Samsara Inc. (IOT) Earnings Call Transcript & Summary

All right. Good afternoon, and welcome to Samsara's Investor Day. My name is Mike Chang, and I'm SVP of Finance here at Samsara. And first off, just thank you all for making the journey out here to a very, very hot Las Vegas to join us in person. And it's amazing to see so many familiar faces in the audience. And for those who are joining virtually, it's great to have you on as well. We have an awesome, awesome agenda pack for you today. We have about 2.5 hours full of content, and we're going to talk about how we're bringing AI to the world of physical operations. Before we get it started, there are a few housekeeping items. The key 2 things is, first, we're going to be assessing forward-looking metrics during today's presentation. These should be taken in addition to -- sorry, these statements contain risks and uncertainties, and these are detailed further in SEC filings and our Investor Relations website. Second, we'll be talking about non-GAAP financial metrics, and those should be taken in addition to our GAAP metrics and reconciliation is provided in the appendix for today's presentation. So at Samsara, we're connecting our customers' operations, we're bringing AI to the physical world, and we're making a huge impact for our customers. Super exciting. We've reached $2 billion of AUR, and we have over 13,000 core customers worldwide, and we're just getting started. So super excited to bring on Sanjit, our Co-Founder and CEO today; Johan, our Chief Product Officer, David, our VP of Products, Amit, our Chief Revenue Officer, 2 of our own customers with Primoris and Performance Food Group. And then also Dominic, our CFO. Just like we do with all of our customers, we're trying to build a long-term relationship with each of you. And so we hope that you'll leave today's presentation with a better understanding of this multi-decade journey we're on. So with that, I'd love to bring sanjit on stage and pass it over.

