You know, Sarah, who was here, Sarah Guo, my friend who interviewed Sam and I. Her feedback to me last night was, "Thank you for inviting us to your rock concert." I was like, I thought I was running a data company. It speaks to the interest and excitement that data and AI have in the enterprise world. I mean, all of you know this, for much of humanity's history, data has been a little bit of an afterthought. It is something that you used to report the news. You know, that is why early writing was invented, to keep track of taxes or land ownership. I was very fortunate to be part of a company, Google, which thought about data fundamentally differently. There are other examples like that. In the world of search ads, knowing how you did was essential to improving what you did.
I mean, think about it, the company itself, Google was founded on the premise that the best answer, the best web page for a given query was one that the rest of the world said was the best answer for a given query, and how you break that feedback loop. In search ads, we had a similar mentality. It's really the combination of the advertiser intent combined with what did users find to be the best answer that really drove how the team thought about what the product should do. In many ways, my data team there was just as large as the product team. Accidental, but it influenced everything that we did. We built machine learning systems at planet scale 20 years, like in 2005, before like from now, a decade before they became wildly popular elsewhere.
I think we are in a moment where the rest of the world, thanks to AI, is waking up to the fact that data can transform how businesses can and should operate. It is that movement from data as an afterthought to data is really important for you to get ongoing visibility and start predicting what you should be doing in the future in your business, as many of our customers like Disney have done to do things like predict next best action, to now thinking about, wait, AI can take this one step further. It can drive business transformation in a fairly fundamental way is what drives momentum for Snowflake. I know I speak for many of us at Snowflake that we feel very lucky to have this opportunity.
As you saw today, we are doing everything that we can to seize this opportunity with all our hands. This is a little bit of a schematic. It's a caricature of how we think about what we call the end-to-end data lifecycle. When is data conceived? Typically in things like transactional systems. When you open your app, that app sends a little message that says, "Aha, you know, Sridhar opened this app." It gets recorded somewhere. We see our technical mission as being there at all key aspects of this end-to-end data lifecycle. When data is born, when it is ingested, transformed, and cleaned, and then when analytics is done on it, and then the predictive analytics where the history is used to predict the future then gets done on it.
What is also unique about this moment is this is the big unlock from unstructured data. You saw a ton of announcements today, whether it is Openflow or Cortex Search. That is because unstructured data has always been a little bit of an afterthought in how people thought about data. It was just too unwieldy. You could not do that much with it. You know, I joke to people for much of my life, Command F is my favorite PDF search engine. All of you write very nice reports, but God help me, until recently, if I had to figure out what did you actually say about Snowflake or a competitor. It is that unlocking where everything that we talk about, whether it is data ingestion or cleaning or processing or analytics, can now act on structured and unstructured data.
In a world where metadata and governance is going to be even more important, you have heard the phrase AI-ready data, both in my keynote as well as Christian's keynote. It's not enough to have data. You need to have additional information about what does it mean. Any of you that has done, you know, written SQL queries for a living or tried to analyze data knows that is super messy. I've spent thousands of hours analyzing data, and we always give like terrible names to all the columns. It's really hard to figure out what the metrics actually are. It's just like, you know, that's sort of how it works out. You have a column called revenue. You do not quite like it.
You'll come up with another column called rev2, and the poor person that shows up two years later is like, who are these people and what did they do? Having that layer of intelligence about that data becomes very important. Of course, finally, the world of consumption. This is roughly how we think about it. When we talk about wanting to play a critical role for our customers into this end-to-end data lifecycle, we talk about playing meaningfully in these different layers. The product strategy very much over the past year has been, how do we release the key components that will let us play this meaningful role? You know, look, we understand that it is easy to say things like all users, all data. That's an aspiration. That's not a plan.
Part of what Christian and Benoit and I and the rest of the team have been doing is be very methodical and deliberate about how do we attack this big opportunity? What are the critical first steps that we should take? How do we always lead with strength? It's easy to aspire to be something else. It's very hard to get there if, you know, a set of companies have gone at some area for 10-plus years. It is not all that easy for any other company to catch up. I think it is really important that we have the big vision, but that we are very deliberate in how we go about seizing that vision. We'll get into some detail later.
For example, in the space of databases and transactional stores, this is everything from recording a session that you're having with a chatbot to how do you have a user store that remembers preferences that a user has. We have taken steps. We've been working on something called Unistore for five plus years now that's meant to combine the best of a transactional store with analytics. Clearly, it has turned out to be a harder project than we thought, but it's out. It's doing well. The acquisition of the finally named Crunchy Data. It's an amazing company, by the way. When I first heard them, I was like, really? You're called Crunchy Data? It's a really good company full of world-class Postgres developers. It gives us an important component in how we go after this transactional market.
Another component is how do we make it much easier for people to ingest data into Snowflake, into cloud storage? That is the Datavolo acquisition that happened six, seven months ago. We have brought it at record speed to public preview. That is another thing that I want to stress, which is we are leaning much more into being effective with acquisitions, getting them out to market, getting that feedback, getting that happy loop of delivering customer value and iterating from it. We think Openflow is going to make a meaningful dent in how people look at us when it comes to getting more data in from existing systems, sometimes legacy systems. Openflow is also much more even between unstructured and structured data. As I said, unstructured data is, I think, a massive unlock in many different areas.
We continue to invest very heavily in the bronze and the silver layers. These are the earlier stages of data processing. Previously, in many of my conversations, Christian and I have talked to you about how paradoxically Iceberg was the aha moment for us a couple of years ago about what Snowflake could do in earlier stages of the data pipeline. We continue to invest very heavily in things like Snowpark. How do we make sure that we are seen as best in class when it comes to early stage computation? The partnership with dbt, bringing dbt right into Snowflake so it feels very much like a first-party product. The ongoing collaboration with them, I think, positions us really well in the earlier stages of the data lifecycle. We have always been best in class in analytics. We intend to keep that lead.
Again, finally, I've worked on databases for a very long time, and I've started working more with the team. It's been fun both to see what they have done and also to lean into what we can do in the future. As I said, getting data AI ready is going to be a key differentiator. We will have partnerships, but we also think that it is an area where we can drive the ecosystem in terms of having customers ask for their data to be data ready. Right now, we live in a world where, yes, you can buy a catalog on the site, but that information is largely isolated from what BI tools do.
If you have some information stored in a particular report, I tell people that every meaningful report is actually an encapsulation of meaning because somebody has taken the trouble to go and say, hey, these are the different pieces of data that need to sit in this report, and it's meant for a particular persona. That is the semantic context of the underlying data. A lot of our work around semantic views and making that commonplace is really based on this belief that data and metadata need to be close to each other. The closer they are to each other, the more value our customers are going to get with it. We also think that this turns into things like data just being ready for agents, data both unstructured and structured data being much more usable and composable.
We lead in governance, and you will continue to see us invest and keep that lead. Finally, we see the world of consumption changing. We are investing into it with the things that make sense. BI as a category, in our humble opinion, it's an established category, and we have great partners that we work with. Us trying to do another BI tool doesn't make a lot of sense. On the other hand, you saw what Snowflake Intelligence could do today. Again, it's not that far to then be able to say, I want to remember all of the questions that I asked today in my conversation and have it turn into a report with these configurable parameters. We see these as democratizing data access and redefining BI as a category. Remember, BI today works in a pretty straightjacketed way.
You have underlying data. You have some genius that knows about the semantics of that data. They have to create the views, and then they have to create the dashboard. If you, as a business user, want something different, sometimes a BI tool will let you do it. Most of the time, you get to wait. To us, the flexibility of AI unlocks those things and lets us think about the category very differently. You're not the only ones to be doing this, but we feel like we can do this with strength. We continue to invest in the notebook experience, in things like machine learning, the predictive part of AI. That is a product where we clearly came from behind, but positioned as a natural add-on to what we do at Snowflake. It is gaining strength.
We also see a world in which things like native applications—native applications are a Snowflake specialty—it is a way to encapsulate procedural logic close to data. It was never meant to be a full-fledged app development platform, but what it does is it lets a partner like S&P Global take their data, attach logic to it, and then ship it to a customer so that they can use that logic. Now they are very excited about the fact that this can be an agentic component that can respond in natural language, and it lets them create value at a whole different level than what they have done before. Finally, when it comes to sharing, you saw Christian demonstrate today already. Things like sharing now joined with AI and semantic models makes a lot of interesting things possible that were not really possible before.
Cortex Knowledge Extensions is a super simple concept. It's a data set plus a search index plus a model on top. This is a product that literally did not take much time to build because all of the components are there. It gets turned into you have all of the data repository of a genetic available as one single component that you can embed into anything that you want to create, any agent that you want to create. We feel very good about how we are set up and how we are executing on it. I place a lot of emphasis with ourselves, with the product teams, on practicing what we preach about AI.
It has already gone through pretty major transformations where the initial versions of foundation models were very good if you could figure out how to stitch things together and send it a particular bit of context, and then they would give you a useful answer. It went from there to these models are now very good at figuring out which data set that they should be talking to. With things like Snowflake Intelligence, we are not that far away from they can begin to replace pieces of workflows that are really important functions, whether it is loan underwriting or like what our sales team does, for example. They should be able to automate pieces of it. We do a lot of practicing what we preach.
