have as they leverage your agents more?

Sridhar Ramaswamy: It’s a great question, but one that we’ve spent a lot of time thinking about. Let’s make sure we examine the counterfactual for some of the early agent products. Several of them were launched as part of subscription bundles and many companies that offer agent platforms see them as an extension of their existing subscription model. At Snowflake, we charge based on consumption, and we, therefore, offer a very predictable model. I’m also of the firm belief that products have to show value right out of the gate. And I can quote you our personal example where our sales agent replaced a legacy dashboarding system that we were paying close to $5 million for. And so it delivered ROI out of the gate because that moved to be a set of Streamlit and Snowflake Intelligence.

And this is where we feel like we are very, very value aligned, but we are not stopping there. We know that our customers will want price predictability even with Snowflake Intelligence. So we will be launching features like a per user cap on top of Snowflake Intelligence, so they can feel like there is a clear upper limit to how much they can get charged with an agent. We think models like this that are consumption-based with clear user caps and account caps offer the best of both worlds, which is consumption pricing with price predictability. And we’ll continue to innovate rapidly in this area because we think these agents can deliver huge value. And absolutely, we don’t want our customers to have sticker shock. We want to be predictable, and we will provide the controls that are necessary to make for wide deployments of Snowflake Intelligence.

We’ve also done things like integrate Snowflake as a whole with identity providers so that even the task of things like configuring users to be able to use our products like Snowflake Intelligence is a whole lot simpler than ever before. Christian and my vision is effectively that every single employee of every enterprise customer we have should have access to a set of agents that provide them with all the key business details that they need to run their part of the business.

Christian Kleinerman: And only get billed for what they use, which is always correlated with amazing outcomes.

Sheldon McMeans: Very clear. And a quick follow-up. So you certainly talked about your robust AI agent strategy progressing well, but there’s also the idea of other Agentic workflows leveraging Snowflake for critical steps in their process. Can you speak to this latter area and how that’s evolving for you? And do you see that as a fiscal year ’27 growth opportunity? And do you see it mainly going through your zero copy partnerships? Or would there be another pattern that would emerge there?

Sridhar Ramaswamy: Could you clarify your question, please?

Sheldon McMeans: Yes, agentic workflow that’s done in a different platform that maybe needs to leverage some data in Snowflake for a step of the process.

Sridhar Ramaswamy: Well, interoperability has always been a key part of how we operate. And over the past 2 years, Christian and I are very proud of the fact that we have executed flawlessly on an interoperable data strategy. We support Iceberg as a first-class construct within Snowflake. We support Iceberg where we manage the rights. In fact, we recently announced we support Iceberg, where we even manage the block storage so that our customers get the best of all worlds. They get the manageability that they get with Snowflake while feeling confident that another engine can read that data. And what we have done over the past year is use interoperability to drive additional workloads for Snowflake because as I said earlier, you can — we can run SQL queries on any open data through things like catalog link databases.

We can also create agents that are sitting on any open data. And this kind of interoperability is really key for Snowflake to succeed. No customer wants to get into a situation where they cannot — where they do not have options. So we offer interoperability at the storage level. Certainly, people can write SQL and access the data. So we offer interoperability at the JDBC level. And one level above that, we make semantic models available to others. We introduced semantic views, but anyone can read semantic views. And finally, our Snowflake Intelligence agents also double up and can be MCP servers that can be used by other agents as well. And so offering interoperability at every layer of the stack is central to what we do. But we also focus on creating world-class products that lead the way, that are easy to use and set up that make all of this way, way simpler than what anyone else can do.

We don’t see any contradiction between the 2.

Operator: The next question comes from the line of Kozi Leva with Bank of America.

Unknown Analyst: I wanted to ask about the $9.8 billion in RPO, which is growing 42%. I mean, really, really nice there. And so instead of asking you where you saw strength, I’m most curious if you could talk about any air pockets where you were surprised that they didn’t contribute more. Why you think that happened? And how you think those pockets get better from here?