Thanks, Mike, and thank you all, again, for making the trip out. What I'd like to do is start at a very high level with what we're about at Samsara, our mission, and then we'll get into products, and you'll even see some demos for me and Johan in a few minutes. So starting again with the mission. We're a technology company. And our role in this is to help enable all the physical operations companies you see here at our conference to be safer, more efficient, more sustainable in the operations that they run. And if you think about it, this wave of digitization has been mostly focused on corporate office space enterprises, especially when you think about the impact that AI has been having, which is, of course, the big topic of conversation right now. And at this moment, we're seeing a transition happen where AI is going from those white-collar industries, industries like legal, customer support or finance or even software engineering now into the physical world. We can now apply AI to industries like construction, energy utility, waste management and field services. And you can think of this as this transition that's happening from the world of bits to the world of atoms. And when we think about the why now, there's a number of just amazing, incredible tailwinds that are back. First, there's a broad interest in digitization. So generational change that's happening, especially in these frontline industries where people are in their 40s and 50s. They're saying, well, why am I not able to see this on my phone in the field? And then, of course, with AI, there's a why now that's happening right now, which is that AI is becoming more powerful and more transformative every single year. If we look at these trend lines, if we look at the graphs, you can look on the cost side, where if you think about a certain level of intelligence, the AI models have gone from being models that cost $60 per million tokens down to just $3.30. That's incredible. It's an 18x cost improvement in a span of about 1,000 days. Very few technologies adapt that quickly. And if that wasn't enough, the capabilities made possible by AI are increasing constantly. And this is coming from all the different Frontier labs competing. We've seen AI capabilities increase from the point where we were a few years ago, where it was a simple chat bot to now where we can build an agent that performs, work, plans out multistep tasks and executes them on behalf of our customer. And that's an incredible breakthrough, of course, that's going to have an impact in the world of our customer. So at Samsara, we are bringing all of these new technologies like AI into the physical world. And it's important to recognize that this is a multistep, multiphase approach. And again, this is very practical. Most of these industries I talked about earlier, they're largely off-line. They're largely disconnected. So we start with the basics, which we call Phase 1. It's about collecting data. This means deploying hardware out in the field, getting GPS trackers on all kinds of devices, getting cameras out there, so we get visual and situational awareness. And then once we have our hands on data, we can go to Phase 2, which is using AI to go find insights in this data, find the patterns in the data and surface them to our customers. And then at the tip of the pyramid, you'll see Phase 3. This is what agentic automation is all about. Now once you have the data, once you've got AI analyzing, you can start to take action. Most of our customers are in Phase 1 or 2. And most of these industries are actually in Phase 0. They have not yet even deployed all this technology. So we're still very much in the early innings, but these industries are absolutely massive, which brings me to my next point. Because if we go and take a look at the base of the pyramid, the opportunity that we're faced with, it's absolutely enormous because these industries are enormous. They make up about 40% of the world's GDP. You see a bunch of different photos from our customers here. Hopefully, this kind of broadens your thinking about what we mean when we talk about physical operations. We do a lot of work in the world of transportation, logistics and food and beverage distribution. We also have customers that run big fleets of yellow school buses, oil and gas companies that are involved in the energy industry, waste management companies, construction customers and utilities. And if you think about these customers, they have operations that are both asset-heavy and labor-intensive. So we want to help them on both sides of that equation. So if you think more about Phase 1, what you realize is very quickly, we need to collect data about the environment. These are not tokens you're going to find online. They're not build on a website. We need to collect these bits of data from vehicles. So that's why we have products like our Vehicle Gateways. We need data about trailers and construction equipment, which is where our Powered Asset Gateways and Asset Tags come in. And then the visual environment has a tremendous amount of data value as well, which we'll talk about over the course of today. That what our AI cameras do and we want to collect data from people. So we have data from wearables and workflows that come from apps. I mentioned this earlier, but this environment is largely disconnected still. If you zoom out and take a look at all the commercial vehicles on the road here in North America and Eastern Europe and just crack them open, see how many are running connected GPS trackers, you might be surprised to see that only about 1/3, about 34% of them are currently connected from a telematics perspective. If you look at construction equipment, powered equipment on this slide, you see a number that's even smaller. It's only about 13%. And then if you think about unpowered equipment, all the different tools and other assets that are required to perform work in operations, it's very small, sub 1%. So there's a lot of market out there. But deploying this hardware requires assistance. We need to help with change management because we're getting all of this equipment connected, which means bringing it in, putting these devices on there. We're getting apps and hands of frontline workers. Many of these companies have tens of thousands of employees. And this is a new technology that's part of their daily workflow. So we help with that. And this is part of our durable and long-term advantage as well. And then at Samsara, we love data. We collect a lot of it. And at this point, when we look at how many different data points we've been collecting into our cloud over the years, it totals to a number that's over 60 trillion. These are all different kinds of data points. We've got data from all different types of assets, vehicles powered, unpowered equipment. And like I said, data comes in multimodal forms. So we have GPS location data. We have engine health and diagnostic data. We have video data, both externally and in terms of driver behavior. And we have this in terms of a lot of different customer industries, customer geographies and customer sizes. So this time series data that we've been collecting for over a decade now is leading us to find insights in a way that was never before possible, and that's servicing value for our customers. Okay. So now moving up the stack to Phase 2 in our strategy. Once a customer's operations are connected, we can find a lot of value in the data I just said. And for the first time, we're able to correlate all the different bits of data together. I'll give you a few examples. So fuel is a big topic of conversation this year. Price of the pump is up 30%, 40%. And so for many of our customers, this is an enormous impact to their operational budget. They are looking for ways to find savings in their fuel spend. We can find in a number of different locations. So one is simply understanding where are they being inefficient. Maybe they're idling their engines too long or maybe they're operating the wrong equipment, they can get better miles per gallon, if they ran something else. Or could be behavioral. It could be training the drivers or rewarding them for eco driving. And it can even be related to where they're buying fuel. Maybe there's a better provider for them if they just drove another half a mile. These are all bits of data that are in our data set. We can find it for our customers, we can surface it in the form of fuel insights. We can do this for benchmarking. We can do it for fleet utilization. We can tell them if their assets are underutilized. Some of these customers, they have balance sheets that are hundreds of millions, even $1 billion plus just in terms of assets. So these are asset-heavy industries. And this list goes on. You can apply this to maintenance, you can apply it to weather, to routing, navigation and so on. And then if you think about this, it's all about ROI. The world of operations, it's an area where ROI is incredibly important. Like I said, these are asset-heavy, labor-intensive industries. And we've seen a survey -- or sorry, a study done by IDC about a year ago where they went and talked to our customers and found there's an 8x ROI that they're getting from utilizing their equipment more efficiently, maintaining their equipment in a more efficient way, fuel savings, accident cost savings. And now as we start shifting our attention to what's possible with agentic AI, we're able to help with task automation, which helps free up some of that labor budget. And we'll increase the ROI as we'll talk about over the course of the afternoon, which brings me to that top of the pyramid, which is Phase 3. We now have the ability to build AI agents that can really perform tasks on behalf of our customer. This morning, I demoed a maintenance agent, and we'll go a little bit deeper into it today. With the maintenance agent, we can do things like understand in a predictive way what's likely to break based on the data we're seeing in the field. We can set up work orders. We can do things like warranty claims, task work that would take hours and hours on a per vehicle basis. On the safety side, we can automate so much of the task work that has to happen to coach thousands or tens of thousands of drivers at scale. We have worker coaching workflows that are automated. We have ride alongs, as Johan showed this morning, we can do post incident reporting. And then we can even change the safety settings on a driver-by-driver or equipment-by-equipment basis, work that would be too cumbersome to perform manually. And then this also applies to dispatch. There are so many different back-office tasks, if you think about it, load assignment, providing better ETAs to end customers and even things like shift length compliance. These are the practical day-to-day tasks that have to happen to keep an operation money. And then on top of all this, we can layer things on like 2-way communication with the front line. We can do that with our systems that are deployed in the cab. We can offer automations like weather alerts and start-a-day briefings and workflows. All of this was stuff we demoed on the main stage this morning, and hopefully, you've had a chance to talk to some of our customers here about it. All right. So I'm going to go a little bit deeper on one of those boxes, which is maintenance. I'm going to put this in the context of the 3 phases to show you how does we collect the data, find the insights and take action for our customers. So the first step, as I mentioned, is Phase I. It's collecting data from the field. The second step will be around finding insights and things like fault codes and diagnostics, predictive models. And then the third phase of this is around automation, so things like warranty claims and order management. I don't have much hardware here on stage. So let me start with the slide explaining what Phase 1 really is about. I think many of you know, we have a Vehicle Gateway product. This plugs into the diagnostic port of a truck and communicates directly with engine computer. It's an amazing source of data because we can get a lot of information about things like fault codes, but also engine performance. How many miles per gallon the truck is getting? How heavily it's loaded when it's going operating different conditions. We also have other forms of data that come in. In a given year, we see north of 300 million vehicle inspection workflows. These are the walk-around inspections that drivers do. And this is a context that's really valuable from a maintenance perspective because you can see the trending, what's happening with the vehicle over time. This is information you're not going to get in the diagnostic report because it will tell you what the conditions look like on the outside. And it's also very unique data. The OEMs don't have it. The service centers don't have it. And then there's additional data from the maintenance shop. We have work orders. We have other kinds of maintenance inputs that come to our API integrations. We can even integrate with OEMs directly cloud to cloud, and it gives us additional context. So this is Phase 1 of the data collection that goes into maintenance. The other 2 phases I'll show you here on the screen. So if we can flip over to the