There's an internal tool called Raven, which is an agent that is meant to surface everything that matters to a salesperson in one interface. We spend a lot of time thinking about how do we make our engineers more productive, our folks in sales more productive. A lot of that kind of reflecting has led to one thing that I forgot to mention at the very beginning, which is we made a big push about roughly six, seven months ago to accelerate migrations, moving data over from legacy systems to Snowflake, again using AI. You saw Christian's announcement about SnowConvert. The idea is how can agentic loops make the very painful process of moving data from legacy systems over to Snowflake.
That's a little bit, as I said, using AI ourselves, using AI in the product is a little bit of a theme that we are constantly pushing ourselves. It's a world that's evolving rapidly. Cursor was barely a company a year ago. Not only is the UX team that developed all those cool UIs using Cursor as their primary IDE, they're being inspired by it and creating experiences like Cursor right within Snowflake. It's an incredibly rapid pace of evolution. One question that you folks will ask yourself, that we ask ourselves, is, gosh, this is a really broad surface. How do we make sure that we actually get things done, and how are we organizing ourselves? Christian alluded to it in some of our slides.
We have done a natural organization both in the product and engineering team, but also in Mike's go-to-market team, where we have pillars of teams organized around everything around analytics, including things like AI-driven migrations. We have another team that's focused entirely on data processing. Everything that you heard around Openflow, the Spark implementations, dbt, and so on is housed within that team. We have a team that's focused on AI products. How should we be thinking about the future of business intelligence or data science? The final pillar is what are the foundational things that we need for people to be writing applications, and how do we make sure that we turbocharge things like sharing and semantic context that can be attached to data? This lends itself to a very clean organization.
Part of the change that we have made over the past year is driving accountability. There are leaders in the product organization that are singularly focused on making these areas successful. They have partners in marketing, in Mike's sales team that know when certain products are specialized and bring along the requisite expertise. Mike will talk about it a little bit, but as you folks know, the specialist motion is one that should be used carefully. We understand the power of it, but we also think a lot about how do we distribute to the entirety of the sales team. Areas like analytics are very much that joint responsibility. We have put in place mechanisms by which the levels of our account executives, the sales engineers, that's the dominant workforce in the sales team, is getting better with respect to the newer areas.
With this larger lens, we feel that we are able to take on a much larger market. This is obviously a thesis. It is based on two things. One is the on-prem to cloud migration that all of you folks are very knowledgeable about. It is also our belief that a platform that is centered on data and AI is going to play a larger and larger role in the world of cloud computing as we know it because we feel that that data centricity is really important in helping customers realize fast value.
The number of customers that I have met even in this conference, or barely a day and a half, that have bet big on Snowflake and have had amazing results, sometimes in six months, sometimes in nine months, to show for it, is proof that we are well equipped to take on such a large TAM and realize a meaningful part of it. You folks ask a lot of questions in all of our earnings calls about what's the thesis for why we believe we are going to be successful. Our consistent feedback to you has been that we derive our strength as a company in continuing to be world-class at analytics. We'll get into a lot of detail about this, but we are, at the very least, many years ahead when it comes to the core technology. Others talk about serverless.
We've literally been doing serverless for a dozen years. The same goes for what we announced today, which is the adaptive warehouse. It represents a step change in how people should be thinking about analytic computation as a whole. We have a world-class team when it comes to how good we are with analytics. Lots of unsung stuff, including obscure but really hugely important things like pipelined execution for user-defined functions. Most people do not know and do not care what that is. If you are doing a Teradata migration, it turns out to be really, really critical. There are a dozen things like that that have taken multiple dozens, sometimes, of personeers to get into the core engine. We feel very confident about leading with strength in this area and things like being able to drive migrations a whole lot faster.
I've met customers who will roughly tell me, "I have 50 on-prem data warehouses sitting in this thing I really do not want to pay any money for. All you need to do is move that data safely over to Snowflake. I can't have you mess any of that up. That money is yours." So far, we've been hamstrung in our ability to scale that quickly because those projects could only go so fast because they were always human-scaled. Part of what excites us about AI is that ability to make that go a whole lot faster. By the way, that means we have to transform ourselves in terms of how do we think about running migrations. We have a professional services team. We are busy thinking about what is the future of that kind of systems integration. We'll talk about partners.
Those folks, the GSIs, are having the very same conversation. When it comes to innovation, I think last year was definitive proof for us that when we put our mind to it, we can really step up on the gas. We went from being, let's face it, a nobody when it comes to AI last year to every customer wanting to know how they can use Cortex AI to not only get things done with AI, but to be able to transform their business. We do it in a position of strength. I talked yesterday. Christian talked about it, where we prize qualities like simplicity and tightly integrated products. You do something with Snowflake Intelligence and create an agent. Everything that you put in with permissions, with data governance, works out of the box.
That is not something that people that essentially think of technology as just adding on services are going to be able to easily catch up. We have a long lead in many areas, but we also bring along a product philosophy that is going to strengthen our position in the market. Mike will talk about this, but we actually think that while we have a healthy partner ecosystem, we can be driving this so much more. I've personally been involved in many conversations with many GSIs and many partners. They can sense change in front of them. They can sense that system integration five years from now is going to be nothing like what it was five years ago. They are excited to be partnering with us through this change. I think it's a point of real leverage for us. We have an amazing team.
Mike Gannon, who you will hear from, comes with a wealth of experience in driving sales teams at scale that I think will stand us in good stead. Vivek became head of engineering some seven, eight months ago. The transformation that he's wrought on the team in terms of driving accountability bottom to top, in terms of how we think about talent, how we think about getting products done, the sense of urgency that he brings to the table. Like me, he's 24/7, all days of the year. I think that intensity shows through in what we do. I'm also very pleased that Mike Blendina joined us. Mike ran the payments group at JPMC. If I remember correctly, some 8,000, 9,000 people processing a mere $10 trillion of transactions every day. It's an amazing team.
Before you ask a question about it, Mike continues to be super active. He is actually not here physically because his daughter's graduating this week. We have many promising candidates for CFO and hope to be updating you folks relatively soon. Finally, you see our event marketing. I am blown away by the scale and just sheer quality of this event. What often astounds me even more is Denny's and her team do 20 more copies of this throughout the world. The energy in Sydney, where I've been at, or in Tokyo, is like the energy in this building. I got to tell you again, it's sort of super funny for the world of data to be this exciting. I'll just spend a couple of minutes on sort of values, which I think are important. We've always been customer-first. I heard this from Frank.
I think I told you folks this. I basically did a tour with Frank last year. Actually, it was the year before last, where we spent five solid days together. It was conversations during those that first brought up the idea that perhaps I could succeed him. Part of what he drilled into me during the time that we spent was how much people were betting when they were betting on Snowflake. He's like, "Sridhar, people get fired if their migration doesn't work. They're betting their careers on Snowflake." That is why you need to take each and every customer seriously. When something does not go right, we need to swarm and make things right. You can sense that again in the conversations that we have with folks. We, even more than before, prize accountability. It has to come with the right structure.
We have key leaders in the right positions that know what success means for them and how that contributes to the team's overall success. We have mechanisms. We use OKRs, at least at the top level, for driving down accountability and alignment. These two go hand in hand. We structure ourselves so that there is very clear alignment of how the parts add up to something larger than the sum. We hold ourselves accountable. The quality that I've stressed with the team is ongoing excellence. My take is that excellence is a way of living. It's not a milestone that you get to. We are exceptionally lucky to be in the position that we are in. It's really going to take us constantly pushing ourselves to be better to achieve the kind of company that we want Snowflake to be.
In terms of what this translates into, things that folks in this room care about, we're super disciplined. Our big areas of investment are in R&D and sales. You will have a question about how many folks we added to the sales and marketing teams. We hold ourselves accountable in every function. In R&D, for example, we realized at some point that we were too top-heavy. We are very intentionally focusing on people that are earlier in their career because in a paradoxical way, they come less encumbered by all the assembly language stuff that's sitting in my head, for example. I did it 25 years ago. It's just not that relevant anymore. We are very thoughtful in how we invest in our R&D teams and in our sales teams.
We hold ourselves accountable, not just with folks like the account execs and the solution engineers for whom accountability is very direct and in your face, but in every function. We spend a lot of time thinking about how do we make sure that these prized teams that directly influence the future of Snowflake as a company are as productive as they can be. I've talked previously about things like stock-based compensation. Some of it was driven by previous policies of not being as frugal as we should be. Both in how we approach R&D and sales, where the majority of our SBC goes, we have a very solid plan to have that be much more manageable. The final point that I will make is that we continue to be very active with M&A, but on our terms.
I don't want to pay a billion dollars or $2 billion for companies making single-digit millions. It's just like the math just doesn't work. We don't live in that world. Thanks to acquisitions like Datavolo, or even Neeva, which is a very modest acquisition, but laid the foundation for what we did in AI and search, we feel more and more confident about our ability to acquire companies, to turn them into meaningful businesses right from the get-go. We are super active in this space. We are open to all kinds of acquisitions, but they need to make sense, and they need to create accretive value.
All of this, the product opportunity, the market opportunity, our shown ability to execute over the past year and more, and the excellence of the team that we are building up to both at a leadership level, but also in the rank and file at every level of Snowflake, combined with what we think are eternal values, cultural values in how we should operate, with an eye towards being sensible in business, make us feel really, really good about where we are as a company. With that, I am going to hand over to Christian to talk to you a little bit more about our product roadmap, what we have done, and what we are aspiring to do. Thank you.
