Snowflake Summit 2026 | Paul Winsor, Snowflake & Gowtham Gundu, Fetch

Clips
More from Snowflake Summit 2026

Paul Winsor

Global Head of Retail

Snowflake

Gowtham Gundu

Chief AI Officer

Fetch

play_circle_outline Fetch Aims to Be the Rewards Destination: Targeting 1-in-10 Users, 13M Daily Receipts, 450B Points

play_circle_outline Fixing the broken value exchange between CPGs/brands and consumers through rewards

play_circle_outline User experience and retention: fun, sticky receipt scanning and gamification drives loyalty

play_circle_outline Snowflake as foundational data platform: ingestion, lakehouses, clean rooms, CoWork

play_circle_outline Agentic Commerce: AI-Driven Personalization, Propensity Modeling, and Dynamic Reward Allocation for Fair Marketplace Economics

Info
Transcript

Paul Winsor, Snowflake & Gowtham Gundu, Fetch

Paul Winsor

Global Head of Retail Snowflake

Gowtham Gundu

Chief AI Officer Fetch

In this interview from Snowflake Summit 2026, Gowtham Gundu, chief AI officer at Fetch Rewards, joins Paul Winsor, head of retail EMEA at Snowflake, to talk with theCUBE's Dave Vellante and Rebecca Knight about how governed data infrastructure and agentic AI are reshaping the future of retail commerce. Gundu, who spent nearly two decades leading AI at Google before joining Fetch Rewards, details what drew him to a platform processing 13 million receipts daily — a transactional data footprint rivaled only by Walmart and Amazon. He explains how Fetch is using S... Read more

explore Keep Exploring

What excited you most about the opportunity at Fetch, and can you describe Fetch and your role there? add

What convinced you to consider (or get involved with) this company/opportunity? add

What makes the company's rewards/loyalty program unique and successful compared with other reward programs? add

Which should be prioritized to grow Fetch's business—foundational technology, marketplace dynamics/user growth, or partnerships—and what role does Snowflake play in Fetch's data platform? add

What does automation in advertising involve, and how will it change the way brands target, reward, and meet users where they are? add

bolt Powered by CUBE AI

Paul Winsor, Snowflake & Gowtham Gundu, Fetch

search

Rebecca Knight

>> Good afternoon everyone and welcome back to theCUBE's live coverage of the Snowflake Summit 2026 here at the Moscone Center in San Francisco. I'm your host, Rebecca Knight, alongside Dave Vellante, co-founder, co-CEO of theCUBE. I would like to welcome our next guest to the show. We have Gowtham Gundu, Chief AI Officer at Fetch. Welcome, Gowtham.

Gowtham Gundu

>> Thank you. Thank you very much.

Rebecca Knight

>> And Paul Winsor, global head of retail at Snowflake. Welcome, Paul.

Paul Winsor

>> Welcome. Thank you.

Rebecca Knight

>> Happy hour is taking place right now.

Paul Winsor

>> I know. Yeah.

Dave Vellante

>> They just want us to lose our voice. That's all.

Rebecca Knight

>> Exactly, exactly. But Gowtham, I'm going to start with you because you led AI at Google for nearly two decades. What excited you most about the opportunity at Fetch and then also tell our viewers a little bit about Fetch and what you do.

Gowtham Gundu

>> Yeah, absolutely. First of all, thanks for having me and what an incredible event. I ran into Sridhar and I was telling him the amazing keynote that he had yesterday and the breadth and depth of the event. This first time me being here, and I was so impressed. And he was like, "I'm also actually impressed." So that's great because obviously Sridhar was our leader at Google and I was always inspired by the way he used to work and I'm sure Snowflake has a great leader. But the question that you asked is a question that I get a lot because almost like being at Google for two decades joining Fetch, but I think there are three important factors that led me to kind of making this decision. The first and foremost important thing is the mission and the vision for the company. Obviously I want to work for a company that has a mission that is like really impactful, but also satisfying when you work on. And also what like Larry Page used to say in the early days of Google, "Work on something that is uncomfortably exciting." And the mission for Fetch is to become the rewards destination for everyone. And right now, Fetch is like serving one in 25 consumers across US and we have the ambition to serve more than like one in like 10 consumers in the US. We're already kind of making strides towards our vision. We are giving away almost like 450 billion points every single year. And then we see almost like 13 million active users every month in our platform. And you would be surprised to know that every single day, we see almost like 13 million receipts. So only after Walmart and Amazon, we are the third biggest in terms of the transactional GMV data that we actually see on our platform. So I mean, as you can see, the mission and given the scale of the platform, that was definitely one key thing. The second thing, which our CEO really describes it as the broken value exchange. Today, the CPGs spend hundreds of millions of dollars trying to reach their consumers on various platforms, but hardly any of that goes to consumers. But then when somebody like a Pepsi or someone spends money on our platform, we actually give away a good chunk of that in terms of points to the users. So what our CO says it as like we are fixing that like value exchange and we are creating a bargain for consumer. And the third thing is because of the scale of our platform, we sit on this huge amount of data which can create a lot of difference for all these brands that are trying to reach to these consumers. So all these three factors and how much AI can impact all of these three things that I touched upon, it was kind of like no brainer for me to consider this.