demo. One sec. Okay. So this is our Samsara dashboard. I'll do some demos here and then Johan and David will be up as well. But this is exactly the same system that our customers use. And I'm going to be live demoing it for you, so you can see how this data comes to life. So for a fleet, you will have hundreds, thousands or tens of thousands of vehicles. Most of our fleet customers, they have been used to operating these vehicles for decades. And the maintenance process tends to be pretty manual. They have a combination of paper checklist, so pen and paper workflows that we're going to go check on the vehicles in the yard, interactions that happen with the driver. So if something is wrong with your truck, you'll come in and let the maintenance technician know, and just expert knowledge that comes from years and years of experience dealing with these trucks. So here, what we've done is try to streamline all of that into a single system where the maintenance tech can come in, in the morning, log in, and they can do this from really anywhere and see at a glance the state of their fleet. And we help them prioritize their work. So you can see here that the AI has figured out, there's a couple of units that need a little bit of extra attention this morning. So I'm going to click in on the first one. And like I mentioned earlier, we're collecting data. That's the Phase 1. We've got data off the diagnostics board. In this case, I picked on a truck that's showing a fault code. And we know the make and model and year of the truck. So we know it's a 2024 Freightliner Cascadia. We also know what kind of engine it has. The Cascadia comes with engines from Cummins. It also comes with engines from Detroit Diesel. This one has a DD13 engine, and it's got a fault 3251, which is related to one of the exhaust pressure sensors. There's a bunch of detail here basically explaining what this fault is about and what some of the recommended actions are. Then I'll scroll down. And you'll see the power of this kind of Phase 2 data that we have and the massive data asset, we've accumulated with the 90 trillion data points. Because we've seen millions of vehicles over time, and we've seen the thousands of different fault code combinations they have in the vehicle performance, and we've seen it over a decade plus, we can do things that tell you how big of a deal a fault like this is. These kinds of complex diesel engines, they have engine faults all the time. Many of these are warnings, informational or maybe emissions related. So they don't result in operational downtime. This one, though, shows up as a moderate fault. That means we're going to want to do something about it. You may be wondering how is it that we know that. And it's because in that data set of 90 trillion data points, we've actually seen 107,000 of this specific type of engine. We know what happens to these engines over time. We know how they perform. And so we've analyzed those. And we've seen that only about 0.25% of them never show this fault. So that's really valuable information. And then like I said, we can piece all that information together. So you can see the fault code chain. You can see here that there's an exhaust pressure sensor issue, and the repair cost is $100 to $800. That's handy for the shop tech to know. And there's a predictive insight there. This says 22.8% of these vehicles progress to the next more severe fault code in 518 miles. So it's very technical, but that now gives me a sense as a shop tech of whether I need to take that vehicle off the road, call the driver up and tell them to turn around, disrupt their whole day or whether we can handle this at the end of the day, maybe at the end of the week, very practical decision-making tool. If I click on this next box here, this is what happens if we don't do anything about it. Default becomes more severe. That's the likely pathway and the repair cost goes up, now it went from a couple of hundred dollars to a couple of thousand dollars. If we don't do anything about that in 175 miles, the repair cost would go up again now to maybe $3,900. For our customers, they spend about 10% of their operating budget on maintenance. That's an enormous expense for them. And so it's helpful if they can get to these tasks early and nip them in the bud, basically save their companies a lot of money in terms of operating budget. Okay. So this is the Phase 2 side of things. This is the insight that AI can give you. Now let me go up and show you how we can start automating a lot of the task work that has to happen to maintenance. So I'm going to click check warranty coverage here. And again, maybe for some context, most of our customers, they procure hundreds or thousands of vehicles every year to keep their fleet refreshed. They do this. And in that process, in their procurement agreements, they have warranty coverage contracts similar to probably the warranties you will have on your cars. But these end up being custom agreements. It might be based on the workload or the environment they operate in. And they have specific coverage areas that they care about. We can upload all those warranty documents into the Samsara cloud, use it as context to power these AI workflows. So what I did is I clicked on check warranty coverage. It fired the engine. It looked at the warranty coverage docs, looked at the fault code and the specific vehicle and said, yes, actually, this is a covered service task. So that means that this customer could get paid back for those repairs because they're still in the warranty period. It understood the document, so it knows that we're within 60 months or 100,000 miles, and it actually tells you what's covered and what's not covered. So you can see some additional notes here. And then you see it's offering to fill out a workflow -- work order. It could do this automatically by the way, before I came in, but I wanted to do an interactive way to show you sort of the steps that are happening. So we can say, -- yes, thanks. Okay. And now what this is going to do is take those documents. There's a lot of additional context in there about what needs to appear in the work order. It can go and query the vehicle, pull in all the information that's needed for work order and then combine it together in the system, very quick process here. So we see unit 1063, it pulled the odometer, it pulled the engine hours. These are all kind of standard fields in a workflow. It's got the description. Let me just go scroll down here a little bit. And then you see the service tasks. So this is what would be required in order to get coverage from the Detroit Diesel warranty. You usually have to provide some documentation, photos, some checklist, things like that. And it's all here. It's all streamlined. And the reason I wanted to go through the step by step is it gives you a sense of how much work the AI is able to automate and to streamline. If we did this the manual old-fashioned way, you'd be cracking service manuals, understanding what that fault code was, you might be calling drivers and asking them to look at stuff on the engine. And then in order to figure out if it's covered by warranty, you're reading through all the warranty docs. And then the last piece here is you have to figure out exactly the checklist. It's probably 2, 3 hours of work. When we talk to our customers, they tell us, they know a number of their repairs are covered under warranty. They probably only claim about 50% of those warranty dollars because this work, this task work isn't happening on a regular basis. Okay. And then one more thing here, I can say, do any other trucks, the same fault? And this is, again, something that our customers will often do. They've been in operations long enough to know that if there's something happening on one truck, it might be happening on other trucks at the same time. Again, the system typically does this in an automated way, but it's helpful to see how we can go through a multistep process. You see it broken down there. This ability to plan is what's made these agents truly possible. So it went -- figured out what the code was that we care about, figured out how to ask or interrogate all the engine computers about it, wrote the code, deployed it, ran it and told us the answer. And is saying, yes, there is another unit that's got this issue. So very practical, very real-world use of AI and agents. But this is the kind of thing that we're excited to get in the hands of our customers because now it's possible for them to automate the hours and hours of task work that would have to happen. And then you multiply it out by the size of their fleet, maybe all our different regional operations, and you can see where the value unlock is coming from. Okay. We can go back to the slides here. So I mentioned earlier, ROI is incredibly important in the world of operations. And we've talked about ROI that we've studied before. ROI that you're going to get from the compliance workflows and the reduced asset downtime. That's about 8x ROI. But then when you layer on the opportunity that comes from this labor automation, the ROI goes up even further. So we can do things like warranty recovery that I just showed you. We can help with things like shop efficiency, scheduling, which order these trucks come in at and whether you combine preventative maintenance with some of these predictive maintenance items. And then there's additional agent ROI opportunities as well. So we think conservatively, this takes ROI from 8x to 10x or even more. And I mentioned that a lot of our customers don't claim all of their warranty dollars. For some of these fleets, that opportunity alone is worth $10 million to $20 million. So there's a significant amount of expense that goes into maintaining large vehicle fleets today. So we're in the early innings of what all this looks like, but we're excited by the initial traction we're seeing. We're excited by the ideas that we're hearing from our customers, about how they want to use all of this stuff. Okay. And then let's talk a little bit about the data. So I went over it fairly quickly. And a few of you may be wondering, well, would it be possible for a customer to go get those insights else? Could you just put an AI on top of their existing data, maybe API it from a different source. And that would be a reasonable approach. And you'd be able to look at your fleet data. So you can see all of your trucks, for example, and all those current fault codes perhaps. And that would give you insights over about 1,000 vehicles. And you might have, call it, 2-dozen different makes and models operating your fleet. But to get the true predictive insight, you need the volume of data that I talked about earlier. We're talking about millions and millions of vehicles, thousands of different fault code combinations. Engine profile data, you can understand how these actually operate over time. And then you can see it over the span of a decade. That is a tremendous volume of data, which is what powers and unlocks the value that I was talking about earlier. So I've talked a lot about what we can do today. But for us, as a company, we know we're in the early innings of this opportunity. We know that this is a world that's going to be digitizing very quickly. And our customers are going to be around. They're going to be relevant because we're always going to have physical operations work to do. There will always be new buildings, new data centers that need to come up out of the ground. The grid modernization product or projects are likely to go on for a very long time, and there's always roadways to maintain. So if we think about industries like construction and imagine what these might look like, call it about a decade or so in the year 2035. We see a picture that look something like what you see on the screen here. This morning, David talked about how we're introducing the new Tracking Label. We'll talk more about it here on stage. We see products like the Tracking Label being used for all the building materials, so you can understand where everything is in terms of coming to the site and when the site is ready to perform the work. We see automation coming in, in the form of autonomy and robots. Think about all the materials that have to be moved on site. The heavy lifting, the dangerous work. We see autonomous diggers, forklifts, cranes but also materials movers, all kinds of handling equipment coming on to the job site. We even see drones and humanoids on the horizon. We're not quite there yet, but it's going to happen. There's going to be so many different ones of these, different makes and models, doing different tasks. And we want to be able to orchestrate that. And we see an opportunity to turbocharge or super power, all the people, to help them orchestrate all this work. We're going to be doing this on different kinds of devices. It could be tablets, could be wearable devices like smart glasses, could be in their headsets where they're commanding all these operations by voice. This is super exciting for us because we know that there's so much opportunity here to have an impact. Okay. To tell you more about how we're going to build for this future, I'd like to turn things over to our Chief Product Officer, Johan Land.

