Good afternoon. Thank you for attending today's Snowflake Q2 and Fiscal Year 2025 Earnings Call. My name is Cole, and I will the moderator for today's call. All lines will be muted during the presentation portion of the call with an opportunity for questions and answers at the end. I would now like to pass the call over to Jimmy Sexton, Head of Investor Relations. Please go ahead.

Good afternoon, and thank you for joining us on Snowflake's Q2 fiscal 2025 earnings call. Joining me on the call today is Sridhar Ramaswamy, our Chief Executive Officer; Mike Scarpelli, 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 second quarter fiscal 2025 and discuss our guidance for the third quarter and full year fiscal 2025. 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 are 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 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, 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.

Thanks, Jimmy, and hi, everyone. Thanks for joining us today. As you've seen by now, we had another strong quarter beating guidance and increasing our FY '25 product revenue expectations. I'm really proud of the team and how we accelerated our innovation pipeline. And our product delivery momentum continues to be really strong. In the first half of this year alone, we brought as much product to market as we did all of last year. We are making Snowflake the best cloud for computation, collaboration, and application on all data. And we are leveraging the power of AI to make all of these easier to create, maintain and use. This is what our team is aligned around. And I can tell you, that our customers are adopting the new capabilities at an incredible base. As I said last quarter, I have three key areas on personally focused on, listening to and learning from our customers, fueling innovation and product delivery, driving execution and alignment within our go to market team. And in Q2, we delivered on all fronts, which you can see in our results. Product revenue for the quarter was $829 million, up 30% year over year. Remaining performance obligations totaled $5.2 billion with year over year growth accelerating to 48%. Given the strong quarter, we are increasing our product revenue outlook for the year. Companies like Capital One, NBCU, Petco, Pfizer, Snapchat and Western Union are all relying on Snowflake to help them fuel their businesses. I'm really encouraged by the strength of our core business under rapid progress we have made on the AI front. I'm very optimistic about where we are going and the opportunities we have in front of us to deliver for our customers. In fact, the more I'm with our customers, the more…

Thank you, Sridhar. Q2 product revenue grew 30% year-over-year, totaling $829 million. Financial services and technology verticals drove growth in the quarter. We continue to see signs of a stable optimization environment. Our largest customers are contributing sequential product revenue growth in line with historical patterns. We delivered strong bookings in the quarter. RPO grew 48% year-on-year to reach more than $5.2 billion. We signed two nine figure deals in the quarter. Earlier this year, we announced FY 2025 sales incentives that would prioritize consumption and new customer acquisitions in order to drive consumption sales reps prosecute new use cases and sell new product features. We believe this focus will convert into meaningful revenue overtime. Our new customer acquisition motion is ramping. We expect it to have a more material impact in FY 2026. In Q2, non-GAAP product gross margin of 76% was down slightly year-over-year. As mentioned on our prior call, we are incurring GPU related costs and Arctic fulfills customer demand for our newer product features. Non-GAAP operating margin of 5% exceeded our guidance, benefiting from revenue outperformance as expected. Non-GAAP operating margin is down year on year, as a result of R&D and go to market investments. Our non-GAAP adjusted free cash flow margin was 8%. We continue to see approximately 80% of our customers paying annually in advance. We ended the quarter with $3.9 billion in cash, cash equivalents, short-term and long-term investments. In Q2, we used 400 million to repurchase three million shares from our original two billion repurchase plan we have $492 million remaining through March 2025. Our Board of Directors authorized the repurchase of an additional $2.5 billion under our stock repurchase program through March 2027. This allows us to use our cash balance and expected free cash flow to manage dilution over this period. Our share count guidance does not include the impact from the stock repurchase. Now let's turn to our outlook. The forecast product revenue based on observed behavior, our FY 2025 guidance includes contribution from Snowpark as previously stated. Our guidance does not include material contribution from a new product features. Our forecast does include revenue headwinds associated with performance improvements. For Q3, we expect product revenue between $850 million and $855 million. We are increasing our FY '25 product revenue guidance. We now expect full year product revenue of approximately $3.356 billion, representing 26% year over year growth. Turning to margins. For Q3, we expect 3% non-GAAP operating margin. We are maintaining full year margin guidance. For FY '25, we expect approximately 75% non-GAAP product margin, 3% non-GAAP operating margin and 26% non-GAAP adjusted free cash flow margins. With that, operator, you can now open up the line for questions.