Thank you, Sridhar. Hello, everyone. Awesome to see many familiar faces, many instances. We're joking with someone that many of our interactions are on Zoom. It's awesome to be able to see you face to face. Let's get started. Maybe the most important takeaway from what I want to share today is aligned with what Sridhar shared, which is we feel really, really good about where we stand, not only in the industry and relative to competitors, but most importantly, in the relationships and the impact that we're having with organizations across industries of all sorts of sizes. The slide that I'm starting with, this is a third-party study. It's a vendor. He was grilling us, and you know who the Apache Spark vendor is. He was holding us accountable hard, like he was saying, "Okay, how do you do this?" And we'll give an answer.
He would say, "This part of it is true, but it doesn't work in this scenario." It is a very technical vendor that went deep and published. There is a blog post on this. We can get you the source. This is the simplified version of that. It highlighted something that we have been very proud of since the early days of Snowflake. It speaks to what Sridhar said yesterday, just retraited here, what Benoit said, which is we believe that we will win in the long run relative to the alternatives by leaning in on simplicity and the ease of use for our customers. Why is that such a compelling value prop? Complexity conspires against what customers are trying to accomplish. We will talk about compute.
I know there are some questions from you, but it is truly dizzying to try to figure out what is the compute instance I need to use to achieve any one task. That is in one cloud, not do it times three. We focus on that abstraction of complexity and simplifying things. That is how we believe is the long-term differentiator and advantage for us. Sridhar mentioned something that is super dear to all of us in the product and engineering organization, which is you do not build an analytic system, a data processing engine to the quality of what we have overnight. Yes, you can get the synthetic benchmark to go and show results in some short amount of time, but there is nothing like the real world.
You say, "Okay, the simple test is going." Once you try to do a migration from an actual system, then you realize that it is hard. Sometimes even math does not add up. I have shared with some of you an instance that happened maybe a year or so ago. We were migrating from Teradata. Results from the two systems were different. Three months into big investigation, it turned out that it was a bug in Teradata. This was core financials of a company. It took a while to go and validate those sorts of things. This is the type of advantage that Snowflake has. It has been battle-tested, used by many of your organizations and thousands of customers around the world. We feel quite good about what our position currently is.
Of course, you saw this morning, and we'll talk more about it, we're not standing still in any way. What the study here ends up concluding is from a pure total cost of ownership, we come out ahead. I think that where you start informs your choices. I have no problem talking to customers about Databricks' audience. They come from a tinker mentality, a very developer-oriented mentality. Removing knobs to make things simpler, it's hard. For us, it's easier to judiciously decide where we have more control because going the other direction is hard because there's backward compatibility, there's unhappy developers, et cetera. It's not just total cost of ownership. I think pure performance also matters. I said on the main stage this morning, I'll reiterate it. All benchmarks need to be framed with the appropriate context.
I think it will always be possible to find the one use case where pick whatever system shows up better than another system. That I granted upfront. What we're showing here, and we showed on the main stage, is we took a set of representative benchmarks. We think that they are correlated with what our customers do on the platform. What you see here are the results of the Gen2 Warehouse. I know some of you have questions on this. Gen2 is more updated, faster instances from the cloud providers, but also a very comprehensive suite of optimizations that lead to this type of results compared to both an Apache Spark implementation as well as a cloud-native database. Probably where the battleground will truly evolve is in the open data. For the last couple of years, there was the, "Ooh, Snowflake is closed data.
They're not truly committed to open file formats, open table formats. That's where you're going to get killed. These are real numbers, similar set of real customer workload-inspired numbers for benchmarks on Iceberg data. We are leading in many, many types of operations, many, many types of use cases. In our conversation with customers, I was meeting with a large bank last night. It's like, we want the level playing field. What open file formats, what open data formats, what Iceberg does is it levels the playing field and makes it easy for customers to switch back and forth between us and BigQuery and Microsoft and Databricks. Yeah, that's what it does. It truly removes the lock-in. Where is the value-based lock-in going to come from? Who's giving the best experience and who's giving the best price performance?
We're ready to go and take on that battle all day long. What I told these executives of the bank last night was we welcome that battle. We want to engage because we believe that we're going to get an unfair share of that type of workload. This is what you see here. I want to preempt some of the questions on compute. Super clear, we have always been, and we continue to be committed to be the leading platform in terms of price performance. There was a lot of questions on, do we announce both Adaptive and Gen2? And why are you announcing two things that are similar but different? How do we think about it? I'm going to preempt some of your potential questions.
In the same way that I said in the main stage, Sridhar just reiterated serverless, which is the notion where customers do not need to think about compute instances, was what we pioneered and what everyone else has followed up on or copied. We wanted to make sure that the thought leadership that we are applying into Adaptive is out there because it is not just an idea. We have customers that are running it. We are running on our internal instance of Snowflake. We call it Snowhouse. That is on Adaptive. The thing is phenomenal. At the same time, tuning a system that is adaptive in nature, that is learning from behaviors, is going to take some time. Before, I do not know, a year from now, we will have the conversation, but Adaptive is not GA yet.
I don't know exactly what's going to be the timeframe, but it's going to take some time. We will continue onboarding customers and delivering value for those customers. That is already, the train is in motion. What happened with Gen2 was, again, there were new instances which were materially more expensive than what we've done in the past. For as long as Snowflake has been in market, 10 years in June, very soon, 10 years in general availability of Snowflake, we've always transparently upgraded customers to newer instances. We all collectively went through the journey of migration from older instances, C2 to C3. Then we did the R migration. What we always did is prices are stable, system gets faster, and we have this conviction that if we improve performance and improve the economics for customers over time, that leads to longer-term commitment and additional workloads.
The philosophy is still there. What happened with Gen2 is the instances were so much more expensive that if we did what we've always done, then it would not work. You would all not be happy. We would not be happy. What we did is we had to introduce a new type of warehouse. It is priced at a higher level than what you see in the traditional Gen1 warehouses. It is so much faster that the price performance advantage for our customers is still there. That is why I have no problem in the main stage telling customers, "Go try it." I got a bunch of emails today already on the, "Oh, we're trying this thing. It looks super fast." Yes, nominally it is more expensive. If you're spending less time, customers come out ahead.
The notion is we had to do this just to let out a ton of innovation, leverage new hardware instances. I do not know if at some point there is a Gen3 or something in between. The future is adaptive, and adaptive is taking all of the benefits of the warehouses that we know to this date. You do not have to know if the size is small or large. Over time, it gets unified with Snowpark Container Services, which is sometimes we gloss over details. There are warehouses, Snowpark Warehouses, Snowpark Container Services. We have done each one of those steps for a reason. We want to go back to our strength, which is if you have too many options, you are making it too hard. Go and remove options. That is what adaptive does. Hopefully, that helps clarify how we think about it.
Maybe something important for all of you. I think in many conversations I have with some of you in the room, there's always been the question on, "Well, but why are you passing all the benefits of performance enhancements to customers?" Both Gen2 and adaptive have now given us the ability to decouple performance enhancements from revenue impact. I'll say we are, even if you go do the math and look at the benchmarks on Gen2, we are keeping some amount of the value. I want us to also collectively understand that if we become a company that is always keeping the dollars fixed and just, yeah, the performance is better, but the dollars are fixed, I can tell you that that's how we become the next Teradata. We are very clear and very determined that's not going to happen.
What this does enable us is to decide when and how much of those savings get passed to our customers, which I think is something that all of you have given us feedback over the years. It is live on Gen2 and continues to be true on adaptive. Sridhar talked about product velocity. There is no two ways to go around it. If it is not evident by not only Summit, and Summit, we usually line up a lot of things, but Summit right now is a milestone. When we do build in the fall, there is also tons of innovation. There have been so many other things that are happening in between our large events that we are still doing launches. Cortex agents, we did it like a month or two months ago. There are other capabilities that we are just launching on a regular basis. Where does this come from?
Sridhar already alluded to it, which is we've organized product and engineering teams by what we call product category, the five areas that you saw, and I'll recap in a short little bit. Each one of those has a clear swim lane and a clear mandate and go as fast as you can and go and win in this category. It is also true that there is AI-driven productivity. We had this fascinating story. Sridhar mentioned our PE team, our product experiences teams, and many others are using Cursor and other tools. There was this funny thing that Benoit himself, he's still coding. He is man plus machine these days. It's Benoit plus AI, Cursor. He's been using it so much that he got throttled last week. This is a true story. He was up in arms.
He's like, "How do we give them more dollars so that my productivity gets back to what it was?" It was a fire drill. We gave them more money. Benoit got the latest thing. He's back to being way more productive than he's ever been. That story is true. It's not just Benoit, but it's engineers along the way. The ability to just go faster is becoming a key asset of how we think about this. AI thought leadership. Sridhar also mentioned we've gone from zero to something in a very short amount of time. What we have done is the original team that we had assembled that really know modeling and AI inference and true AI research, they continue to build models. They were the ones that built Arctic and Arctic Embed. They continue to build models for certain small areas where needed.
Like Document AI now has a new Arctic Extract model. That is small, efficient, better than anything else out there. We can do that. We are also looking at where the state of AI can be improved. We have an AI research team blog. You can go and look at the details on all of these. They are producing world-class results and sharing with the community for us to go and be part of that AI leadership. The examples here are a little bit in the weeds, but how do we make inference faster, which is what many of the slow chatbots go through, or it is too expensive? We did some research on SQL generation. The moment that we published it, we were number one on the Spider SQL query benchmark, the standard benchmark everyone is playing.