All right. Thank you so much, Sanjit. And [indiscernible] it's great to be here with you. Let me back that one. So it's great to be here with you. And I'm excited to talk to you more about physical operation. It's a really, really important part of society. And what we are doing is that we're building the agentic layer for that. As you just heard from Sanjit how agents are now automating tasks, and I want to go deeper into this and talk about the platform that makes this possible. Now before we go there, I'm just going to spend 1 minute. I think this is the first time we meet. So just about myself briefly. So I've been in tech for the better part of my life. And for the last few roles, it's really been in various forms of operations. At Waymo, I built the commercial PM team. Think of that as the -- we built a service on top of the vehicles. So lost and found, repairs, positioning of vehicles, maintenance, charging, refueling, but also the mobile app acquisition, pricing and all those things you need to build a service on top of it, and took that to first revenue. And my last role was TomTom, which is navigation and routing, both together with OEMs, but also for enterprise. And excites me -- really excited about Samsara is kind of the platform and the ability, like the connection with customers, the deployment of devices, the gathering of data. And that's really the key thing here. It's the data. And we just have an enormous -- and it really means there's enormous amount of data. And effectively, one way of thinking about it is that over the last 10 years, we've pretty much put sensors on everything, millions of equipment and tools and vehicles, and they are sending data back to us. And we're gathering this enormous amount of data that we then can use to build really exciting things. This all sits inside of this platform. And as you can see on this chart, this is the data gathering is just accelerating. And then we use this data to train models. And as we get more and more data, the models get better and better. So that's the basic principle of it. Now Sanjit, Dave and I will cover 3 new products for you now. And the first one is agents, and that's exactly what Sanjit spoke about. The agents sifting through this data, extracting conclusions and taking actions for our customers. And secondly, I want to talk about how we're extending this to operational AI, like these are new capabilities that we're getting from this data and the AI that opens up new problem areas and new markets, things that we haven't done before. And I want to talk a little bit about that. And lastly, David will talk about the Tracking Label that you saw earlier today. And to put this in context, the new products that we have, the new emerging products are already more than 20% of net new ACV. So this is not something that we're exploring. This is already a key growth driver of the company. And in addition, these new products opens us up for new pricing opportunities, in particular, consumption pricing. And what this creates is an ability for us to really go deeper into the problems together with our customers, solve harder problems that creates more value for them, but we can only do them because we can now build as the AI and the agents take actions and create value. That's the other new thing here. Now we really came from a place, focused on driver safety to the left here. And we started there naturally because that's what our customers asked us to do. Like they get into a lot of accidents and making their operations safer, brought great ROI for them. So we followed the customers. And that's always been the logic. We collect all this data, and then we solve the most important problems and pay points for our customers. And that includes things like drowsiness and mobile phones, following distance and what not. And we're still doing that. And by the way, it's a pretty good market. It's a growing market, and there's lots of room to grow in there. But today, I really want to focus on the left-hand side on this chart, like the new horizon that is opening up because we have all this data from the sensors that we've deployed, like new operational AI detections. And here we use these cameras, right, and combine all of this together with improved AI. And when we do that, just like magic, to call it, whole new use cases open up. These are things like vehicle pushback on a tarmac or road defects that need repairing, and detecting road blockages that we then can route around like things to make it more efficient. And today, I'm going to talk specifically about 2 of these. The first one being waste intelligence, garbage trucks and such and the second one being ground intelligence. Starting off with with waste. And by the way, these industries, they are huge, right? They are often very, very labor intensive, which leans itself really well to AI and agents. But Waste Management starting off is a $1.6 trillion global industry. And we've already got $7-figure deals in the pipeline on this one. And I want to show you this. If we pull up the dashboard, I can show what this looks like, let's see, there we go. So this is American Waste, and this is their operation in the Tulsa region. And what you see here is it's the routes that they are driving or have driven. There's nothing really new here. We have been doing routes for a long time. But the new thing here is that we're integrating into these garbage trucks, literally connecting to them. We can see things like when the arm goes up to pick up these garbage cans and dump them, of course, we install cameras in them all around up to 10 cameras, by the way. And what this gives us the ability of doing is to see what's actually happening. Like this. There's a pickup, we get the video. We know that the pickup happened because we're connected to the arm that picks up this. So we know therefore that pickup happened. We collect all the data and observe it, right? That may not sound like a lot. But the things that in these types of operations, there's a lot of stuff happening. So drivers would miss a stop like this, or maybe the customer forgot to put out the trash cans. And the way this is handled today is that the customer will call in and say, "Hey, you didn't pick up my trash, right? But because the customers -- the operators of this garbage truck, they don't know why, so they end up sending out another truck, don't worries, we'll send up a truck to pick it up, right? Very simple, which is additional cost for them. But the thing is this. It's actually a billable event. They didn't pick up there -- if they didn't put out their trash cans for pickup, there should be extra charge for that extra drive. So leaving revenue on the table and they're taking extra cost because they don't have the visibility. Another example of this is overfilled bins. It looks like this. As you can see, it's overloaded. And garbage companies, they have policies for this. It's either 3 or 6 inches, you can overfill. But if you go above that, you got to pay for it. But the thing is that the drivers, they're just getting on with their day, right? They're just trying to get through. A Side loader can have 1 -- upwards of 1,000 pickups in a single day. So the drivers getting through, trying to get through their day. The company doesn't have any visibility, so they don't bill for these ones, right? They just pick it up, take the extra cost for the extra garbage without billing for it. And this enables them to actually charge for it. And by the way, these problems -- another one is -- the list goes on, contamination, customers throw the wrong trash in the wrong bin, they just pick it up and hand it on the backside, by sorting, the list goes on. The thing is that these are problems that they currently cannot solve, right? They're unsolved because they couldn't connect et cetera, and the AI wasn't there, right? The data wasn't there. And by the way, this $1.6 trillion, like, this is what they do. They drive garbage trucks to pick things up. And these things that I just showed you, they're integral to that process, right? It's not a small thing. So this is just 1 example of us taking all this data and the AI and taking the advantage that we have from everything that we already deployed and going deep into a vertical, solving problems and creating value. But another example of this, we're doing this industry by industry. But another example of this is public sector. In public sector, there are many things that are happening, and we're knocking them off one by one. But one example is potholes. And pothole is a $3 billion or so kind of damage that comes from people driving into them. And the cities are actually liable, if someone drives into them and get damages on the vehicle or even worse. So therefore, they are eager to repair them. It's also very visible to the citizens. Mayors are very focused on this. Those of you from New York. This is like Mamdani's first win, if you saw one. So like mayors really -- we got some laughter. So it's actually very visible, and that's also we approach the mayors for it because they want to show wins to their citizens, and it's also visible to citizens. We don't like to drive into potholes. Now the way that cities handle this today is that they get a 311 call, they answer it, someone claims there's a pothole there. They send a vehicle, drive out and check it and make a determination of whether they want to repair it or not. That's quite a lot of work. So I'll show you what this looks like once you connect it to -- once you use our data, our sensors that are already out there. If we flip to my computer here. So this is -- we call this ground intelligence. And the first thing out there is really like the pothole detection, right? But that is material in itself, but that's just the first thing, like we can use this to see like broken guardrails, we can see low-hanging tree lines or power lines. We can see graffiti, encampments, you name it, like with cameras, and we drive 99% of the roads in the U.S. So we can see it all, right? And by the way, also, we've already ingested this data, right? There's no deployment. I'll go -- get back to that. But here -- so here's the city. This is Kalamazoo. Yes, this is Kalamazoo. So in this case -- and this is a customer that's testing this out right now. So as you can see, there are 4,000 pot holes that we detected here from 28,000-or-so observations that we've gathered. So there's a lot of potholes even in a small city. Imagine answering those calls, sending people out to checking all of these. It kind of doesn't scale, but we have visibility on all of them. And then we can actually in this tool, we can filter down. Let's look at some potholes here. And by the way, there are different flavors of potholes as far as I've learned. But we want to just look at not just crackings and whatnot, but the actual potholes. Let's take a look at this one. So here -- so here's an example. So here, we're driving past, and you can see the potholes there. Now we detect this not just from the camera but also from the integration into the vehicle and all the other sensors we have. In this case, we're using GeForce to detect it. And then we run AI on top of that to identify the potholes. Now in this case this was on April 29, but we can actually here see how this has progressed over time. And as you, you saw it was a little bit rainy here. Portholes grow. They grow day by day when under certain conditions. So you're sending a vehicle out to check it is not enough. Like you need true visibility onto this. But we can see here how it's developing over time and how it's progressing. You kind of see it's getting deeper here. This is a month later, but they can grow even faster than that. And then towards here, if we scroll down, here, they've actually repaired it, or they have patched it. It's called they have just put -- eventually, they're going to have to repave this one. The point here is they don't have this visibility. They're sending vehicles out to do that, that cost money. They can't do it enough, so they actually don't know where they are. These are progressing quickly. That's a point for them. The point for us is that we already have the vehicles. The vehicle is already camera. We've already taken the cost to upload and handle all the data, right? So this we can deploy and sell without hardware, immediately deployed. We could build this for every city. [indiscernible] go approach them to just show them where the potholes are. So it's a software product in that sense. So let's see. Yes. So exactly. So that -- so this is just like some examples of the type of operational things that we can do. And we're now working through this, like to see what are the most important problems, what are the verticals that we should go after, et cetera. but it's all enabled by the platform, devices, the sensors that are already out there. Okay, cool. So with that, how about to hand over to David, and he'll talk a little bit about the Tracking Label.