Brian Robins: Kozi, this is Brian. There wasn’t any — we called out the big contract in the quarter for over $400 million in the 7, 9-figure deals, but there wasn’t anything in the quarter that happened where I thought there was areas that we over exceeded or underperformed. Overall, we had a good sales execution quarter. And the RPO, as we talked about a little earlier, is just really points to the business outcomes that we’re driving for our customers and then buying into Snowflake long term.

Sridhar Ramaswamy: Overall, I’m just — I have to add that I’m incredibly proud of our sales team for delivering both across consumption in terms of driving use cases both the wins and our services team for driving more and more of them to production. And of course, what the sales teams got done in terms of these monumental contracts overall, it was a stellar year by those folks, and we are all very grateful.

Unknown Analyst: Yes. And maybe just a quick follow-up here. I wanted to ask about platform usage visibility and predictability. Maybe compare and contrast today versus a year ago, if that has changed at all? And if it has, what has been driving that change?

Sridhar Ramaswamy: Could you clarify your question? What did you mean by platform usage and visibility?

Unknown Analyst: The usage of your platform by your customers, how much more predictable is it today versus a year ago, if at all?

Sridhar Ramaswamy: We continue to have among the most sophisticated systems for consumption prediction. And we obviously calibrate ourselves on how well we do, something like a 0.5% deviation is one part in 200. And for us, that’s sort of a big deal. That’s the level of sophistication that there is and there is similar methodology that is being applied for contract prediction, the TACV prediction as well, and it’s an area where I expect us to see — where I expect us to get better and better over time. And another area that we are actively working on, which has a little bit less predictability is one that goes from use cases to consumption. It’s an active topic for us. It’s a little bit of a research project because we are not always privy to what our customers do.

But we feel very good overall about our ability to model the business and be able to see where it goes. Of course, you also want the surprises that are not part of your models. There is no model that would define the birth of Cortex Code or its adoption by 4,400 customers. We are happy when things like that happen. But when it comes to the core, we are very, very buttoned up among the best teams that I’ve worked with. I worked with a lot of them at Google and other places when it comes to predictability of our business.

Operator: The next question comes from the line of Matt Hedberg with RBC Capital Markets.

Matthew Hedberg: Congrats from me as well. You guys are checking a lot of boxes. You’re accelerating at scale. Sridhar, you went through a number of new AI product announcements. And it looks to me like you’re starting fiscal ’27 organically a couple of points higher than you did at this point last year. So I guess investors want to know, is AI-related products, is that some or all of the kind of the upside that you’re starting to see in this model? Because it certainly feels like you guys are well positioned from these trends. I’m just wondering, is it starting to inflect in the model?

Sridhar Ramaswamy: So the other side of this is that our models are pretty based on observed behavior. And we think that there is a lot of upside. As I said, there’s no way that they can take into account the impact of coco because the historical data simply is not there. We see the benefit of things like coco vividly because we can see how quickly projects finish when they’re being done by our services team. We also see when our partners take these products and are able to do truly transformative things. And you can ask me, am I overusing that word? I point you to a blog post that one of our partners James Dinkel wrote, where he said that they were basically moving their business model as a whole from charging for time to offering fixed fee services.

And a lot of that predictability came because they use Cortex Code to drive the vast majority of the migration. So we see a lot of upside to where the business can go. And on top of this, part of what we have learned even over the past few weeks with Cortex Code is the impact that it can have on every function within Snowflake. Our product managers now have their own version of this to be able to predict — to be able to look at everything from what are the launches coming out next week or what are the bugs that have been filed against their products. There’s even someone that wrote a Christian feedback bot to give them feedback about how Christian would react to a product proposal. The level of innovation that we are seeing across the company is pretty inspiring, and that gives us a lot of confidence about how we approach the year.

Matthew Hedberg: I’ll just squeeze a quick one in.

Brian Robins: Please go ahead. I was just going to add on to what Sridhar talked about prior. Go ahead, Matt.