I was asking Katherine, "Hey, can you confirm that we're still number one? Because the world of AI is moving so fast." Yeah, as of this morning, we're still not only number one, but we also are number two. We are far ahead on the spots of this. We will continue to invest in AI core technology shared with the world out there because we want to be part of steering where AI is going. The same is true of Iceberg. This one warms my heart in a very significant way because we are shaping where Iceberg is going. We have hired a number of PMC members, the program management committee folks, the ones that vote on the proposals. We co-chaired the Iceberg Summit that happened here in San Francisco a few months ago.
If anyone gets a recording of the presentation from Ryan Blue, who is the chairman of the Apache Iceberg committee, he was very upfront. These two proposals came from Snowflake. These two proposals got feedback from Snowflake. These are shaped by Snowflake. With you all as the audience, I just enjoyed to know and be able to say we are steering Iceberg and we did not pay the $2 billion that Databricks paid. It is true. That is the reality. We are steering it. We are part of it. We are partnering with those guys at Databricks. We are making good use of our shareholder money. The roundup, this is the similar slide to what I had in the main stage this morning. The roundup on types of data is very important. Analytica has been a story for a long time, of course, the beginning of Snowflake, hybrid, the HTAP, Unistore.
Yes, we underestimated it by a lot. Yes, I came and told you it's the Holy Grail of database. And it still haunts me that it is the Holy Grail of database. But it's working. It's going better than we thought since it hit GA. GA took a long time. Someone stopped me in the hallway an hour ago and said, "Can we please get Unistore on Azure? I have the use cases." I think we have a capability. Now we're rounding it up with the Snowflake Postgres that comes from the Crunchy acquisition. If I line up the different enhancements from this morning, this goes now framed into the five categories that we mentioned. Sridhar has already reinforced the importance of some of these enhancements. We think Openflow is a major capability for us just in being able to make more data available to Snowflake.
Also, it gets us more upstream into the data lifecycle. This is very important. If you think that data is being born in devices and IoT and OLTP databases and somehow they land and then they go get processed by Spark and Databricks and then show up in Snowflake for analytics, this helps us be way more upstream on that lifecycle. We're early. We help customers with data as it's born. We help customers with data when it's consumed. It might as well go do the middle thing with Snowflake. There's a lot of strategic value in how we do this. AI SQL, we spent time in the keynote this morning. There was a demo.
Hopefully, all of you got to appreciate the piece, which is we're bringing AI to our core audience, the SQL-savvy persona, the analyst that just now has this very strong toolset with multimodality. It's not just text, but it's images and audio. We'll do video, that's coming soon. Snowflake Intelligence, we talked about it. You saw it. I'm not going to spend time here. Sridhar also mentioned the AI-ready data that matters a lot to us. Last couple of slides. A year ago, I stood here in front of many of you. I gave you a taxonomy of the different efforts we had. I said, "We think that these are going to be material for FY2025." Chatting with Jimmy and Katherine, he's like, "I'm not going to stand in front of you and not hold myself accountable to what we said.
At least the three that we say would be material for FY2025, they've largely turned out to be the case. We have stopped looking at unstructured data as its own category because at the end of the day, unstructured data is generating value for us through one of the two, either Snowpark or Cortex AI, Cortex Search in particular. If you look at what we share with you in earnings and calls and callbacks, yeah, the first two, Snowpark and Cortex, are doing well. Also, Dynamic Tables played out quite strongly. Iceberg, which was a conversation that we spent maybe way too much time, has materialized the way we originally would have wanted it, which is the tailwinds and the opportunity of Iceberg far exceed and outpace the potential headwinds of Iceberg. Jimmy framed it at the beginning of this conversation.
Of all the product enhancements that we are working on and future innovations, we do want to think of them in these five product categories: data engineer, analytics, AI, applications, collaboration, and platform. You will see us talk in these terms and frame all of our efforts in this same taxonomy. At the end of the day, I will end in the same state where I ended it this morning, which is the AI Data Cloud. I want to invite Mike Gannon onto the stage. Thank you. Great to see you all.
Thanks, buddy. Good afternoon. Welcome. Hopefully, you are feeling the energy in the conference this week. I am curious because this is my first summit. Show of hands. First summit for everyone else in the room? Okay. About 30%.
By way of introduction, Mike Gannon, 80 days on the job as Chief Revenue Officer for Snowflake and feel absolutely thrilled to be here. The energy and excitement that I'm getting from this conference is revitalizing why this decision, and this is only the third career decision I've made in my life, is probably one of the most exciting chapters of my career. I want to share with you a little bit of background. First chapter of my career started in 1997. I drove from SUNY Oswego, New York, to Hopkinton, Massachusetts, for an interview with this company called EMC Co. It was probably the more unpopular route to take. Most of my friends had gone into finance and insurance. I went a different route and took a run at a technology company and had the good fortune of landing a job there as an inside sales associate.
Number one job was setting up 15 qualified sales calls a week. In 1997, EMC was not quite a household name. We were trying to get customers to decouple a storage decision from buying a compute, which traditionally was a peripheral. Learned a lot during that time. One was first and foremost, we had to define a market. EMC was really defining an element of what was considered the enterprise storage market. Over time, we had great success. I certainly learned the role humility plays and being hung up on 50 times a day, but continuing to pick up the phone with a smile and getting the next person on the phone. Eventually, we started hitting stride. The tectonic shift that really occurred in the market that really brought prominent role for enterprise storage was the internet. It had really started taking shape there.
One of the lessons I bring to the discussion today that we're having with customers is let's make sure we've got a monetization plan in place before we go invest millions and millions of dollars into a data architecture. Because you can tell when you remember the dot-com boom, there were people spending hundreds of millions of dollars on a business model that never had a revenue-generating outcome. You bring those learning lessons through time to have conversations with customers. The other thing I learned during that time, and again, this is the first chapter of what was the first 15 years of my career, was maintaining relevance with customers. At the time, we had a single product family. We had some software that helped customers with replication from a disaster recovery perspective. We were not capturing the low-end and mid-market.
We made an acquisition with Data General. Again, our relevance was important because customers had to have buy criteria for not only the high end of the market, but the low end of the market. Relevance had to expand beyond that. We made an acquisition. That is where we first met Frank and Data Domain because customers wanted a different way and a more transformational way of doing backup and recovery. We bought this little company called VMware, which is an interesting one, head-scratching back then. Today, boy, what an acquisition that was. One of my learning lessons coming out of that first chapter of my career was really what it took to define a market, but then what it took to maintain relevance as a leader in that sector. At the time, Pat Gelsinger was the President of the company.
He had been moved over to run VMware. I started to get an appreciation for how software was eating the world. I took a run. I had the good fortune of being mentored by Bill Teuber, our Vice Chairman at the time. I said, "Hey, I'd really like to go take a run at getting some software experience." He got me the right introductions. I landed a role in strategy and operations. At the time, VMware was still an incredibly successful product company. They were also faced with staying relevant. My first task at VMware was to build a specialist salesforce that was going to help define what was a software-defined data center, ultimately the elements of a virtual cloud.
If you look at any of the hyperscalers today, there's some pretty common architectural decisions, which is x86 is the atomic unit at which all software functions run on top of. We virtualized and we abstracted both compute, storage, and networking. That was my job at VMware, to build a specialized salesforce that can go and help customers build out a private cloud function that was as real and as priced and as performant as what they would get in the public cloud. Public cloud was obviously just starting to take shape. Customers started to figure out what was the role that public cloud was going to play because they've got all these assets obviously running and fueling their existing business models.
They saw this great opportunity from a time-to-market perspective where they could swipe a credit card and have a developer up and running in the cloud. During that time, again, it was a learning lesson of defining a market, staying relevant, and making sure that we gave our customers the greatest hybrid cloud experience. After 11 years, and we had just finished taking the organization through a big integration through an acquisition, I started to recognize that data was starting to take over the world. That, to me, was when I started to really start looking at the landscape of what was the next chapter of my career going to be. I typically do not move around often. If you look at the LinkedIn, you will see it has really been three jobs in three decades. I started to look at the market.
I had some conversations with some senior executives that I maintained close contact with. It became really apparent that Snowflake was an exciting company. There was an exciting tectonic shift happening with AI that was going to obviously put a huge dependency on having a good data governance and data foundation that was going to fuel the next generation of companies. Very much like if you think back in 1997, no one would have thought of a business concept like DoorDash and Uber. That foundational technology afforded new business models to emerge. That is the same opportunity we have today, sitting with customers and talking about how we can unlock the value of the data that they have, the assets they have to open up and exploit new business opportunities, monetization and commercialization of data.
That, to me, is one of the most exciting conversations I continue to have with customers because it's very easy as I come into this role, 80 days in, to try and wrap my brain around all the concepts and the data and the products. It could just be an overwhelming amount of information. Our customers are feeling this as well. There's a lot of noise in the market. There's a lot of competing strategies. My first role is really just asking the foundational questions when I've met with 35 customers and 30 partners. Why? Why did you make this decision? Depending on what industry vertical you're talking to, for the most part, and as you've seen time and time again, it's easy. It's connected. It's trusted. I had an opportunity to sit with a customer about a month ago.