[Operator Instructions] Our first question is from Keith Weiss with Morgan Stanley. Your line is now open.

Excellent. Thank you, guys for taking the questions and congratulations on a solid quarter. Really good to hear about the optimization starting to normalize. You guys are - it seems to be settling into just about like a 30% product revenue growth rate over the past couple of quarters. There was a lot of concerns coming into this quarter about impacts from Iceberg tables, there was concerns that accrued during the quarter about the data leakage. That wasn't your fault, but it definitely was a marketing headwind. There was concern about the CrowdStrike cybersecurity incident may be impacting consumption. Were any of these outsized impact, were any of these additional impacts on the consumption in the quarter versus what you guys were expecting when you originally gave us the guide?

I would say the cybersecurity incident was still - it really had no impact on us at all from a consumption standpoint. And the CrowdStrike outage was minimal. It was a day with a few customers, but nothing of any substance. And it never really impacted us itself. Our production does not rely on Microsoft for that. So we didn't have.

Got it. And then was Iceberg tables in line with your expectations?

You know, Iceberg tables is rolling out. As we said, we had 400 customers that are using it. We haven't seen customers move storage out of Snowflake, but we are seeing a lot of our customer. We mentioned 400 that we know that are actually using Iceberg with the new workloads are trying that out. And we're very pleased with the progress we're seeing there. Storage is still running about 11% of our revenue.

That's super helpful. Thank you, Mike.

Our next question is from Raimo Lenschow with Barclays. Your line is now open.

Perfect. Thank you. Two quick questions. Mike, can you -- the gross margin this quarter were better than what a lot of people have modeled. Can you speak to some of the factors for that? And then Sridhar, if you like that around the Iceberg ecosystem, there's obviously -- there was a lot of change this quarter with the Tabular acquisition by someone. What do you see in terms of attracting talent to drive that roadmap forward here? Like how is your positioning in that Iceberg ecosystem evolving? Thank you.

On the margin side, the margins were slightly better than what we had forecast internally, but it doesn't change the guidance that we've given 75% for the year. Part of that is we're still waiting and some deployments for GPUs that around the world that we don't have yet that we're anticipating would have come in this quarter. That's really the -- on the margin side, the gross margin side.

But then on the Iceberg side, it's important to understand that the acquisition of Tabular, the company has no impact on the Iceberg project, which is an Apache open source project. This has contributors and program committee members from a number of a cloud companies, the hyperscalers, yes, but also other companies. And we also have members within Snowflake. So we very much intend for this to be an industry standard that we take a pretty significant role in shaping. And so from that perspective, we actually feel that the Tabular acquisition in many ways is a vindication of our strategy to bet on Iceberg because that was the format that was truly interoperable -- hopefully, this is the end of the Betamax wars and everybody with everybody centering around the one format that has broad support. And as I said, we will continue, to be a key player in this ecosystem to ensure that the format truly serves everybody and moves the industry forward.

Okay. Perfect. Thank you. Well done.

Our next question is from Mark Murphy with JPMorgan. Your line is now open.

Okay. Thank you very much. I'll add my congrats. Mike, you mentioned a couple of nine figure deals in the quarter. I'm curious if those are renewals with expansion or perhaps if they're related to anything else for instance? Iceberg tables unlocking, you know, new business where companies want to tap into some larger data sources that are in an open format or just whether there's anything else to call out on the nine figure deals. And I have a quick follow up.

Those were existing customers. You're never going to see. I never want to say never, but it's highly unlikely you're going to see a nine figure deal from a net new customer, but those are really as part of the renewal process and but expansion in both of those customers as they're looking to do more. I can't specifically say that Iceberg on either of those that I'm seeing, but they're -- they both plan on doing more with Snowflake.

Yes, I understand. Okay. And then as a follow up, as we start to think forward into the next fiscal year, I think we're trying to balance out the large slate of products that recently reached GA that you mentioned, Sridhar, and might start to contribute. Then on the other side of the ledger, you know, potential for any discrete headwinds from hardware and software, software improvements that you pass along. You know, storage, compression, the in the past, you had auto warehouse sizing. Any high level thoughts on Sridhar on how to pencil that out in terms of new products? You know, ramping and then on the other side, some of those improvements.