Dave Vellante

>> How did you achieve ... I mean, I understand your ambition is to get to 10%, but you're at 4% now. How did you achieve that? What was the secret to achieving that, even that scale?

Gowtham Gundu

>> Yeah, that's a great question. A lot of it predates to me, so I don't even take credit for that. But I think if I understand the history of the company, they've tried a few things during pre COVID, but I think there hasn't been a company in the reward space that makes it both fun at the same time while like earning the points. For example, imagine this experience of somebody coming, opening up the app and then scanning a receipt. It's much more than just the points. And then they also get points for every single receipt. So I feel like they've figured out a way of doing loyalty in a way that is like fun and adding value to the activities that you're daily doing. You go and you shop. Now at the same time like scan. Oh, you go and play games. So why don't you open Fetch and then download a game and play and whenever you reach certain levels. So we're trying to meet user where they are. I'm sure I'll touch upon this like later as well. I think because it's so sticky and it's creating this fun and joy moment and also giving a rewards, I think that's how we have gotten here.

Dave Vellante

>> Paul, we love the points, right? The airline points, our hotel points, our credit card points, give me more points. And if we can make it more fun, is this the future of that business? How do you see it? What role does Snowflake play in Fetch's success?

Paul Winsor

>> How do I see, sorry, what?

Dave Vellante

>> How do you see the future of that business, and what role does Snowflake play in Fetch's success?

Paul Winsor

>> Yeah. Well, I think really good comments in terms of the way that Fetch is using Snowflake today. I think the opportunity to use the platform to really understand how customers are using those rewards is really, really important. We've got a lot of customer conversations we're having today around how consumers are thinking about how they save money and how they spend their money from regards to their disposable income. So understanding and attracting conscious consumers is something that we are speaking to a lot of our customers about. So I think the customer side of understanding that is really important.

Rebecca Knight

>> Gowtham, can you talk a little bit about what you're actually doing in terms of targeting customers and learning about what their preferences are and how they are spending money and saving money and therefore which brands they gravitate toward and which ones they avoid?

Gowtham Gundu

>> Yeah. Is the question more targeted about the technology or the trends?

Rebecca Knight

>> Yeah, both.

Gowtham Gundu

>> Okay. Yeah, I think this is a marketplace. Like any online marketplace, you have user on one side and you have the partner on the other side. So how do we get the right kind of balance within the marketplace? And of course that involves using AI as a technology and something like Snowflake as a foundational data layer to be able to get there. So what we try to do is really through the data that we have tried to understand more about user, and we realize to be able to get to our mission, we need to be able to meet user where they are. So when you think of that, it is like less about like traditional personalization. Of course, that is table stakes. But also like really understanding the deeper user propensity, like which cohort of users care for like what kind of behaviors. And then on the other side, one, you're kind of like getting the brands and providing them the value, but then also you need the right kind of demand so that you are able to provide value across all the user base. So that's why we are into various kind of CPGs, the restaurants, the fuel. And then now we are going from enterprise all the way into the mid-market and SMB. And as you know, this is a 10X and 100X number of customers that we have to reach when we go from enterprise to SMB. This is, again, where we are like resorting to AI and some of the technological advancements that Snowflake and others are enabling us to be able to get to that scale.

Dave Vellante

>> Paul, what do you see as the bell curve, if you will, in terms of, let's call it the haves and the have nots? The companies that are democratizing AI, driving AI throughout the organization versus those that are maybe saying, "Well, let's wait and see. It's kind of confusing." Can you discern any patterns at this point?