Good afternoon. my name is David, and I'm Vice President of Products and Engineering for our Connected Equipment business. I want to take a step back to sort of Phase 1 and show you what Phase I actually looks like on a map. So this is the Samsara network. We talked a lot about what this data asset looks like, what we can do with it. This is what generates it. And it's kind of easy to not appreciate what you're seeing on the screen. The depth of it are remarkable. Each one of these dots represents a snapshot in time, a 1-hour snapshot of what our network our installed base looks like. These dots, they're everywhere. From this Zoom level, it's a little tricky to see, but they're on the road, of course, they're in residential areas, from buses and city vehicles, but they're also in intermodal yards. They're in airports like ground service equipment, they're everywhere. This network, as it turns out, is one of the key enablers for us to build new products. So you're selling ground intelligence, that's one of the products you build off of it, Asset Tags and another. Today, we're going to introduce a new one, it's the Tracking Label. But what is each one of those dots? What's actually sitting underneath there? Well, each one of those dots have gateway on there. It's got a camera. we can leverage both of those to build new products. So you just saw some of the ways we can use the network of cameras out there, there are millions of cameras that are buses and bulldozers and trucks. They also have gateways, which have Bluetooth. And so we can use that to build products like the Asset Tag that we launched a couple of years ago. So these 2 together can work to create net new experiences that would not have been possible several years ago. We are able to now change the paradigm from building a product for one vehicle for one customer, to a product or a suite of products that we can now work at scale across multiple. So of course, a few years ago, we introduced Asset Tag, Bluetooth-based device, that was really purpose-built for durable equipment, construction equipment, things like generators, concrete saws, fiber splicers, things that go lost and it's all about theft and lost recovery. A couple of months ago, we introduced a new version of the extra small, XS. And with that because our customers said, we love Asset Tag, but we've got even more equipment out there. Every time we shrink the device, we discover this whole new universe of assets. That's things like PPE and fire extinguishers and trucks and portable gas meters and the like. And we've seen that together, these products have really add a lot of value to our customer, getting ROI through. That's a loss recovery but also through managing this inventory at scale. Over time, we've also adopted the networks to be more than just location. So of course, you've seen the cameras, but we can use the Bluetooth for data. You can actually transfer data over this Bluetooth layer. And so we introduced a tank level monitor, and that does location and sensor data. Then we introduced a wearable. It does location and sensor data, even voice over Bluetooth. So this network asset is a tremendous backbone for us to be building on top of, and it's been really valuable for us and frankly, an accelerant to build new products and deliver more value to our customers. But today, we launched our brand-new Tracking Label. So this is the TL11. And the way to think about this is that it's a single use purpose-built Asset Tag for shipments. So think about situations where you have one way of visibility. You can consider things like GPUs need to make it to a data center or pharmaceuticals [indiscernible] to make it to hospital, really high-value, really important things that need to make it from point A to point B, where customers are not interested in the reverse logistics or the return of that asset. Just got to go to point B and then never think about it again. and this is a really acute problem in the industry. Customers actually use Asset Tags to do this, but they're just not purpose built for it. So today, we've introduced this thin flexible disposable tag, last 45 days, once it's activated, it's got about a 9-month shelf life when you turn it on and activate it with our new shipment app. The last 45 days, provide visibility from origin to destination. There's no lithium in this. There's no hazardous materials. As soon as it gets the destination, we automatically stop tracking it because we've geofenced that destination. The network density allows us to discover it's been delivered. We stop tracking it, and it can be disposed off safely. We've made it realistic and economical for our customers by putting it on a consumption-based model. So this is not a subscription based, it's a consumption-based model, which is -- makes it palatable for our customers. One of the interesting aspects of the Tracking Label is it's opening up a new market for us. So we primarily or historically have thought about customers with fleets of vehicles and assets. There's a whole other aspect to this world of physical operations with organizations that don't necessarily have heavy infrastructure in trucks and trailers and so on. They are Shippers, they're actual manufacturers of the goods that this whole ecosystem sort of exists and move their goods around from point A to point B. There are folks in retail like Nike, of course, and electronics, NVIDIA, there are folks in automotive, like Bosch, and these organizations have a real problem on their hands. They've got situations of cargo thefts. We all hear about that. There's also operational visibility and downtime and decision-making that we have to make. And so these are really the motivators for us to think about how to actually go about building a product like this and what caused us to go build Tracking Labels in the shipment center that I'm going to show you soon. We are also using this ourselves. So we have an aspect of our business that is logistics, and that's shipping our hardware devices to our customers. So we've actually been eating our own dog food on us for the past several months and partnering with other customers as they deploy this and getting into beta. And we've seen tremendous impact of us. We're able to make better decisions as a result of knowing where things are. We shift the game from being reactive to proactive. Instead of finding out a shipment is not going to make it on time, we know it -- [indiscernible] the fact, we know about it right now, we can make a decision, maybe ship a new thing, maybe call customer, let them know. But either way, we're not now encumbered by barcode scans that are 12, 18 hours, maybe a week late in certain cases. So how do we do this? Well, obviously, it starts with the network that I just showed you. It starts with the visibility that comes from that Bluetooth network. So we're not now relying on barcode scans at a cross dock or loading dock. We know exactly where it is all the time, and that makes the difference. Then on top of that, we're able to layer our operational visibility. So that's the millions of cameras that are out there that are able to give us context, well, okay, I see where it is on a map, but where is it really? What's going on here? Why is that shipment stopped? And then we're really fortunate to be building this in the age of intelligence, as Sanjit mentioned. So we can do this with AI. Now we can make better decisions. We can take all that data that these shipments are generating, sift through it or allow the AI to sift through it on our behalf and make decisions. For example, how many warehouses do I need? Which carriers do I want to leverage out of which warehouse? Am I serving my customers appropriately? These kinds of decisions, which tremendous -- historically have been tremendously difficult to make. And so now we're able to do it in a matter of seconds. So I'm going to flip over the laptop and show you kind of how this works. So this is a shipment that came here. We shipped it of our Louisville, Kentucky warehouse and it came to Las Vegas. You see the origin in Louisville, it's to destination is Vegas. That's great. It was delivered. So that's good to know. We're aware of that. We see the contents. It's got 1,000 of these labels in it. We've actually integrated this with our ERP in our warehouse management system. So one of the key investment areas in making this product is how do you make it frictionless and must be frictionless and our operations team really worked with us on this. So this is great. We can see where it is, where it started, where it ended up. But let's look at how it actually gets here. Let's just pause for a second. So that arrow, that is the Tracking Label. Now some context, we intentionally didn't ship this with the same Samsara customer. We want to see what the worst case would look like. So there's no dedicated network infrastructure on here. There's no Vehicle Gateways attached to this, providing this connectivity. This is that label on pallet, inside of a box truck, made out of metal, shooting Bluetooth out of the side, getting picked up by that network that I showed you. And that is why that network is so critical and so important. This is a product that anybody can talk about, but it's much harder to walk the walk. And our network allows us to walk the walk. Now these other gold dots here, these other little pink dots. That's the carrier update. So that is what traditional scanning looks like without a label. You see a scan at a barcode, or barcode scan at a cross doc or loading dock, you see arrived at facility, left facility. That's the visibility that our customers have today. You can appreciate why seeing real time might make a difference. So we'll finish this journey out. We'll see if it came all the way here and delivered. So what are we able to do with this? How do we actually think about this? Well, as I mentioned, we've been using this ourselves. So this is Samsara on Samsara. These are all the shipments going out to our customers right now. And we're not clicking on individual shipments here. We're not going to go individually inspect each one of these. We're going to manage by exception. And so we can do things like look at delayed shipments, and we can click on this one here. And we can see that this was supposed to be delivered a couple of days ago, and it's not yet been delivered. So okay. what do we do about this? Well, all of the AI that we've invested in, all the automation we invested in, this feeds right into it. So our ops team can trigger an agent or have an agent configured, automatically e-mail our customers. It can slack them and let them know, hey, you might need to make a new order for this customer because this is going to arriving late. Now there's another thing that those cameras out there can help us with. Sometimes shipments get stalled. We don't really know why. We don't really know where. And so as I mentioned, we can actually put eyeballs on these shipments. These photographs are very similar to the photos that Johan showed you and you've seen with StreetSense and Weather Intelligence. These are taken from other vehicles in the vicinity, in the vicinity of both time and space to where the shipment is. And earlier today, we investigated a shipment that was stalled, it was stuck. And we were able to see it was stuck in a gas station, slept there overnight. For us going to the gas station is a 15-minute affair. For a driver oftentimes they're parked overnight at a gas station. And then you wonder why this cargo theft. Well, if we see something is stalled, now we can say, "hey, you're not in a secure facility, move that trailer, don't stay at the cargo -- at the rest stop overnight with this precious cargo inside. So this is how Bluetooth and cameras can all start coming together with shipment visibility and provide better and richer context. Now One of the other use cases for us has been actually making these operational decisions. So how do I actually now they've got the base layer of visibility, the camera layer of operational visibility. How do I leverage AI to make better decisions? Well, one of the questions that we like to answer and think about is how are we performing as an organization. Our customers are asking the same question. So you can imagine that on a quarterly or by quarterly basis, our operations team likes to think about how we do [indiscernible] a report card. We want to see if we're doing all right, we've appropriately balanced our inventory across these warehouses and if we're serving our customers appropriately. Now this type of analysis to ask how my warehouse are performing is really, really burdensome for people. You have to go get CSVs, you have to get G Sheets. You have to get third-party information. And even in the age of Claude and chatGPT, managing this is a pain because you have to go find the data. It's really difficult. It takes several days to put this kind of report together. Well, in a matter of seconds, the AI is able to put this together for us and it's able to benchmark how these warehouses are doing. And we can look at things like, yes, we are doing way more volume out of Kentucky. And that makes a lot of sense because most of our customers -- more densely populated areas on the East Coast. But we may want to reload or rebalance a load. So we have to go deeper and ask, which carriers are best on time, do we want to use different ones for different routes so we have the best possible service for our customers. So these are just a couple of ways we can go back to the slides that we're leveraging Tracking Labels internally and our customers are leveraging them to get ahead of issues. So in closing, I think the really interesting and important part here is we're seeing an accelerate in the way we're able to develop these products. The products we ship, go out to customers to generate data, which enables us to brand-new products, just like with Asset Tag, just like the tank level monitoring and now, of course, just like Tracking Labels. So we're really excited about this. We're excited about the ROI that we think we're going to see out of this for our customers and the new customer expansion into shippers. And then I'm going to hand it back over to Amit.