I'll start the answer and Christian should and Mike should chime in. First of all in all, we obviously can't say anything about next year. It is next year. But wherever possible there we have an indication about how new products are going to perform. We certainly told you about it. We have given guidance, for example, about Snowpark is going to do this year. And then similarly with the AI products, as I said, we are seeing broad adoption. And we expect that it will begin to contribute materially to revenue next year. With respect to performance optimizations, so I would say that's more of an ongoing thing. We have talked to you about the things that we have on top for this year. It is important to understand that these optimizations turn into massive cost savings for our customers and make the core product strong. And it's really important that our teams continue to do that, because that's what protects our overall base. Christian?

I would just emphasize that our leadership position on price performance continues to be a priority for all of us. And at this point, you've seen several years of us continue to innovate, but also deliver growth and additional usage from our customers.

Thank you.

We have a question from Kirk Materne with Evercore. Your line is now open.

Yes. Thanks. Just two really quick ones, Sridhar, actually a lot of your partners and customers at the Summit event that talked about the excitement around that's in these newer products like Cortex and Snowpark. But one of the refrain was we just needed better roadmaps to understand how to use them. And from an industry perspective, that came up a lot too. So kind of curious, where do you think you are in your product or your product so that they can understand the ROI that occurs?

Kirk, you're your line is breaking up and we can't really hear you.

Is that better now?

Getting a really bad echo. And it's hard to hear you.

I'll stop here.

Yes. Thanks.

We have a question from Kash Rangan with Goldman. Your line is now open.

Hi. Thank you, guys. So one for you Sridhar and one for you Mike, Sridhar when you look at the product portfolio, so clearly initiative is to get these services out in quick cadence. And as you pointed out, the net new Internet services, one of the conversations with customers like when they are discussing new service with you or does it take on where are we going to be a year out with the consumption profile of an average Snowflake customer and kind of the business, I think on average? How do you see that mix changing between the cores if you want just bluntly call it, warehousing, storage revenues versus unstructured data, literally uncluttered Cortex AI. And the other emerging buckets, how does that mix change for customers as they start to appreciate the net new products you have coming out? And one for you, Mike, you said the sales force compensation drove towards consumption is still in early days, but you also intimated that in fiscal 2026, we could start to see the fruits of all this. So help us to understand what you mean by that and what are the KPIs that you'd be internally monitoring to inform you in there for us that that tilt towards getting more consumption within your customers is actually working toward to your advantage? Thank you so much.

Thank you, Kash. On the product side, I would say that our customers go through a journey, typically starting with a desire to have a really good data platform that gives visibility. They end up adopting different architectures, but often the enterprises most like pristine cleaned data, the goal layer, so to say, is the one that's put into Snowflake and there are lots of customers that have standardized on Snowflake as that key data backplane. Next, usually, there is a leaning towards collaboration because all companies exist in the context of an ecosystem, they have partners, they have customers, and collaboration of various kinds, certainly starting with data sharing becomes the key next thing that they adopt broadly. And you've often heard Christian talk about things like stable edges, which is a metric that we track because it creates value and it obviously also creates a network effect. Our overall strategy at Snowflake is to make sure that all of the sort of data workloads that a company has is satisfied by Snowflake. And this is where things like data engineering, which you have played a pretty significant role in for a while, has been an investment for us. And this is where things like Iceberg become pretty key because all of a sudden, the universe of data that can be acted upon by Snowflake goes through a large expansion because -- precisely because not all data needs to be ingested into Snowflake before things happen. I would say AI is a little unusual in this front because of obviously the industry excitement around it. But we approach it very much from a viewpoint that I think this was part of the previous question of how do we go about creating utility to our customers. We just don't go in and say use AI. We talk about how it can be used to derive much better insights or unstructured information, for example, by using an LLM function for doing data transformation like sentiment detection. And easier access to data, whether it is with the chatbot or text docs or using something like analysts to give a business user access to structured data, those tend to be the follow-on applications and really we feel comfortable enough to be able to be investing in these in parallel and driving revenue growth. I would say it's too early to talk about x percent for this versus y percent for that. Our goal is to be relevant. And for me, relevance is lots and lots of our customers, thousands of customers using our products and driving meaningful 9, 10 digit revenue for us.