Matthew Hedberg: You can finish your answer, Brian. I was just going to wonder, it looked like gross margins are down about 1 point this year. And I’m curious with all the investments that you’re making, do you feel like mid-70s is kind of a stable place for kind of gross margins, especially as we look a couple of years forward?

Brian Robins: Yes. Great question. One of our objectives when we launch new products is really, first and foremost, is to build great products. Two, we want to make it easy to use; and three, we want to drive revenue after that. Once we get there, we’ll look at optimizing the margins for that. We have launched a lot of new AI products. The margin profile for those right now aren’t as high as the core business, but we’re offsetting that by finding more efficiencies in the core business. And so that’s really sort of the component of that. We’ll do what’s right to drive growth, and we’ll balance it all the way down the line at the operating margin level.

Sridhar Ramaswamy: And things like margin improvements are coming both at the gross margin level, but definitely also at the company level to just tell you folks about a couple of projects that we did that have had a big impact. One of the folks basically optimized all our free pools across all our deployments using AI because they got way better visibility into that data. That actually. Yes. Free pools, basically, we have to maintain free pools of compute so that our customers don’t have to wait when they want to spin up a new warehouse and somebody found out a very clever way to look at the data and to optimize it or we’ve done a number of things around things like storage life cycle policies, when does the table need to be in nearline storage versus more like facial storage and things like that.

So there are a lot of wins to be had with AI, both above the gross margin line, but definitely at an operating margin line as well. To be honest, it’s a matter of prioritizing what you put your time into because the world is so rich with opportunity.

Brian Robins: And Matt, just to emphasize that point, just in fourth quarter, we saw a lot of benefit with AI that we had a small reduction in force and about 200 people in the company were impacted. So if you look at our fourth quarter net adds on a headcount basis, we only added 37 people. So AI has really changed the framework for investing in growth. It’s no longer tied to headcount.

Operator: The next question comes from the line of Brent Thill with Jefferies.

Brent Thill: All the SaaS names are selling off on the big AI labs taking the stack, as you know. I guess when you think about the advantage you have with the platform of having Gemini, OpenAI and Anthropic available natively, first, do you think your customers understand that yet? And second, I guess, are you seeing that show up in demand given that you have all 3 of the top supported natively?

Sridhar Ramaswamy: I think it’s useful to step back and look at the impact that AI as a whole is having on software. We spend a lot of time looking at this. We live this — and our take is that overall, the winners are going to be the companies that provide that single source of enterprise truth. No AI model is going to help you if there are 4 sources of the truth. Similarly, having built-in security, auditability, trust or even governance over access, who can access what data set is critical. Obviously, you do need the best models, but there are at least 3, if not 4 best model providers right now, and we work with all of them. And I think our secret sauce, which has existed since the beginning of the company is packaging all of this into a cohesive product that is easy to use.

And you see this play out with things like Snowflake Intelligence and Cortex Code working together which is Snowflake Intelligence is a pretty cool product, but Cortex Code makes it 4 to 10x faster to be able to deploy those agents. I think we are really seeing a lot of nice synergies come together as we go into this journey of agent AI. And it is this combination of capabilities plus the fact that we have always been trustworthy stewards of all enterprise information that I think make us a great party for every single enterprise to be working with.

Operator: The next question comes from the line of Ryan Weiss with Wells Fargo.

Ryan MacWilliams: Just excited to see the progress around Cortex Code, and it seems like you’re combining the best of what AI can do today along with the best of Snowflake. As it makes it a lot easier to build agents on the Snowflake platform, it seems like there’s a lot of different vendors that are trying to be the place for users to build agents. So from a [indiscernible] perspective, what do you think are some of the advantages that Snowflake has to be the best place for users to build agents? And then have you seen any increase in query volumes from Cortex Code users today?