Thanks, David. Good afternoon. My name is Amit Vyas and I'm the Chief Revenue Officer at Samsara. And today, I'm excited to talk to you about our go-to-market momentum. As you see here, we are seeing a lot of success selling into various industries. But what I find most impressive is that we are selling to the market leaders in these various industries, companies like Alaska Airlines, Hertz, even Ecolab, all have complex operations at scale and rely on Samsara to solve their problems. So what problems are we solving. When I'm out in the field, these are the 4 things I hear most often. First, rising cost, fuel cost, as Sanjit mentioned, insurance premiums. And these are essential for our customers. They cannot operate without them. So any chance they get to reduce these costs, they're going to take it. Next is safety, not just for their employees, their assets and their equipment but also the general public. Our customers run critical operations and cannot afford any unexpected downtime. That will cost them millions of dollars in certain cases. So they're always looking for ways for preventive maintenance. All 3 of these things converge to basically an emerging need for digitization. For so long, these customers and physical operations have been ignored by tech and they are now with AI, as Sanjit mentioned, are really reaching out to say how do we go lean in and digitize our operations. And we're able to show ROI relatively quickly. For example, SD saved $3 million in fuel costs, USC had a 98% reduction in insurance claims and Maxim Crane, saved $13 million on maintenance costs. So one question that I always get from investors is you are solving customer problems, you're showing ROI very quickly when they deploy, why do they choose to do a phased rollout. So to explain this, I'm going to walk you through the customer buying journey at Samsara. So we'll start with the rep qualifying the opportunity. They will do a demo of our dashboard, and then they'll pitch a free trial. Now the free trial is our most powerful sales tool. It allows the customer to try Samsara within their own organization and compare it to the competition. From there, they'll do an initial deployment. Now I used to sell IT, routing, switching. And when you're selling IT, you're just replacing the incumbent in a networking closet and you're able to sell the complete solution. We sell to physical operations. So you have bulldozers out in the field. You have cranes being used. You have vehicles transporting goods. And our customers choose not to stop their operations and install, they rather do a phased approach. To illustrate this point, I'd like to talk about a well-known home improvement company that bought some -- sorry, back in 2019, they bought -- they started with 2 products, safety and telematics. And as they started with a small deployment and then the following year, they bought safety and telematics, but they also bought a third product, Asset Gateways. Now over the next few years, they continue to buy more and then they bought a fourth product, Asset Tags. And then just this year, they bought a fifth product, Connected Workflows. Now when you think about this, my sales rep sold the initial deal. They're also getting to know the customer better, establishing the relationship, but understanding their complex operations better as well. So they can sell and provide solutions as they identify them over -- as they roll out. The other thing I want to point out is this shows how our enterprise customers buy. Look at the confidence. They come back. They do an initial purchase. It's small and then they continue rolling out, but they're adding other products, as they continue their Samsara journey. I'm going to switch gears and now talk about 3 initiatives that I launched at the beginning of this year. We are seeing success in the enterprise. We are continuing to move upmarket and getting into global accounts. So now we have a global sales team that is dedicated in calling into these accounts. These are our most tenured reps. We also are seeing a lot of success with emerging products. So Johan mentioned, we're rolling out quite a few new products. I don't want to overwhelm the sales team, the main sales team. So we launched product specialists at the beginning of the year. This is a dedicated team with their own quota that sells specific products, co-sells with the main sales team. And this is working out really well. In fact, Q1, our emerging products was over 20% of net new ACV. And lastly, I'm always looking for ways to increase efficiency of my sales reps. I want them to be able to spend time selling and not doing admin work. So we've launched a lot of homegrown AI tools so that it can automate things like follow-up e-mails, et cetera. I've been at Samsara for almost 10 years. And I have a habit that after every end of the quarter, I call a few select customers and ask them, how did my rep do? How is your buying experience? And tell me why you bought Samsara. And for the last 10 years, the first 5 have been consistent with the newer 1 being actionable AI insights. At the end of the day, I would say Samsara customers buy Samsara and it's working because we have the right people. I know exactly which profile rep to hire for each of my segments. We have the right product. Our product road map is built by customers with their feedback, solving their problems and it's the right time. As I mentioned, there's an emerging trend to really lean in to digitize their operations. This is why Samsara is winning in the market. Thank you. And now we have a quick break for 10 minutes. [Break]