And on Kash on your question for me on the sales comp changes, what I was meaning by that is, we really bifurcated our sales force into the acquisition reps. And those customers that are landing today, we really have them focused on trying to land the right type of customer that can grow. We think they will have a meaningful impact on revenue next year with those new ones. And then also the new muscle that we've been building in the sales organization where the reps are just being paid on consumption is what is driving them and it's really the growth within customers' consumption. It's a new muscle for them to learn how to go and find and help forecast new workloads coming online and that muscle they're developing we see is going to have a really positive impact in 2026 for us.

Excellent. Thank you so much.

We have a question from Karl Keirstead with UBS. Your line is now open.

Okay, great. Hey, Mike, I'd love to just ask you about usage trends as you closed the July quarter and what assumptions you're embedding in the second-half product revenue guide? I think we've -- everybody on the line is has had fairly ample evidence from Microsoft all the way down that it's a tough IT spending environment. So, I'm just curious, as you said, the 3Q and 4Q product revs guide, what you would call out as the key variables or maybe changes in macro-related assumptions that you embedded in that guide? Thank you.

Yes, I'll just say, orders showed from a bookings standpoint that it's a normal environment and we are very pleased with the deals we closed in the quarter. I don't see it any worse. It's not euphoric or anything, but it's very stable customer buying pattern we're seeing. And in terms of consumption trends, obviously, we just guided our revenue right now for the quarter. We've beaten, we've raised the full year as well too, and that's seeing the consumption trends up through this week. So, we're pleased with that right now with what we're seeing.

Okay. Mike. And then just maybe a follow-up. I know that you've embedded in your guidance the assumption of some degree of runoff of the storage revenues that you just repeated earlier still represent 11% of revs. Is the expected pace of that storage runoff in this new guidance tracking similar to what you embedded three months ago or is it a little bit lighter or a little bit faster?

Yes. Well, what I would say, as I said earlier, we really haven't seen any storage leave Snowflake yet, but that was always forecast to happen in the second half of this year. And we are expecting that is -- some of that is going to happen and that is factored into our guidance.

Okay. Thank you, Mike.

You're welcome.

Our next question is from Brent Thill with Jefferies. Your line is now open.

Thanks. Sridhar, can you give us an update on the adoption of Cortex and how you're seeing that trend? And for Mike, just on RPO, it was good to see really good sequential growth and the acceleration of RPO, but the gap between revenue and RPO continues to be one of the highest we've seen. Is there anything that's going on that we should consider there? Is this just similar consistent patterns you've seen in the past?

I'll actually answer that first just while it's on my mind, as we said before, we have customers that sign long-term contracts. If they have consumed everything under their contract, they have the ability to buy monthly. We have two of our top-10 customers right now that can continue to buy through the end of the year and we're seeing that. So they're in their top-10 customers and think top-10 customers are roughly in the $50 million to $40 million range. Those aren't reflected in current RPO very much because they're just buying as they go.

On the quick question here on the adoption of Cortex, I believe we have a number of product capabilities under that umbrella. Cortex LLM, which references the different language models that are available. That one, the adoption is quite strong, lots of the use cases that alluded to on tech summarization, tech sentiment analysis, that's very strong, but also we introduced at Summit both Cortex Analyst and Cortex Search as a way to enable the users across all types of organizations to be able to chat and interrogate their data whether it's structured or unstructured. Those two are important for you, but the adoption at this stage is quite strong. We have quite a bit of demand for -- going to general availability and maybe the last one that I would call out is Snowflake Copilot we see a lot of usage on customers getting assistance on how to write better SQL queries, which also drives consumption back into Snowflake. So all up, a strong product suite and interest and adoption across all of them.

Thanks.

We have a question from Brad Zelnick with Deutsche Bank. Your line is now open.

Great. Thanks so much for taking the question. It's really a bigger-picture question for Sridhar or maybe Christian, if the world increasingly interacts with data through generative AI and LLMs, how does the role of the data warehouse as we've known it for decades evolve from here? And why is Snowflake well positioned to help enterprises bridge these worlds, leveraging mountains of enterprise data to build newer generation AI applications? Thanks.