Paul Winsor

>> Yeah. I mean, I think 2026 has been a year where customers that we're talking to today, the opportunity to now sort of speak to your data and turn it into a conversation, and have these personal agents kind of working alongside knowledge workers. We're seeing that as a really exciting opportunity for a lot of customers. We heard them talk about it today. I think we had Samsung talk about it on the keynote today, as well as Under Armour in terms of how they're adopting this intelligence around interacting with your data from a conversation perspective. I think it's time to insight. I think this is what is really, really interesting is how quickly you're seeing that knowledge workers can get answers and the context to the answers quicker through things like Snowflake CoWork. I'm really excited about Snowflake CoWork. I've been in the retail industry for 40 years. So I go all the way back to mid 1980s working for a retailer out of the UK, which is one of the largest supermarket companies. And if you wanted an answer to a question way back when it was just the batch era, it would take you weeks to get that answer. What we're now seeing is this opportunity to be able to ask a question, get an answer, get the context of that answer before you've even pretty much started your day's work.

Dave Vellante

>> I mean, state of the art back then was placing beer next to diapers. Remember?

Paul Winsor

>> The classic story.

Dave Vellante

>> Right, the classic story. My wife the other day was so frustrated. She got to the cart, whatever we ordered, only one came. And I said, "Why'd you only order one?" She goes, "Well, I got to the cart and I realized that they told me that it was out of stock so I had to go all the way back." I said, "Just give that some time. We're going to fix that. AI will take care of that." So having spent all this time at Google as its head of AI, you can see the potential. We were chasing AGI for a long time. Let's put that aside. I like to call it enterprise AGI. We're going to fix all these enterprise problems. So what do you see as the future of retail maybe generally, but specifically for the rewards business? Is your intent to consolidate that. 1 in 10 is still pretty distributed. So what do you got to do to get to that 10%? Is it partnerships? Is it better data? Is it better product?

Gowtham Gundu

>> As always, it's an and. It's all of them. I think we have to get better at the foundations and we have to get better at improving these marketplace dynamics, but also the partnerships. So there's user growth, there is partnerships, and then you have the technological layers. And maybe I can start from the bottom, which is again, Snowflake. Fetch's data platform is purpose built for scale, and Snowflake is at the center of it. And today, it adds value in multiple different dimensions from the data ingestion. I think we already touched upon that. We see like 13 million receipts. Receipts are completely unstructured and unorganized. There is no one single catalog that can capture all the retail networks in the world. And then when you look at a receipt, they will have words like PPSI-CH. So it means like Pepsi cherry flavor probably, you don't know. It's very subjective. Imagine starting from these messy unorganized data and to be able to get to this data that acts as backbone for serving all of them. So Snowflake and the other data ecosystem, the lake houses, the clean rooms and all of these things of course enable us to kind of do a lot of stuff. Then on top of it for users, it is really the personalization and the experience, the new kind of experiences. For example, because of all the things that AI agents can do in terms of like support and everything, like for example, somebody's not getting points for their receipt scan, they have to file a ticket. But now we do have agents that can automatically look at them and then probably even try to resolve themselves and reward the points, right?

Dave Vellante

>> Yeah. Why have a human do that, right? I mean-

Gowtham Gundu

>> Exactly, exactly. And on the partner side, to Paul's point, you have data. If you look at the layers from the data you have insights. Then you have the measurement MMM and then activation. I think through AI advancement and then what some of Snowflake kind of technologies are all enabling, you can have impact across all these layers. And I think we are focusing on, like as I said, always given a marketplace, both on the user, kind of provide them more value and do everything to kind of increase the kind of experience that user has. And again, I'll keep on saying meet user where they are.

Rebecca Knight

>> So what is that though? In terms of what Paul was describing with at the grocery store, being able to see the insights, now we can get the answers. So what does the brand customer relationship look like when there is more context, more personalization beyond just, "Oh, I didn't get points for that. I need to be reimbursed." But what is it really going to change?

Gowtham Gundu

>> I think, in some ways, the industry has been pushing for this, which we call automation. If you remember a few years ago, it was like the advertiser or the brand has to come and say, "I want to target people who are in market for X, Y, Z, but this is the amount of money that I want to spend." And there was a lot of controls. But then later they realized that is not creating optimized experience for users, so this is what we call automation in the space of advertisement. So if I have to simply describe what it means, what the future holy grail is, it's like unimaginable amount of automation. For example, if you take Fetch Rewards, it's a place where we want to be, you get up in the morning and we automatically be able to give you reward for certain kind of activities like walking and then watching your favorite show. Or you go to a grocery store and we know that based on your history of purchases that you actually need to buy some stationary today and we are going to automatically give you that kind of suggestion over to you. This is exactly what I mean by like meeting the user where they are. But once we saw this, brands are also able to achieve what they want, right? So I think that's the relentless focus. But then to get there, there is layers of like data understanding and then matching the offers to users, and then also creating an economic system that can self-learn when to reward and how much to reward. Like for example, if somebody's already buying Pepsi versus like you want to sell this to somebody who hasn't tried Pepsi, you need to kind of provide them different kind of rewards system.