I asked why the decision was made to go with Snowflake. I'm still in that kind of learning mode asking the five-year-old questions, why, why, why? This is when it really dawned on me the impact we're having. We had one of the largest contracting suppliers in the country basically walk me through what a bid process looked like. A blueprint gets submitted into a portal. A bid analyst downloads the blueprint. They start looking at putting a bill of materials together for the amount of lumber, the trusses, the doors, the windows, the thresholds, everything that goes into speccing out and building that house. They submit their bid. They get a 20% hit rate. It's okay. Where's the opportunity here? How can AI, how could we unlock more value for this customer?
We had industry vertical specialists sitting next to this customer and basically start prototyping what the value and opportunity could be. We took that unstructured data source, which was not in the platform, brought it into the platform, married it up with the contextual data of basically what they already had in their structured data set, and basically started prototyping an AI model that would go in front of that bid manager. We can get through 100 quotes in one day, which typically would take three days for a human to scrape through that architectural plan and put a bid in place. It was the power of unlocking the opportunity to say, basically, let's look at making sure all of your data sets come into this to make an informed decision, but increase productivity. We take it a step further.
We say, okay, once you win that bid, now let's have the next discussion around supply chain. Now you want to set the expectation with your customer on how long is it going to take for these supplies to arrive on site because ultimately, a builder wants to get this house built as quickly as possible so we can sell it. We start introducing partners like Blue Yonder into the conversation who are going to help customers with that supply chain component. All of a sudden, we've unlocked a business outcome that had never been conceived before. That, to me, is one of the most exciting elements we have in front of our market today, to help customers think differently about how they unlock value from their data. Again, these are business discussions.
This is a discussion I can easily have with a customer. Obviously, it takes a different way of thinking about their business models. I think it's most certainly true now, and you've heard this before. It will be the fast that eat the slow. It is most definitely not the big that eat the small. We're excited by helping customers unlock that value and giving them velocity in their business plans. Easy. You've probably heard this multiple times today. We can get a Snowflake instance up and running with incredible ease. It doesn't take an army of engineers to get this instance up and running. We can help you get your data into the platform very quickly, structured, unstructured. We marry it. We'll analyze it. We'll warehouse it. We'll build AI prototyping models and machine learning prototyping models against it. We'll start unlocking value. Connected.
We have instances where one of the largest investment banks down on Wall Street had a conversation about how they get their data from market feeds. It became obvious to me when they said, "Hey, we basically have this connection with Moody's and Standard & Poor's where we can get instantaneous access to data to start running our models against." We have one of the rating agencies that still continues to send us FTP data seven hours every night. That is where the slow is going to obviously have a disadvantage. You start thinking about the connective tissue of connecting a buy-side and a sell-side customer together and say, "Hey, how could we share data to accelerate the business value so that we can be more informed?" We start to see this kind of network effect happen.
That's one of the most exciting things I'm also starting to unlock is the value of the Snowflake network that we can provide to customers. Trusted, critically, the most important element of our value prop is the governance, the security, the data protection, everything that we put into making sure that the data, once it's in the Snowflake platform, is a vault. That is the, I'd say, largest apprehension most executives have when they think about AI is security. It's data privacy. It's governance. We spend a lot of engineering time making sure this is the most trusted platform in the market. This is the theme that we get from customers. When I ask those questions on why they're making a Snowflake decision, the most common patterns we see is it's easy, it's connected, it's trusted.
I have the good fortune of running the go-to-market organization. It's not just sales. That is not the only people in the organization, obviously, that we, I'll show you the organization a bit. We spend a lot of time with the sales organization. One of the things that I look at when I was interviewing for the job and I talked about with Sridhar and some of the board members about what I believe Snowflake needed to get to the next chapter of growth, it was scale and velocity. Scale in the sense of we could double and triple the salesforce, of course, to drive more revenue streams. We've got to act smarter about how we scale an organization, which is why I'm going to talk a little bit about our partner community and our alliances in a little bit.
Velocity is where we're starting to eat some of our own dog food. You saw today there's some announcements around Snowflake Intelligence. It was referred to earlier as Raven. I hired over 500 sales and marketing folks in Q1. Getting them as productive as possible, as quickly as possible, is truly important in a hyperscale market that we're operating in today. What typically would take a year to get a rep fully productive, I need in six months. How do I onboard a rep and put them in front of a platform such as Snowflake Intelligence and allow them to run an instant large language query about a customer that they're going to go visit? I'll use Caterpillar as an example. I've got a new rep on Caterpillar.
He goes into Snowflake Intelligence and he says, "Hey, tell me about the profile of my customer." It will go and dip into our database and say, "Here's the consumption they have on your platform, compute and storage. Here are the use cases we're using with this customer. Here are the use cases that have been identified in pipeline because we're looking at Salesforce Data. You've got a couple of open support tickets you need to be aware of." It starts to build a complete picture of what's going on with that customer.
Now, an exciting opportunity exists where you incorporate a technology like Canva into the equation and say, "Build me a curated executive summary so when I show up to that customer, I can show them a good profile of what we do and the business outcomes that we're delivering, as well as what we're doing with some other companies in the manufacturing sector." That's a productivity boost we could never have done before if we weren't eating our own dog food. I'm excited by what we're deploying. My goal is to make sure our salesforce is the most productive people that are leveraging AI and machine learning to make sure that they show up and they ramp up quickly. Our partner community is critically important to scale. I talked about scale and velocity. The two are the most important elements of, I believe, my job.
We get good scale from partners. We have our hyperscaler partners that are clearly the Amazon, the Microsoft, the Google, great partnerships with each of those organizations. We get some scale there. Our global systems integrators are giving us some scale. We've got three of the five that have committed to, in practice, a billion-dollar organization to basically build a practice around Snowflake because customers are drowning in data. They're looking to unlock value. They're trying to deliver business outcomes, but they can't glean enough intelligence about the data. I'm excited to see the kind of lift we're getting from our global systems integrators. We've also got a number of customers that have built a business on top of Snowflake that gives us great scale.
Companies like Fidelity that started as a great customer have unlocked and monetized their data and are using that as an opportunity to drive a new revenue stream. Helping a technology office now become a revenue stream through data and with Snowflake is an incredibly exciting opportunity. I have built an organization that basically supports these, what we call data cloud providers. We've got hundreds of these data cloud providers. The one area of the business I'm looking to unlock and activate is a true distribution network where I get what was those true value-added resellers. These are connections I've made over the years with VMware. I learned how you scale an organization through the channel. That is one of the areas I'm spending a lot of time focusing on, how I activate a distribution network through value-added resellers.
It's something we really, frankly, haven't done a good job at. I know that that will be the future on how we scale this organization. Around specialization, I've had a lot of experience not only building specialist organizations, but also understanding when you sunset a specialist organization. If I go back to my time at EMC, we had brought in Data General. We had mid-range storage specialists that were going in front of customers and talking about the differentiation of that particular technology. At some point in time, that became a core element to what our core sellers did. We didn't need that specialization anymore. Instead, we had just acquired Data Domain. We want to invest in specialization around backup and recovery. At times, one of my roles here is going to make sure that we've got the right volume of specialization supporting the business.
I'm spending a lot of time right now on investing in both technology specialization so that we can help customers unlock and build prototypes around AI and ML because they need help understanding the value and the vision around what we can do, but also our industry specialization. Having someone who's been in the corporate capital markets, who has been through a manufacturing supply chain discussion, bringing those business-minded specialists into a conversation so that we can unlock the value and vision of what we can do for them with data is hugely important. I'm getting really excited about we have an existing team, but putting a little bit more wood behind the arrow to make sure that specialization isn't just product. It's also the industry and the outcomes. As I've told my sales teams many times, good people sell product. Great people sell outcomes.
We're intent on making sure we're delivering outcomes. I've been confronted with a lot of complex issues throughout my career. I was taught this very early. There's a really very simple decision support matrix I will go through. Whenever someone comes to me with an alliance question, a partner question, a compensation question, you name it, it's pretty simple. What's the right thing to do by the customer? What's the right thing to do by Snowflake? What's the right thing to do by the individual contributor? It's very simple. 99.999% of the time, you're going to get the right outcome. I bring to this role a very customer-obsessed focus. Everything that we do is built around making sure our customers are successful.
Are we bringing in the right level of resources to deliver the right outcome so that we earn our seat at the table for the next workload? In a subscription model, unlike anything I've sold before, hardware product sits on the floor for five years. Software perpetual license contracted for three to five years. They own it in perpetuity. Then you're just selling maintenance and possibly upselling them more capacity. In a subscription world that we live in, you're only as good as your last use case. We have to show up every single day making sure the customer is getting value of this platform. That's where we're spending a lot of our time. Sridhar talked a little bit about this earlier as well, accountability.
Everyone in my organization, and by the way, when I got here, there was only half the team that was on a variable compensation plan. Compensation always drives behavior. The behavior we are driving is a customer-obsessed mindset. We are moving the entire organization, with the exception of customer support, to a variable comp plan because accountability has to be in their compensation. We have key metrics that are in place to make sure that the sales in the SE world is, quite frankly, very simple. It is a very binary outcome. Are you driving bookings? Are you driving consumption? Our partner community, our alliance community, they all have to have metrics. Our inside sales teams need to have metrics where we are looking at productivity on a weekly basis to make sure we are getting yield out of those assets.