I mean, first of all, Snowflake is often at the center of all of the interesting data in companies. Absolutely. We started out as a warehouse. But increasingly, we are the data backplane that provides a single unified view. Obviously, this comes from production systems, but also increasingly from multiple connectors we have to other systems, whether they are Salesforce or Segment or SAP or any of the other applications that you use. And what we are doing with additional capabilities like machine learning, like AI is to be able to act on that data to then drive operational systems. Disney, for example, uses Snowflake to do a lot of in-park optimization, and we see a lot of people that bring supply-chain data into Snowflake and then optimize within Snowflake. And obviously, there are also partners like Blue Yonder, which is a supply-chain company that platform on Snowflake and then provide additional capability to their customers so they can combine the supply-chain data with other data that there is. Our bet is really that AI and machine learning are going to go where the data is, data is going to have strong gravity and this is the reason why we are seeing such broad adoption. And by providing easy-to-use products with Cortex AI, for example, any analyst that knows SQL now is able to use language models. You don't need to go buy new systems, set up new things, and things like data transformation. Text transformations become as simple as writing like a two-line SQL. It is really this ease-of-use that makes Snowflake such an amazing platform to be able to do all these value-add applications in addition to the core analytics applications and the data sharing that people have done on Snowflake.

Thank you very much for that, Sridhar. Maybe just a quick one for you, Mike. Am I too optimistic to think that NRR can stabilize here, especially in light of all the new product that you guys are bringing online? Thank you.

No, As we've said before, I'm not really going to guide NRR. And I've said, it will over time converge with our revenue. And clearly, we'd like to see NRR stabilize at this level, but we'll see at the end of the quarter. I'm not disappointed with our NRR given our revenue growth, I'm happy with it.

Makes sense. Thanks so much, guys.

Our next question is from Mike Cikos with Needham & Company. Your line is now open.

Great. Thanks for taking the questions here, guys. The first one for Sridhar, and I just wanted to come back to Keith's question at the top of the Q&A. I know we're saying that there wasn't an impact on consumption in the quarter related to the muddled headlines around security. But has Snowflake noticed any change when thinking about the timeline for sales cycles or the demand gen as it relates to some of those headlines that were out there? And then the follow-up for Mike would be, when we look at the reiterated margin guidance that we have today, is there anything to think about or was there any delayed spend that's now expected in the back half of the year just given Snowflake's outperformance on a year-to-date basis through the first half?

On the first one about the model of security headlines, obviously, we make sure that we bring up the topic when we -- whether we talk to an existing customer or to a new customer, we point to the security capabilities that we've had honestly for close to a decade and try to ensure that all existing customers, for example, follow best practices. They're simple, like MS, like multifactor authentication, like network policies when it comes to security. And our sales team as well as I, whenever we have conversations with potential new customers bring this up, front, and center. And part of the reason why we are slightly muted about this is like these are our customers that got -- the people that got breached, these are our customers and we want to work closely with them to make sure that they get out of the difficult situation that they are in and both existing customers and new customers appreciate that spirit of partnership in helping them get through difficult situations. We had them talk to our CFOs, we had them talk to our security field CTOs, advise them on best practices. So like roughly, I would say that there's not really been any noticeable effect or delay in things like our ability to sign up customers or get -- I mean, sign up new customers or get existing customers to deploy new projects. We just need to be more proactive about having the security conversation and we absolutely do that.

In your question, Mike, on whether there's any delayed spend or not having an impact. There's no delayed spend, but I will tell you is we are looking at accelerating and it doesn't change our guidance. Our guidance is the same what it was. We are looking at potentially accelerating some of our hiring in the second half of the year, in particular for the sales force in some of the areas where we want and that's factored into our full-year guidance, which is guidance we gave last quarter.

That's great. Thank you. Appreciate the additional color.

Our next question is from Joel Fishbein with Truist. Your line is now open.

Thanks and congrats on the good execution. Sridhar, you spoke about this hospitality customer using Iceberg. Can you give us some more color about that use case and maybe ROI? And maybe if that may use case may be used as an example to attract other customers. Thanks.

Can you repeat the first sentence in your question, what customer?

You talked about a hospitality customer using Iceberg for a specific use case. I was just hoping to get a little bit more detail and color around that. And then maybe if that can be used as an example to attract other customers.