Rebecca Knight

>> Right, right.

Gowtham Gundu

>> How do you automatically generate these points in a way that's still like fair to the overall ecosystem?

Dave Vellante

>> And you can rapidly test that. Go ahead, Paul, please.

Paul Winsor

>> Well, I was just going to say, I mean, data collaboration again, 2026, it's not going away. Data collaboration is accelerating. So this opportunity to kind of share customer data from a very privacy PII perspective, but through data clean rooms with brands is happening in 2026. We're seeing some very, very interesting partnerships taking place on that. But also on data sharing just on the ability to understand how customers are shopping products as well. This is something that's accelerating through the platform. And it was great, again, to see today in the keynote that we are accelerating our data sharing capabilities for multi-partner data sharing. So I think that's going to be very interesting going forward as well.

Dave Vellante

>> How do you think about the future of retail? I mean, we talked about Walmart and putting batch jobs, beer, diapers. They survived the storm and are thriving, no doubt. Amazon, obviously a juggernaut. But every time we get one of these waves, things change. And you've seen some of the smaller malls, retail's Been hurt, a lot of transition.

Paul Winsor

>> Yeah.

Dave Vellante

>> How do you see the future of retail?

Paul Winsor

>> Well, I think the biggest kind of talking point at the moment is the agentic commerce era. And I think we are having conversations with our retail customers today about how do they get prepared for this agentic commerce era where you're going to have consumers are going to have personal agents, perhaps doing the shopping for them and choosing which brands that they're going to purchase those goods from. And so as retailers, you've really got to get your inventory data in a really good space, your customer data in terms of understanding everything about the customer behavior and the purchasing and the history, as well as all of your product data as well. So this is going to be really important. Your data is going to be absolutely critical to the next evolution, which is going to be agentic commerce. This is what's going to be next. It's just being considered now by a lot of retailers, and we're talking to them now about where do they need to get to be ready for this.

Dave Vellante

>> And that's end to end, from supply chain all the way through to the customer.

Paul Winsor

>> All the way through.

Dave Vellante

>> I mean, you go to NRF and it's just such a big important supply chain discussion.

Paul Winsor

>> This is where it kicked off at NRF this year. It kicked off. And I remember seeing a quote that if an agent cannot read your data within milliseconds, then your product ceases to exist. This was a comment that was made this year. And if you think about that, that's how important getting your data right, good quality data, being able to kind of read it very quickly is going to be really important.

Gowtham Gundu

>> I totally agree with that. One of the first things I worked on when I joined Fetch was to provide like natural language interface to our data. So I can totally relate to what like Paul just said. We have this huge amount of data, but then if somebody is like Pepsi is coming and typing a question of like, how does my buyout rates change or how is our basket size being impacted in like last few weeks, or things like that, you need to be able to respond very quickly, which means to be able to access this data. So there's a lot of handshakes happening when somebody's asking this question. First you're converting that question to like a SQL query and then running the query, getting the data and again converting that back to a response. So unless your data layers are like really prepared for that agentic moment, I would call. And then the other thing I would add is some of these protocols that like Google and others are pushing in the space of UCP and all.

Dave Vellante

>> A2A and MCP. Yeah.

Gowtham Gundu

>> I think like the dilemma is when is the agentic commerce moment coming, right? That's what everybody is having. But I think Paul hits on the head of the nail by saying like the minimum bar is to prepare your data for the agentic moment. In terms of your strategies and all on top of it can be like bargain, but like the data layers and some of the core technologies should be there.

Rebecca Knight

>> Excellent. Well, Gowtham and Paul, thank you both for coming on the show. A really interesting conversation.

Dave Vellante

>> Thanks guys.

Gowtham Gundu

>> Thank you. Thank you so much.

Paul Winsor

>> Appreciate it. Thank you. Thank you.

Rebecca Knight

>> I'm Rebecca Knight, for Dave Vellante. Stay tuned for more of theCUBE's live coverage of the Snowflake Summit. You're watching theCUBE, the leader in enterprise tech news and analysis.