Because when I go to Sridhar in a couple of months and I ask for potentially another couple hundred to a thousand heads, I need to have a strong ROI statement against that. I'm bringing a very strong operational and disciplined mindset to the business. Every person we bring into this company has to have a strong ROI. We can't have people sitting around being complacent in their role. This market is moving way too fast for us to be complacent in any way. I'm spending a lot of time right now talking to the teams about accountability. We're putting strong metrics in place. We're looking at this on a weekly basis so that people know how they're being measured. It's the most important thing you can do is making sure people know what to expect and how they're being measured. Absent of that, you get complacency.
We talked about the organization. We made a decision about a couple of months ago to bring the customer support organization into go-to-market. To me, this was critically important, not just to make sure the customer feels smothered with support, which is truly important in a subscription business, as I stated a little bit earlier. We got to make sure that there's a strong alignment with our customer support engineers so that after that use case goes live or after we onboard a new customer, they immediately know who to call if something goes wrong because something's going to go wrong. It's technology. We know that. It breaks. We got to make sure that that team is completely integrated and ingrained in the customer fabric and the customer experience. We made a purposeful decision to do that.
It's also critically important that those support engineers give Christian and Vivek the feedback loop on what we want this product to do. That is an important element to making sure we stay relevant, is that we're continuing to deliver value to customers and we're giving them the innovation they need. We've got a large organization built around marketing. Everything that you're seeing here this week, which is, to me, an inspiring collection of partners and customers coming together, I'm overwhelmed to see 20,000 people here. In my last role, I think at the peak at VMware, we had about 18,000. This is our sixth summit. We're already at 20,000. It just tells you how exciting of a marketplace we're in.
The draw we're getting for customers that are trying to figure out how they can put a strong data platform together and also leverage AI to unlock business value. We've got a strong professional services organization that helps get customers up and running once a contract gets signed. I'm being very clear with Ted, who runs this organization. We are absolutely not competing with our channel. Professional services' job is to help customers drive consumption. If you've got a competent partner that's involved there, my first preference, let the partner drive the consumption. There's going to be certain instances where it has to be my organization that does it in a highly regulated environment. Maybe on the first use case, we've got to prove it out. We're going to go in and do it.
My job, and what I'd say Ted's job is, is to make sure our channel is completely enabled and competent, make sure they're certified, they're competent, they know how to get this up and running. By the way, we've got hundreds of professional services. We call them true blues out there driving the platform and the consumption. To me, my first preference is get those partners activated, get them enabled. This is not a profit center for us. In my previous lives, professional services was a profit center. That is not the job of this organization. The sole purpose of this organization is to drive consumption. We've got a strong partner community. This is an area I'm going to continue to invest in, is making sure our partners feel like first-party citizens, if you will, as important as my customers.
I have to treat them that way. We have sales development, which is really a part of the initial interaction of we get 450,000 impressions a week. That leads to 13,000 interactions, phone calls, which leads to 4,000 in-person meetings, which inevitably leads to a new logo. We have a good supply chain. We're looking at areas in which we can leverage AI to increase those hit rates in terms of 450,000 impressions leading to 13,000 meetings. I want to get that number up because that's going to lead to new logos. We're doing a great job leveraging our technology to increase that count and that productivity.
One of the things, as I've come to learn, is it's truly important you've got a balanced business, balanced in the sense that we're driving net new bookings through new customer acquisition and balanced in the sense that we also have continued strong consumption. We actually have got to a place where we feel really good about our operating model. We've got certain skill sets and I'll call muscle mechanics built around how to acquire a new customer. It takes a certain mindset of an individual and a certain amount of resources and talent to get that new customer acquired and onboarded. As we get through that transformation and we start looking for how we're going to continue to maintain relevance, it's truly important we surround the customer with even more resources to drive expansion.
We've got a good model set up where we've got people that are just focused on purely driving acquisition as well as purely driving expansion. We've got a specialist organization that, again, goes through not only product specialization because we still are always going to have to compete for the next workload, and we can differentiate our value proposition from a technology and product perspective. Then bringing the industry organization people to sit with a customer and talk about in a manufacturing use case, "Hey, have you guys considered a supply chain outcome?" If we're sitting with a financial institution, "Hey, here's what we're doing with another bank to look for fraud detection or anti-money laundering." There are real use cases and real outcomes that customers want to hear from each other. The technology, I'd say, is kind of the easy part.
As overwhelming as it is, that's kind of the easy part. It's really unlocking the use case that customers want to learn from each other on. We're excited about the organization and how we're set up. It's yielding great results. I think that's a wrap for me. Again, I thank you for giving us the opportunity today to talk to you. It's my honor and privilege to be in front of you as the new CRO 80 days in. I've got a certain accountability on my shoulders, which is to continue to give you great earnings on a quarterly basis, which I'm dead set on doing. We feel really, really good about the trajectory we're on. We also feel really great about the vibe and the community of customers and partners coming out to this conference. Hopefully, you see and feel that too firsthand.
Thank you. Are we staying here?
Yeah. I'll invite Sridhar and Christian up to take a seat. As I mentioned, just raise your hand and the mic runners will find you. We have Mr. Scarpelli on the phone.
Here. Eagle in the center.
Right there. Thank you.
Good.
Hi. It's John Difucci from Guggenheim. And question is sort of for Sridhar and Christian. You both talked a lot about unstructured data in AI. You have also been clear on the conference calls that the strength in the business is really from the core data warehouses and analytics business. I'm just trying to understand all the work you're doing. Sridhar, you come from Christian, at least as long as I've known you, you've been here. That's what I think of you. Sridhar, I know you come from a different world. I'm just curious if you feel like you're there now where you can sort of play on a level playing field against what's at least perceived as you being the data warehouse in the cloud vendor. There's a lot left there.
Your competitor, Databricks, I'll say it because Christian did, is really viewed as the data science company. Is it more just perception now, or is there still work you need to do on the product side? Thanks.
Gosh, there's a lot to unpack in your question. What we have consistently said is that we have incredible strength in the analytics side of the business. It's the foundation, as evidenced by our continued growth, things like our net retention rates, and that our newer product efforts, whether it's in the earlier phases of the data cycle when it comes to things like ingestion, early data engineering workloads, that they're gaining strength, as well as in areas like AI. This idea that AI is the purview of a data scientist is very 2020 because AI is now the purview of the CEO because they want real transformation to happen. Similarly, if you're the Chief Data Officer at a company, you want to use everything possible in order to drive the kind of outcomes that you are looking for.
Just as a matter of practicality, unstructured data was harder to deal with before. No one in their right mind says, "I want to write a lot of code in a notebook on files that are sitting in S3 for me to understand what a piece of text meant." What we are able to do instead is tell them, "You can, using Openflow, connect Snowflake to a SharePoint repository, have data flow in, be able to create an index on it. These are one configuration screen, one command." All of a sudden, you have an interactive chatbot in the data that you can drop into a data agent, which now knows how to intelligently choose from, does it go against a quant data set, or does it look at the unstructured data. With respect to capability, I feel very confident about where we are.
If anything, I would say with things like our marketplace or the knowledge extensions, the stuff that you saw today, we are leapfrogging because, again, no one wants complexity. They want outcomes.
Nothing to add.
Not simplicity, ease of use has always been important and a strength of yours. That seems to play into everything here too. Thank you very much.
Absolutely.
Great. Hey, guys. Brad Zelnick, Deutsche Bank. Really great event. The innovation customer success is palpable. Christian, I tuned into something specific that you said when you were talking about Gen 2 and adaptive answering the age-old question or speaking to the age-old question of separating performance from revenue impact where Snowflake gets to keep some of the value and avoids becoming the next Teradata. I'm trying to reconcile that with open data formats and AI helping to reduce switching costs for customers. I may have missed the point, but why isn't competition and lower switching costs going to dictate price?
To a degree, it's not competition. How I framed it was we want to be the leader in price performance, which is why my statement is if any of you goes and update models and say, "Oh, we're not going to go get the performance headwinds," that would be the wrong conclusion. That is why I was very clear on the no, no, we'll still continue to pass on benefits to our customers. What we did with both Gen 2 and adaptive is now we're in control of when and how much. What is really going to guide it is our position on the price performance leaderboard. That is how we think about it. If we need to go pass it all overnight the way we've always done, we can do it. If we do not have to, we do not have to. Does that help?
It helps. Thank you.
Hey, guys. Right next to Brad, Alex Zukin with Wolfe Research. I want to maybe ask from a little bit of a different lens. You've got two tech questions. It does feel like the vibes are different at this conference this year. It feels like the spend intensity. It feels like the demand environment. It feels like something has changed. I'm curious, is it as simple as, "Hey, these AI budgets are having downstream significant impacts. Customers are realizing that they build AI on data, not on models." Snowflake and your peers are both kind of rising tide benefiting. Is it a product intensity, innovation velocity that folks are reinvigorated by that you can attribute to? Maybe just explain kind of the post-stress moves and why this is a really important part of the dynamic this year.
Lots to unpack. I'll start on the first part, and Christian will add on and speak to post-stress. There's always feedback loops in life where if your team, you yourself feel like work that you're doing is having a positive impact, that spurs more work because you can see that positive impact. I've been open about the fact that we went through a tough phase last year. I'm pretty proud of the fact that we stood up to it and produced better outcomes with the things that we knew best, which is the hard work in sales to get things landed, to deliver value, and then the things on the product side to create world-class products and then do the hard, sweaty work of telling every customer about it. To this day, whenever I meet a customer, I tell them, "You should let us do a half-day workshop.