Yes. I think the general pattern that we saw there and we see another use case from other customers is the desire to adopt an architecture that is based on open cloud formats. And usually the adoption starts with a small use case, here's some incremental spend validate interoperability between engines and then you can go and deploy a larger scale. And I think that's the pattern that we see.

Okay. Thank you.

We have a question from Patrick Colville with Scotiabank. Your line is now open.

All right. Thank you so much for taking my question. So, one is for Sridhar, I mean, I want to ask about kind of classic data analytics and warehousing. Could you mind just commenting on the pace of migrations, new analytical workloads coming online, query pricing, competition in analytics now versus six months ago? Is there anything to call out there?

Yes. In the core analytics space, we are the best in the world that there is, especially when people consider migrating from complex on-prem systems. We have a professional services team that is exceptionally skilled at it and a very large ecosystem of partners that have been battle-tested with massive migrations. And we have done migrations from on-prem workloads that end up saving something like 60% of the cost the customer has to bear and their Snowflake implementations end up being very, very efficient and low maintenance. This is an area that continues to be important for us and we see migrations from a wide variety of legacy systems and there is also increasing interest in having AI-aided tools. We have a tool called SnowConvert that is used both by our professional services teams, but sometimes also by our customers. And especially in the world of AI, we are investing more into tools like these so that migrations can be faster. I would say these kinds of data migrations from legacy systems remain an important part of both new customer acquisition, but also driving substantial consumption increases in existing customers. Christian, you have a lot of background in this -

Maybe the additional color is that some of the systems run the most critical processes in organizations. Typically, they're closing the books on this. So, oftentimes a big portion of the migration cycle is validation and ensuring that the results are correct. And as Sridhar said, we are constantly looking at how to create technology and ways to accelerate the process, but there are some parts of that where the validation of testing is very important to customers.

And your question, Patrick, you specifically were talking about the cost of queries and how that has gone. We don't price on a per query because every customer's queries are different. We do a per credit pricing. You get so much compute. That has remained stable quarter-over-quarter. Sequentially, it actually grew year-over-year at 1.3%. But what I can say is the price performance has gotten better quarter-over-quarter for customers and that continues to be.

Yes. And then we have been public about a Snowflake performance index where we take not synthetic benchmarks, but actual customer workloads and we measure the performance improvement over time and that's how we give assurance to our customers that the economics are getting better on a regular basis?.

Super helpful. Thank you for answering that question. Can I guess I want to ask a quick follow-up. I'll leave it open to whoever wants to answer. It's about the prepared remarks saying you're not factoring in benefit from new products and then only minimal benefit in fiscal '25 from Cortex and Snowpark. I think when could those hockey-stick and more materially drive product revenue because the previous answer was that the call is rock-solid and very healthy. So what about the new stuff and when that's going to really inflect? Thank you.

The only newer product that is in the '25 guide for the full year is Snowpark as we said at the beginning of the year. And we said that's going to be about 3% of our revenue or $100 million and it's tracking to that nicely. And the newer products are not factored into our guidance until we see more history, I do expect it will have an impact this year. I don't know yet, but 2026 will have an impact.

That's clear. Thank you so much.

Our next question is from Michael Turrin with Wells Fargo. Your line is now open.

Hey, great. Thanks. I appreciate you taking the question. Sridhar, I want to go back to that last point and just ask your perspective, you mentioned innovation is a key focus area. So, as we're thinking about those newer product efforts and you're having those initial customer conversations, particularly around Summit, what are the couple of product areas beyond Snowpark that you see the most early demand signals or just customer conversation around?

So not only have we had customer conversations, but we've also developed specific cadences that teams are implementing in taking new things to market. So we place a special emphasis on what we call data engineering. Iceberg is definitely a part of it where Snowflake can be used for different kinds of computation that is distinct from the analytic workloads that we've been running. Christian talked about how a hospitality customer is running Snowflake on Iceberg tables that were not created by us. We very much focus on that. We focus on other things like streaming ingestions. So data engineering is one workstream, definitely a big focus on AI. We have basically a war room, a top-to-bottom team from engineering to sales that is focused on how we take AI products to market. And we expect our notebooks, for example, to hit GA in a few weeks and we will be making an effort around making sure that those get into the hands of data scientists so that they can run the most complex machine-learning algorithms that they want to run on-top of Snowflake. And so we have a pretty methodical approach to how we are taking new products to market. But across the gamut, whether it is AI or machine learning, or more sophisticated data engineering operations, including with unstructured data or things like notebooks appealing to a very different persona, that is broad interest. It's a matter of organizing ourselves to put the right product offering in front of the right customer at the right time.