Sridhar Ramaswamy: Our mission for a number of years has been to be that data platform that makes data easy to get value from. This is what we did when Snowflake first came out. This is what we have always been doing. In fact, our motto always has been easy, connected and trusted so that data within an enterprise is easy to use, but also present wherever you need it to be wherever you need it to be present. And it’s that thing that I think — it’s that quality that gives us an advantage when it comes to creating agents. As I said earlier, we are also believers in interoperability. It is perfectly fine if someone wants an agent and be able to use NCP to call into a Snowflake intelligence agent. But I think we are uniquely positioned to be that central place where that 360-degree view is possible for a number of our customers, we are stewards of their most important data, the gold layer, as it is called in analytics.

I think that positions us exceptionally well to also be the ones that are providing agents for accessing that data. And we are heavily leaned into technologies like NCP. NCP works both ways. You can use NCP to read from an agent, but we can use NCP to read data from other systems. And we are beginning to see use cases like that come alive as well. And we have done a number of studies. Snowflake Intelligence absolutely drives more usage, more queries. And — but we tend to focus on what’s the value that we are creating. At this point, I’m slightly indifferent about whether we get more of Snowflake Intelligence revenue from running a query or from running the model. It’s all about creating amazing experiences and making it easy to do so. Christian?

Christian Kleinerman: We definitely see the telemetry activity on the platform being increased based on the ease of use that both Snowflake Intelligence and Cortex Code brings.

Operator: The next question comes from the line of Alex Zukin with Wolfe Research.

Aleksandr Zukin: Maybe Sridhar, a quick one for you and then a follow-up for Brian. Last quarter, you spoke to kind of how January and February consumption trends would be the most important to determine the fiscal year guide. Maybe just talk specifically about kind of what you saw post holiday in January and specifically even coming out of February that give you the confidence on what looks like a stronger guide this time versus last year? And then I’ve got a quick follow-up for Brian.

Sridhar Ramaswamy: Well, Brian did say earlier that when we guide, we try to take every ounce of data possible into that guide. That’s what we have done. And we also clarified that the guidance process is a pretty strict one that focuses on historical information and our ability to reliably predict the future. So in that sense, it is taking everything into account. And if you were to ask me what’s the difference between last year and this year, at the beginning of last year, Snowflake Intelligence was a glimmer in our eye. And 1 year later, not only did we launch Snowflake Intelligence and get it adopted, we’ve also — we are also being at the forefront of how you use agentic AI to massively accelerate how a data platform is being used. I think all of that is going to culminate into how we perform this year. But as far as the guide is concerned, it is very much about using every bit of data that we have until this moment, Brian?

Brian Robins: 100% correct. What was your second follow-up question?

Aleksandr Zukin: Yes, I was just going to ask if any update on the Snowflake AI ARR and then the free cash flow margin guide, obviously, digesting the Observe acquisition, maybe just the puts and takes there and how we should think about that trajectory.

Brian Robins: Yes. Just on free cash flow overall, the seasonality will follow prior years. We collect the majority of our cash in the fourth quarter. It’s been greater than 16% in the fourth quarter for the last 2 years. Observe, we guided to 23%. Observe was a 150 basis point headwind. That’s included in our numbers. The revenue is included, the op margins included as well as the free cash flow. And then we just wanted to give guidance that we felt comfortable with that we can perform against.

Operator: That concludes today’s Q&A session. I will now hand the call back over to Sridhar for closing remarks.

Sridhar Ramaswamy: Thank you, everyone. Snowflake remains at the center of the enterprise AI revolution, and we see significant opportunity ahead. To recap, AI has moved from promise to reality, and Snowflake is built to win this era by combining trusted enterprise data, governed metrics, secure execution and broad model choice so that customers can deploy AI and agents safely at scale. We are rapidly transforming from the platform for governing and analyzing data into the platform where customers build and run AI native applications and workflows, making it easier for both business users and builders to go from ideas to production. This strategy is working. Our rapid pace of innovation and strong go-to-market execution are driving continued product revenue growth, and we see a long runway of sustained durable growth ahead. Thank you.

Operator: That concludes today’s conference call. Thank you. You may now disconnect your lines.