It'll cost you nothing. It'll show you what's possible with Snowflake and AI. Then we'll talk about where we can create value. Thanks to things like that, I think there is a real momentum for Snowflake both inside and outside the company where the team believes they're headed to a better future. All of our customers have the same belief as well. Remember what I told you folks earlier. Every person that's betting on Snowflake is kind of betting their career on Snowflake. The more they see, "Oh, wow, here are all the other things that I can get done as a result of my bet on Snowflake." In fact, it's going to help me navigate the AI wave as opposed to me having to stitch stuff and answer awkward questions about what I'm doing with AI. Snowflake is very much a part of it.
I think that's the message that you see resonating now with every constituent. You see this outside. You see this in the announcements. We get customer feedback from tons of people, including you folks that helpfully compile feedback and send it to us. We see it in that as well.
Yeah. Briefly, I think the journey customers have been on is also informing how some of the items that we say resonate. A year ago, we said, "There's no AI strategy without data strategy." Some people say, "Oh, that's because you have nothing in AI." We also said, "Oh, you want your AI to respect your governance, and governance matters, and permissions matter." People, "Yeah, yeah, I can put my ChatGPT on the data." You realize all these things are true. That is where the post-risk conversation comes in. "Oh, I want to be able to store state. I want to be able to personalize my agent." Where do you do that? That is where all this.
I think there's been a, I don't know if it's, maybe it's the hype cycle, but a point where people now have tried to roll out AI, and they realize that many of our core thesis is true. Governance matters. Not taking data out matters. Not making copy matters. I think that's what you see reflected now.
Being world-class is essential. You folks know this. We have now gone through what appears to be, what, three generations of vector databases. These are all companies that purported to be the next Snowflake. I mean, I wish them well. A vector database is much better as an addition to Snowflake where it's like one command to create an index. We have to walk the walk. That's where we're able to lead with the strengths of what we as a data platform offer.
I think you also look at the customer stories that we're starting to generate. I talk about that supplier organization where we unlocked tremendous value for them where their bid officer can put together 100 bids a day versus one every three days. Those stories are getting out there. Customers talk to each other. The most powerful sales force I have is my customers. When they talk about the success of what we've unlocked for them, the ease of use, the connected tissue, the trust, those stories are probably my greatest sales assets out there. It is also how my team shows up. We show up talking about the technology. I won't say we've lost it. That is the rocket science. That is kind of the easy part.
It's how we help customers think about their processing of the data, the people, the business ideas that are going to inspire the next revenue stream. That is a conversation we have to start with. I feel like we tend to end there. I want to change the way we show up for our customers and have outcome discussions so we can unlock value. We can show them great reference use cases. Then we get deep into tech, and we have to validate it.
Thanks very much. Kirk Materne with Evercore ISI. If you came here a few years ago, I think it used to be we would talk about data warehousing, and there's always importance of data. It seems today that through the simplicity of the platform, you all are able to talk about solutions and business outcomes. The question is really, are the people you're talking to today versus three or four years ago different in terms of their ability to make higher-level decisions, move faster with you? Many used to be sort of, "Well, we'll move one data set at a time." Are they looking to move at a faster pace because they now understand the business relevance of all this? I guess for you, Mike, how much farther do you have to go in terms of specialization, whether it's verticalization?
I'd also just love you to just chime in really quickly on Europe and what you're doing internationally where you're still obviously a lot to go there. Thanks, guys.
I can't speak for last year in terms of the buying profile and meetings we've been on. I can tell you within the first 35 customers I met with, I think only one of the meetings had a Chief Technology Officer in the discussion. It was lines of business. It was the CEO of a large publishing company in New York who wanted to have a discussion around, "Tell me how I can monetize and commercialize my data." Again, how we show up and the people we're talking to are people that are trying to drive new revenue streams and optimize the organization. In terms of the investment specialization, we have foundational people in place today.
We are definitely going to be investing more in those specialized business analysts or specialists as well as data specialists so we can show up and not only talk about the business outcome, but actually have people prototyping within a couple of days how we can unlock value and showcase that outcome and how we can actually let the data unlock value. Europe and APJ are two growth markets for us. Those are my fastest growing markets. Again, a large portion of our revenue still comes from the U.S. We're investing heavily in Europe and APJ. Europe, we've got a European leader that just onboarded. We've got 23 new managers, so we're expanding. It's obviously 138 countries we have to cover. We've got a great base of individuals there. Again, my ramp-up and productivity is my primary interest in those markets.
How quickly are we bringing on people to those markets, and how quickly are they being productive and having a business outcome discussion and then bringing in the technology discussion to marry it up? Same goes for APJ. Those are unbelievable growth markets for me.
Good afternoon. Brent Bracelin with Piper Sandler. I wanted to go back to the pricing conversation around Gen 2. What's the right value capture for 2x higher performance? Can you capture 20% higher price for 2x higher performance? Is that the right way to think about it? One. Then two, maybe more importantly, what portion of the installed base is performance the number one priority? Is this something that maybe half of the installed base as you contemplate might move to Gen 2 versus sticking with the Gen 1? Or is it more of the installed base that potentially would move towards kind of Gen 2 and eventually these adaptive tiers?
Yeah. Two thoughts, Brand. One is the price incremental from Gen 1 to Gen 2, that's a fixed ratio. That's the easy part of the equation. The benefit is not the same for everyone. That's where it gets difficult. That's why I said, "Hey, the benchmark and the number I'm sharing is representative for some workload." For example, if someone has an update-heavy workload, they'll see numbers way better than what's in here. If it's a workload that is more read-only, they'll see slightly lower than that. That's the piece that makes it hard. We've calibrated in a way that customers will for sure see better price performance and ideally see an economic benefit relative to where they are. It's always easier for us to lower it later if we want to deliver more.
Back to us being in control on the when and the how much, going up is always complicated. I do not know that there is an easy, "Yes, this is the split," because it is a customer-by-customer conversation. We are starting at a point where we think is conservative and roughly revenue neutral or a little bit lower. We can go and adjust based on what we see. The performance is going to be materially, materially better for pretty much all of our customers. My expectation is that everyone is on Gen2 or everyone, like 90%, 95%, 95%, a large percentage of our customers are there within months.
The only thing I'll add on to this is that there are second-order effects from performance improvements because people are willing to do a lot more. Things finish faster. They're able to get more work done for the same unit amount of either time or the dollar that they put in. It is our ability to roll these out that is more important than exactly how much performance because, as Christian said, that varies depending on the workload.
That's a great point, which also talks about your second part of your question on sensitivity to performance. I think the willingness to open or bring new use cases to Snowflake are influenced by that. I just met with someone after the keynote. A large company said, "Given some parameters, I'm going to use my internal customer support agent, like internal traffic, on Snowflake. If the latency is lower than some other thing, I'm going to put a customer facing.
Thank you.
Hi, Kash Rangan here with Goldman Sachs. Zelnick and Zukin are typing their research reports, and you'll have to use AI to interpret what they said. First of all, congrats on Summit, Christian. Great to see you present today. And Mike, welcome. Scarpelli, hi.
Where's Scarpelli?
Hello.
There you go. Got you, Mike, to say something while posing me the question on the mic. Scarpelli, I will spare you of the question. Mark Murphy challenged me to keep my question to one question. Let me see if I can win the bet. It is a glass of wine, Mark. Consumer tech hides all the complexity. You have seen OpenAI. You had Sam Altman. It has had so much massive adoption, captured the world's imagination in such a short span of time. Sridhar, you have been in consumer tech before, at Google. Christian, you too. Why has the magic of generative AI, or maybe there is a magic formula that Snowflake has? This is a question I have been asking other CEOs as well. Why has it taken so long in enterprise with respect to adoption of AI?
What do you think is the key to unlocking the magic that ChatGPT and other consumer apps have unleashed in the consumer world? Thank you so much.
Yeah. It's a great question. First of all, I would say that when it comes to things like meeting preparation, when it comes to a lot of things that we do commonly, I think generative AI has basically penetrated most companies, most of our consciousness. We all get the utility of it. Until we launched Anthropic inside Snowflake, for example, obviously, you can't take something that's confidential and put it into ChatGPT, even if you're paying for a subscription. I think it is easy-to-use products like Snowflake Intelligence that don't require you to make compromises that I think are truly going to unlock large-scale use. OpenAI is doing a little bit of it. They have a set of enterprise deals for their enterprise products that are driving broad adoption.
To me, the magic of AI in the enterprise is going to come from all of the data sets that matter to you, whether it's sales information, consumption information, HR information, or stuff that's sitting on Drive or Box or Dropbox. When all of that becomes accessible, I think that's when the magic truly unfolds. That's what we are excited to be driving. I don't think it's that far away, Kash.
I would add one quick thing. I also agree it's not far away. I also think the tolerance for incorrect results is lower in a consumer world than in an enterprise world. We all remember the original ChatGPT. It was hallucination left and right, and it was OK. In the enterprise, that is less OK. A lot of what we're sharing here at the conference and our overall strategy is, how do you do AI, but with the trusted piece that Mike Gannon talked about? That is the piece. I think they're all finally coming together. Now you have the experience in Snowflake Intelligence, as Sridhar said. The time is near.