That's helpful. I also just in summarizing a lot of the little comments on rest of your guidance, just go back to that point. We initially started the year with, I think it was around 625 basis points of potential impact contemplated from the mix of things you've talked about throughout the calls. Is that still the right zip code for us to think about? And is it fair to assume that you're leaving room for storage to come down from that 11% level in the second half, but you're not seeing that at this point. Is that the right takeaway for us?

Well, the storage, we always plan with Iceberg, though that was going to be the second half of the year, and we still think there will be an impact to that and that's factored in. But our -- every year we look at potential headwinds associated with performance improvements and other things. We don't update that on a quarterly basis. And we're not going to try to reconcile back to that. We never have and never will.

That's fair. Thank you.

Our next question is from Tyler Radke with Citi. Your line is now open.

Yes, thanks for taking the question. It's encouraging to hear that you didn't see any negative impacts from the cyber headlines that hit intra-quarter. I was wondering if you could just sort of talk about some of the offsets though to the strong consumption in tech and financial services customers that you saw. The magnitude of the beat was smaller than you saw in Q1. And I know there were some maybe positive trends in Q1 that didn't continue into Q2. But if you could just help us reconcile the magnitude of the smaller beat from Q1 to Q2, that would be great. Thank you.

Tyler, we've always said that we try to manage this business that's at a 3% to 5% beat is a big -- a good beat and we were 2.4% just as I don't get excited if it comes in at 5%, I'm not excited if it's at 2%. I just really -- we're trying to manage this business for the long term. I think what's more important is the fact that the beat there, the raise that we're putting through the full year, I think speaks more to what we see in the second half of the year happening now.

Yes. No, helpful commentary. On cash flow, it -- I think collections were a bit lighter. I think that could be a function of the go-to-market changes you talked about. The full-year cash-flow outlook was maintained. Can you just talk about the moving pieces in cash flow? Is there any impact from lower CapEx just given some of the GPU availability you talked and are you seeing kind of the billings terms compress more than expected?

No, the billings terms remain consistent, as I said, 80% of our customers are annually in advance and actually Q2 collections were actually pretty much as we planned. Payables were a little bit higher and the timing of payables. But our Q4 -- remember, we have a lot of seasonality in cash flow. Q1 and Q4 are going to always going to be our strongest free cash flow quarters. And so I'm comfortable keeping the 26% for the year right now.

Great. Thank you.

We have a question from Matt Hedberg with RBC. Your line is now open.

Great. Thanks, guys. Mike, for you, there was an early question on gross margins. It sounds like you still have some GPUs to procure in the second half. But I'm curious, now that you're kind of thinking through the impact of AI on Snowflake, confidence in sort of the medium term that gross margins have in fact bottomed and I think we could see -- start to see an upward trajectory here at some point.

So, first of all, we talked about procuring GPUs, we don't actually buy GPUs, we rent them. So, I just want to clarify that. And we have a lot of demand from customers outside of our major regions in Asia and even certain areas in Europe that want us to have GPUs so they can be using some of our newer products. And unfortunately, some of the cloud vendors are just not available yet. And when I say they want the H100s, the bigger ones, you can get the smaller GPUs. And is this the bottom of margins? I'm never going to say it's the bottom because I don't know what's in the future with new products that could have an impact on the margin. I'm not talking next year, but I feel good about the 75% for the year as we've guided.

Got it. Thanks, Mike. And then with U.S. Fed being a big quarter for you guys next quarter and sort of, I guess, some of the uncertainty on the U.S. election, have you discounted some of -- some U.S. Fed expectations for 3Q? Just kind of curious on, sort of, how we should think about that vertical going into next quarter?