All right. Let's get going here. So hey, [ Theo ] and Eric, like thank you so much for being here, like really appreciate it. Sitting down with customers, it's like some of the best things that I do. It's truly an honor to speak to true -- 2 true leaders in the industry. So thank you so much for that. Now you need quite complex operations as such, like I would love for if you could start off with us introducing yourself and maybe a little bit about your companies, maybe starting with you, Eric.

Sure. My name is Eric Amlee, Senior Vice President of Fleet for Primoris Services Corporation. We're a utility scale contractor. We do EPC projects from renewables, solar, power generation, heavy civil. We're basically scaled all over North America, roughly 24,000 pieces of equipment, 8,000 of those are rolling stock units with dual-facing cameras. And my role within the organization is a cradle-to-grave fleet management. So we're a centralized fleet model. Anything from business cases to repair and maintenance, procurement, disposition, all that rolls under my group.

My name is Tom Olitsky, referred to as TO here.

Sorry, it's just that everyone in the industry that knows him, like, that's TO.

TO. That's fine. That's just my initials, but that's -- I've been going by for about 10 years. With Performance Food Group out of Richmond, Virginia, we are a food service distributor, Fortune 100 company, serving over 300,000 locations, including chain, independent restaurants, schools, health care, hospitals. We do movie theaters. We do gas stations, convenience type of stores. So it's pretty much what we like to call the away -- North American food-away-from-home market, right? Anything in the North American group that does food-away-from-home. We employ about 43,000 associates, over 150 locations with an annual revenue of over $60 billion. In my role as a VP of Safety, I'm responsible for the health and welfare of all of those associates, including protecting the motoring public from what our drivers are in, and those communities that we serve. On the transportation side, we have about 12,000 delivery associates and as many vehicles.

That's some incredible scale here, like you're really powering society here. It would be interesting to hear, like how are you using Samsara day-to-day? Maybe starting with you, Eric.

Yes. So we started using Samsara roughly in 2018. We've grown exponentially since then through acquisition. We currently have 8,000 dual-facing cameras, as I've stated before, so protecting our drivers and the public. We've also used a lot of -- we use yellow iron. So it's equipment that's powered through multiple OEMs. We use trailer tracking devices, asset trackers. And really, the challenge we run into every day is that we use dozens of OEMs and where do we put all that data? How do we aggregate that? Since that's where our partnership has really, really blossomed and developed.

Yes. TO, how are you using it?

We are a Samsara customer for the past 5 years. We did come from a competitor that we've had 10 years prior. So we've been in the camera safety business for a while. But I think that our journey with Samsara has taken us to a totally different level. Safety is paramount for our drivers and for the public that we serve. And we took the platform from going from reactive to proactive, with all of the data and information that you're able to provide on the safety side. We also are a food company. So we have refrigerated and frozen foods. So we use environmental door monitors and temperature sensors because food safety is critical, and we have to be compliant with the FDA. So that helps us also in a different use case. Sometimes we run out of space in our buildings, and we use our trailers for storage, and they become an extension of our buildings. And Samsara helps us be able to monitor those because before, it was all manual. We just had, hey, some go out and check the refrigeration units every 4 hours, 24 hours a day for 6 months because we have 25,000 turkeys on them. Now, we're able to be able to do that remotely, and that helps us immensely.

And I know that in our industries, like outcomes is what truly matters. So can you maybe share a little bit about how does the ROI work? And what kind of ROI have you seen from using Samsara? Let's start with you, Eric.

Yes, sure. So we -- for years, we used just the GPS telematics and we track drivers through driver scorecards. And then we saw some pretty good impact. But really, the change was when we adopted dual-facing cameras. We ran through a pilot program, and we're a pretty robust company. So we have a lot of union, non-union companies. So working through the unions, getting agreements and proving the value to the employee, not using it as an oversight and like a big brother scenario. And really, we've seen a huge impact. I mean, roughly, we're down 66% in total events, 40% in severe speeding and 42% down in crashes overall with 8,000 units. So that's equating to today roughly $5 million a year in savings from crashes and claims. And really, it's unbelievable to have met the amount. And the adoption rate of that through our organization has been phenomenal. Everyone, at first, is definitely afraid of it. They're like, oh, everyone is going to quit and leave. We didn't have one person quit and leave. Honestly, we've worked through it. Everyone complains they have some disorder that can't be recorded, and we work through it and explain to them like, it's actually a value to yourself, right? We don't record folks. So everyone thinks that it's being recorded, and we proved everyone. So we actually developed a one pager and showed them like, hey, here's our change management. Here's what we're trying to do, and here's also what we accomplished. And we shared some success stories, really, of allowing how we've been able to get drivers out of infractions because -- to prove that they weren't at fault from where an officer thought they were at fault for it. So when you share those success stories, the adoption rate becomes -- it's just easy. It's naturally you're done, right? And really, the indirect of it is we've reduced our idling by 30%. So with today's fuel cost, it's millions of dollars in savings. So it's been a very, very successful program for us.

TO, how are you experiencing ROI?

Well, as a safety leader, ensuring our people make it home every night is most important to me, obviously. But in the past 10 years, we've seen -- we're battling insurance inflation, catastrophic loss and claims, nuclear verdicts, all of that is a headwind for us. And it just continues to rise at an alarming pace. I would argue with anybody that the system pays for itself in just pure exonerations. From a financial group, I don't know if you count that as an ROI, right, something that didn't happen. But I can tell you, I could look at 10 videos in a year that would cost 3x as much as we're paying you guys just to have the system. So exonerations is important. But other than that, we forecasted an increase in insurance costs this year for varied reasons, a lot of inflation of $20 million, we expect it to increase $20 million. Now, our fiscal year ends this week. We're a July to June fiscal. And as we finished the year, we realized that we only needed to spend $10 million of that. So that's a weird type of win, but planning for a $20 million loss, and only $10 million is very successful to us. But that doesn't happen easily. It's not luck, right? So it's our dash cam footage. Over the past 3 years, we've seen a 26% decrease in total event rates. 60% in collision risk, 70% traffic signals and signs all through the tech, 23% decrease in harsh rate reduction. One of my favorite, 90% decrease in speeding over 10 miles an hour, 90%. That's just grassroots coaching, which is great. 33% in seatbelt. But the most important part is repeat behaviors, right? And we love that you give us a repeat behavior percentage. And that has improved by 67% in just the last 12 months. So bottom line is we're moving in the right direction. All the data that you give us helps us be better stewards of our drivers and helps us coach.

Yes. Those are some quite impressive numbers. This takes partnership. So I would love for you to explain -- share a little bit, what is it like to partner with Samsara? And maybe also, what is it to be a part of the broader community? Maybe starting with you this time, TO.

Okay. Yes, I cannot -- I get excited about the partnership. It is so fun to be a part of this -- I feel like I'm part of the company. And I've watched you grow over the past 5 years so greatly. But we came -- again, we came from another vendor for 10 years, and we were -- at that point, we were unheard, and we're a big company. But it became to the point where if you have a suggestion, it was, well, it's going to have to benefit everybody. I'm like, okay, well, it probably will. But then it would just fall flat and nothing would happen. You guys listen and you react. And it is a community. I could call up anyone, even not in my industry, and talk about -- because everyone's -- road safety is the same for everybody. I don't care what you're delivering or installing or doing things. So it is a community, but we have chances to communicate it beyond. We're on the Customer Advisory Board, which is fantastic, peer-to-peer meetings. Some of my favorite is I sat down twice already with two potential customers in the morning and this afternoon just to talk about -- because they're on the fence. What do I do? Should I go here? Should I do it? Should I not? How hard is it going to be? And I love sitting down with those customers and telling them, yes, you have to. It's -- you have to take care of the public. And some people will go into this thinking, I don't want all that data because I'm afraid what I might see. That's not the reason to not buy the system, right? Don't say that out loud again to me. You have to face what's out there and you have to coach and make it better, right? Everybody's families are on the road. It's important for everybody.

Yes. So maybe -- by the way, thank you so much for the partnership. Maybe on the topic of new products, Eric, for you, like looking at our road map, what are some of the newer products that you are most excited about?

Yes. I think the 360 camera for us is going to be something that we'll be able to use just to get that bird's eye view. Some of the manufacturers have something of that, but there's a lot of OEM partners that are way, way behind. So we need something that we can put in place today, the avoidance of person detection, object detection, just the amount of -- just in injuries and incidents alone, not just the cost of the unit getting damaged, but taking that unit out of service and disruption of operations and sometimes, they are very hard to measure, and that's really where it affects your margins and your profit margins. And it's all about utilization and getting those going. And really, we're really after connected assets and using the maintenance program. So we're really ramping up all of our maintenance schedules, preventative maintenance and everything. So really, pushing that through one system. We have -- there's just so many we use today. Having to get fault codes from different systems and different manufacturers and having one place to go and being able to act on that in real time, it just -- that's really, really where we need to be and where we're going to ramp up for sure.

Yes. That's great. What are you excited, TO, about the road map?