I think there's a great most executives worry that they don't have the prerequisite skills they need to drive an AI strategy. That's one of the things that I'm most excited about is kind of simplifying. We can help solve this for you. Let's talk about business concepts first. We'll make the tech the easy part. I think there's also generally a skills discussion that most executives are worried about. Secondly, to what Christian said, security, data privacy. Let's make sure we have all those controls in place. That's truly important because we all know we've been around tech long enough. When it comes to warehousing, garbage in, garbage out, what data is going to be most critically important data, but accurate data that we're going to build our models on?
We're spending a lot of time talking to customers around the data and the governance of the data that goes into the platform. Obviously, spending a lot of time talking about how simple large language models can make unlocking data or unlocking value from that data really simple and easy to understand. I think it's skills, and I think it's privacy, security.
Karl Keirstead at UBS. Sridhar, what do you think of the SaaS firms moving into the data layer? Obviously, Salesforce derives $800 million-ish from data cloud now, just offered to buy Informatica for $8 billion. Bill McDermott at ServiceNow talks about workflow data fabric. SAP is pushing their business data cloud. What is the signal there, in your judgment, that we should all recognize? In particular, are there data use cases that, in your judgment, the SaaS vendors have a legit shot at winning? Conversely, are there use cases that will never truly go to the SaaS vendors that need to go to data specialists like Snowflake?
This is a great question. We spend a lot of time debating this question. I think a lot of their behavior is driven by AI. I think in the case of SaaS vendors, that makes software that assists humans in getting things done. Customer service is a classic example. I think there is an existential risk to that model of making that kind of software, which is that software that is based on things like agentic AI can be a wholesale replacement for not only the software that they make, but also the people that are doing this. I think for a number of SaaS companies, getting into agentic AI is not an add-on to their business model. It is a threat to their business model. I think that, in my mind, exemplifies the interest that they have in the area and why data becomes so important.
Look, we collaborate actively with all of these folks. We have bidirectional data sharing with Salesforce, with ServiceNow, and we are working on a close partnership with SAP. What we offer is a place where all data can be brought into one place. That is what we do. We are very good as a data platform. It is a hard skill. Just like making a model that is world-class is a hard skill that only some companies can truly be world-class at. I think building a general-purpose data platform is very different from building a SaaS application. That is something that we are very good at. I think the kind of use cases that you will see SaaS companies capture will be along the lines of what I described, which is agentic versions, like insight into data about processes that they drive, generally speaking. Obviously, they have bigger aspirations.
Part of the reason why we make this bidirectional is we need these partnerships to be win-win so that they feel like they're getting what they want out of it as well. You can move data from Snowflake onto one of these data clouds. We feel confident about the value add that we bring to the table as that central place that can offer the 360 view, be able to act on it. I see interoperability happen at multiple levels. With Fabric, for example, we collaborate at a bottom level where we are able to read a table that is sitting in OneLake, or somebody that is in Snowflake can say, "Hey, I want to store this table in OneLake." We also work with them at the top level of agent components that are created with Snowflake Intelligence. We'll interoperate at that level.
You're going to find this kind of cooperation and competition at both of these levels.
I'll add one thing. Hi, Karl. Most organizations have two or more of these SaaS applications, which creates a problem. Because Benioff says, "Oh, you want to combine our data or the data you have with Salesforce with the rest of your data? Give me the rest of your data." Then McDermott, an hour later, goes and says the exact same thing. You have CIOs, CTOs, and CDOs saying, "You want me to copy all of my data into these different platforms?" That's a massive no-no from efficiency, cost, governance, et cetera. I talked to the CTO of a bank recently and said, "There's no way I'm going to go put my data multiple times in one of each of these platforms." I think that is the single biggest gap in that thinking.
Thank you both.
Mark Murphy with JPMorgan right here in front. Mike Gannon, I wanted to ask you. Great presentation, by the way. You mentioned it, but the number of sales and marketing hires in Q1 was rather stunning. When you look at the history and the sequencing, can you clarify how much of that reflects just there's been a step up in the pipeline, so we need feet on the street to go get it, versus maybe you coming in and saying, "Our ratios have been off for a while." Can you speak to that type of hire? Should we look at that and say, "These were the specialist hires coming in where they would take a while to ramp"? Should we look at it and say, "These are people you're bringing in people that sell farther up the totem pole and would take longer to ramp"?
I'll just provide one bit of historical context, and Mike will answer the question. I want to make sure that all of you juxtapose the hiring numbers that you saw in Q3 and Q4 with the numbers in Q1. We went through a period of just cleaning house. We called it a get-fit initiative, where we really focused on accountability and performance. That was an ongoing thing. It was combined with the numbers that came out in Q1. That context is important for you to remember.
Yeah. It's a great question, though. 70% of those new hires were additional capacity. 30% was really what I'd say performance management, people that were just complacently sitting back waiting for business to show up on their lap. That was a large part of the activities that happened in Q3 and Q4 of last year. In terms of the question about the hiring profile, we're actually leveraging some modeling right now. We're looking at our most productive reps, the reps that are consistently overachieving. We're actually scraping through LinkedIn to figure out the hiring profile. What's the most common, most successful backgrounds that these individuals are showcasing? That then becomes the new informed profile of the rep that's going to carry us into this data and AI space. We're spending a lot of time. We're a data company.
We do this really well, where we actually can do a lot of intelligence gathering on the most productive reps, and frankly, even the reps that are maybe struggling. There is a consistent profile we see there. We are spending a lot of time looking at this. I mean, productivity, to me, in this space that is moving at a blistering pace, we cannot afford a mis-hire. We are spending a lot of time on not only just getting the right people hired. That is critically important step one. Step two is leveraging the productivity tools that I talked about around Snowflake Intelligence to get them productive within six months, as opposed to what historically may have taken a year. In terms of the growth projections we have put out onto the market, that has all been factored in based on the capacity I shared with you just now.
That has been factored into the results that you're expecting to see at the rest of this year.
Gregg Moskowitz from Mizuho. Christian, just getting back to the notion of keeping some of the value for adaptive compute specifically, some of our conversations with partners and customers have indicated an expectation of a significant cost savings just because they can purchase smaller clusters and flex them. I completely appreciate that this should drive more consumption on the platform, which is terrific. Might we see an initial revenue reset once we get to the point of adaptive being GA? How are you sort of thinking about this?
We have not finalized the pricing level of adaptive. One of the reasons why I caveated that we are going to go slow and see how this works, part of it is a technical reason, but the other one is we want to make sure that we understand that consolidation effect. Right now, all the guidance and numbers that we provided are factoring all the information that we have. We do not expect to go and see the top line change dramatically because of this. We believe that a lot of the value of adaptive comes from the ease of management, the TCO, et cetera, not necessarily because we need to go and effectively lower the price.
I think Jimmy says last question.
Yeah, one more.
Last question.
All right. Patrick Colville here from Scotia. I'll make it a good one, Jimmy. Two years ago at Summit, you coined the AI and data cloud message. Back then, it was right at the beginning of this kind of AI adoption. We're a bit further along now, and we're starting to get a sense of these kind of reference architectures in AI. You can't walk one block in San Francisco without talking about MCP. You guys and your closest peer bought a Postgres database. OpenAI was on stage yesterday. Sridhar, you're very deep in this ecosystem. I guess, where do you see the AI reference architecture as of today, and where will it evolve to in 12 months? Where can Snowflake kind of move towards the puck where it's going in 12 months?
This is a great question. I think foundation models continue to evolve their capabilities rapidly. I think what they are able to achieve with post-training, achieve with some amount of time, like what they can do with deep research, is I think truly, truly remarkable. I continue to be blown away by how that kind of tech is able to do so much just acting on the open web. From an enterprise context, I genuinely think that having these world-class models then paired up with the right sources of enterprise data, with the right reference architectures for interactive applications, which are fundamentally just incredibly different from, "Here, do some research for me, and it's fine if you take 10 minutes." Those are just not the same things. I think that's the stuff that'll get settled over the next few quarters.
At one level, yes, we want AI SQL to happen because that is data processing at scale where you're able to crunch through a million rows. Whether you take three minutes or four minutes does not quite matter. On the other hand, if you are building an interactive chatbot that you want to power customer support, then every decision, how long that search index takes, what is the right-sized model? You do not want to go pick the biggest model for that one because it is no longer quite so interactive. Or what is the underlying store for analytic queries, but now being served out in real time because you want that data fast? I think those are the knobs that are going to get tuned over the next few quarters. That is also how we internally think about this.
Part of what, like even with Cortex Analyst, it takes a few seconds to give an answer. It cannot be that way if I truly want all of you to be able to use it freely and fluidly for the kinds of things that you want to do. I think there will be a set of these foundational components. There will be things like tool calling. I think web search continues to be a really powerful repository for anything that is recent, anything that is not part of your enterprise scope. I see very much composite tooling coming in, which knows what to talk to where. Where we need to partner for these cases, we will absolutely partner. We will continue to partner with the foundation model labs that continue to outpace others in terms of what they're able to bring.
When it comes to all of the data that is proprietary to enterprises, that matters to enterprises, we feel very good about being able to figure out what's the kind of product and interaction model that is needed for them to be effective in creating products for enterprise use. That is what we are driving towards. Jimmy, come on up. I think we're done.
Thanks so much, everyone. Katherine and I will be hanging out after to answer some questions, but they have to go to a customer event, unfortunately.
Thank you all. It's great to see you.