Well, I will say our federal business is our smallest vertical. And as I said before, it's only upside for us. So there's not a lot of expectations in our numbers at all for federal. And I do expect we'll close some deals this quarter. And by the way, public sector is pretty good worldwide in what we do, but your question was on U.S. Federal itself. I do expect some deals, but in terms of impact on revenue, it takes time. That won't be until the future once we close those deals. We do have FedRAMP high now. We're working on some other things as well too. And so I do expect to close some deals this quarter in the U.S. Federal space.

Got it. Great color. Thanks, Mike.

Our next question is from Alex Zukin with Wolfe Research. Your line is now open,

Yes. Hi guys, thanks for taking the question. Mike, maybe to the -- and Sridhar, maybe to the question you got a little bit earlier, obviously, really great commentary on bookings growth, RPO strength, the raise being more than double the beat, which is I think the strongest raise in a while for product revenue. But to the point like what -- it's a little bit surprising to hear kind of technology being a strong vertical for you guys. Can you maybe talk about what you're seeing from the financial services and technology verticals that is maybe different? Is it just the macro in those areas? Is it use cases? And maybe where it was a little bit on the other side, a little bit weaker and kind of what you expect there?

Yes. I would say on the financial services, we just have some banks that are in the midst -- early innings of migrations like one of them has grown, I think, 400% year-over-year and they are going to continue to grow and that really is driving a lot of that. And on the technology side, we just have some -- a number of other companies that are growing nicely in the tech space for us as well.

And a different consideration, which we actively work on is the -- like the size and quality of the teams that we put to play in the different areas. We have among the best teams in financial services. That's where we were strong to begin with. And we are early pioneers of things like data sharing where that -- the positive momentum fed on itself. There are other areas like the federal business, for example, where we are still building out a crack team and that's driving growth in as much as one likes to think that it's the macroeconomic differences between sectors, what we can do often ends up influencing how robust the growth is.

Got it. And then maybe just one more. If you think about the competitive landscape that's kind of been a little bit of a charged topic this year, particularly for investors. What are maybe some of the major misconceptions or misunderstandings that you would say are out there and how or any changes you observed in the competitive landscape over the course of the quarter or even the first half of the year or from a pipeline perspective?

I'll start. Christian and Mike should add on. I think I've spent a lot of time on the road. Just this quarter, I've met with over 100 of our biggest customers one-on-one. And the thing that clearly stands out from a like an analytic capability, core data capability, data sharing is that we are bar none. We are the best platform that there is and our customers absolutely recognize that. And we see more-and-more of their data workloads move over to Snowflake. That doesn't make the process easy, migration of a large complex system that's managing the books of a bank so that they can close it every month there's nothing to sneeze at. It's a multi-quarter project that has to be done exceptionally carefully. So one thing that I wouldn't call it a misconception, but the one thing that I would reinforce is the strength that we have in our core capabilities. When it comes to newer things like AI, we've been open about the fact that we were a little behind early last year in terms of how much we were invested in it and the kind of products that we could rapidly moving field and the kind of products that we could deploy. But even before my coming, the management team here recognized the opportunity, invested heavily in it and things like the Neeva acquisition through which I came were an accelerant to things that were already in place at Snowflake. And I would say the change over the previous quarter is that we can tell our customers, we can tell you with confidence that our AI products are world-class And honestly, that we are much more reliable than building products off of APIs that you can get elsewhere because we pay a lot of meticulous attention to how we craft products. And it's that combination of reliability and ease-of-use on that we are turning into a major strength for us in AI as well.

Yes. One quick last comment on-top of what Sridhar said, our product philosophy on ease-of-use, smart defaults, how do you make things work easily out of the box, we hear competitively, it's a strong advantage for us across areas for analytics, for data engineering, and for AI, And we'll continue to invest with that philosophy in mind.

I promise, last comment on this to give -- to talk about something very concrete that comes with the Snowflake platform, especially after the incidents of the summer like CrowdStrike, one of the hot topics has been how we set up replications at Snowflake. This is now a battle-tested operation and most customers are shocked to find out that you can run a full replica of an important deployment at something like 15% the cost of the original deployment because the replica is just -- is basically keeping up, it's not running any of the workloads that are going on in the main one. It's feature is like that, that continued to drive our strength in the core.

Super helpful. Thank you guys.

We are out of time for questions. So thank you all for your participation. You may now disconnect your line.