I'm like a kid at Christmas. I mean there's just so much good stuff out there. I would agree with Eric that I think multicam is a game changer, having 360-degree visibility around truck and you're now going to have bird's eye [ view ], it amazes me that you can look down and around because we love to pull -- our businesses are in parking lots, right? That's where customers are. So we're pulling in and we're knocking down wires and hitting overhangs and gas stations and tree branches. We're always looking ahead when no one is looking up, but you get that over, the view around the whole truck really makes you realize what you're about to get into [indiscernible]. So I love multicam and the birds eye. AI-powered ride-along is fantastic. We require our supervisors to go out at least once a year and do a ride with every single driver. Now, this kind of -- it will take us to a different level of, do we really have to go out with all 11,000 drivers? Or can we just go out with the 25% of the drivers that really need a little more TLC? Because everyone else could be monitored through the system, and we could look at that. And then if you could go out with a ride-along if that trend comes back as a little bit rough. So powered, that's a great one. Two-way conversations is fantastic. We don't want phones in their hands. Phones are an addiction. They love to pick up the phone. I've asked some of our coaches, when you need to get hold of the driver, what do you do? Well, I call them on the phone. Exactly what we don't want to do, right? We tell them don't use the phone and you're calling them on the phone. So this now gives us an avenue to be able to talk to them or be able to talk to them if they have fatigue issues. There's just a lot of tech that's coming out that's really exciting.

Eric, you have started testing autonomous equipment a little bit. So I would love to hear like, how is that working? And also, how do you see Samsara fit into that?

Yes. It's been a challenge. We use a lot of it today based on excavators. So mounted on a CAT excavator, we can overlay a trench and say, dig this trench and it has obstacle avoidance on it. It's been a little bit of a challenge because it's not necessarily faster than an operator is today. The hope is to be able to scale that one operator can run five or six machines. A little bit of difficulties just with getting into the weeds, material sloughs off behind it. The machine doesn't necessarily know that. So I have to go back and clean an edge or if it hits a rock formation and what it has to do to get over that rock. So that's where the operator is a little bit more efficient, but the technology is definitely rapidly approaching. Where it's very successful today is in our mining facilities. So it's going to load a truck, and then it's going to dump in the same location every time. So like a point-to-point, and that's a fixed location. Some of the industries will get challenged on is in utility structures -- or utility infrastructure, you don't really know what you're getting into until you get on site. So power poles, residential areas, if you have to get traffic control, delineation, things like that, that's where it's kind of a slow progression. But that's definitely the progression of that. And really, where I see Samsara is fitting into that is all this is being engineered by multiple OEMs, multiple third-party companies. And again, we're trying to not get multiple platforms and more platforms. We're trying to condense all that. And really, where I see Samsara is really fitting into that as being the single pane of glass for everything we're doing for maintenance, GPS utilization, AI platform and autonomous platform all being housed in one location.

Yes, that makes sense. Have you started testing any of this, TO, in autonomous vehicles?

It's an interesting conversation for our industry. And I've been in food service distribution for over 40 years. And we're doing exactly today what we did 40 years ago. We are going from point A to point B with 800 cases on a truck, that we have to touch every single case to go to 20 different customers, up and down a ramp. So it's a very -- it's a human operation. And so we have not tested it yet because if we do -- now I fully, especially from a safety perspective, believe in the technology. It's safer. If everything was autonomous, we wouldn't have accidents because accidents are caused by humans. It's not because of a failure of a vehicle. So we need someone to offload. So I would have a driver sitting next to an autonomous vehicle and then they get to every stop and offload. So we're not saving labor. And the cost is probably 50% more of a regular vehicle. And I'm worried about the infrastructure. So we're cautiously optimistic about it moving quick. But again, we'll always have the labor side in front of us that we're going to face.

Yes. And how would you see Samsara fitting into that when it eventually comes?

You have to build those robots, that's for sure, that are going to offload those trucks. Okay. I don't know what else to do. But no, yes. I know you will find a way. We will work together to figure it out.

It's good. And on the subject of AI, TO, so you've been early adopters of our agents. So I'm curious to see, what are you seeing there?

Drowsy detection. So that's been one of my favorites and one of the things that keeps me up at night. When you see videos of people starting to fall asleep at the wheel going 60 miles an hour, it scares you to death. So now we get alerts, right? We get alerts back at the shop, we get alerts in the truck. Depending on the type of event, if it's egregious enough. We might pull a driver off the road. When we first started getting the drowsy events from Samsara, what we did is we probably had six drivers we pulled off the road and send them for a fitness test to see if they have obstructive sleep apnea, which tons of people do. And all six came back with obstructive sleep apnea. And all six went on a machine to help them breathe better, and none of them had a distracted event in drowsy again since that. So it's really about the health of the driver, which is fantastic. So I think it's very powerful.

That's cool. And Eric, switching gears here a little bit. I think you've been testing out the asset tags, specifically when it comes to theft and loss. What have you seen there? And what are some stories? Do you think they recovered or...

Yes, definitely. So in the construction industry, a lot of our stuff is parked at a hotel at night, someone's house. It could be a job site that's not secured. It's on the side of a highway, city street. So we do, unfortunately, see a lot of theft, specifically in small tools, whether it's a jackhammer, a saw, and we've been able to put small asset trackers on all of those devices. And in addition to our mini excavators, backhoes, things like that, they are also getting stolen as a redundant device. So we put an earth magnet on that AT11, just put on the machine, whether it's a rental or not, and it's allowed us to recover multiple devices. So for us, recovering one mini excavator or backhoe pays for almost the entire program as a whole to be able to afford us the opportunity to invest in more AT11s, AT12, AT13. So just last week, we had three separate incidents where trucks were broken into. They stole some partner saws and small hand tools, and we're able to recover all of them. And two of the thefts that were separate were actually recovered at the same site. So it tells you that most of these aren't just random people in the neighborhoods doing things. These are professional rings that are going around. Not necessarily professionals, I wouldn't call them pros by any means. But that's definitely a group of people. And we're rapidly expanding that, especially with the new devices we're coming out and testing. And we've been an early adopter and an early tester of that, and it's -- we bear a lot of fruit with it.

That's great. Hey, I want to thank you so much for a great conversation. And more importantly, thank you so much for your partnership. Thank you so much.

Thank you.

Yes. Thank you. Appreciate it.

Right. I think that's over to you, Dom.

Great. Thank you, and welcome to Investor Day. I'm Dominic Phillips, Samsara's CFO. I have about 20 minutes of content talking about durable and profitable growth, and then we'll bring everyone back up on stage, and we'll do a Q&A for about 30 minutes before wrapping for the afternoon. So before I get into the reasons why we think our growth can be durable and profitable going forward, I first just want to level set with where things stand today. So we had our Q1 earnings call a few weeks ago. And on it, we announced we're doing roughly $2 billion of ARR growing 30% year-over-year. And you can see on this chart that as we've scaled our ARR over the last couple of years, we've really been able to stabilize and level off our overall ARR growth rate at roughly 30%. And the reason we've been able to maintain this high level of growth as we've scaled is because over the last several quarters, we've been accelerating our net new ARR growth. So this is a look at our net new ARR growth by quarter compared to the same period in the prior year. And you can see that for the last three quarters, net new ARR growth has accelerated compared to last year. And even over on the right-hand side, for the full fiscal year, FY '25 net new ARR growth was 16%, and then we accelerated it to 21% in FY '26. And the big reasons we've been able to drive this growth, first, emerging products. For the last two quarters, emerging product net new ACV has contributed more than 20% of the overall net new ACV. The second is large customers continue to be our fastest growing. So 62% of our ARR comes from our 100,000-plus customers. That's up from 58% a year ago. And then within that, our $1 million-plus ARR customers now contribute roughly 1/4 of our overall ARR. And the ARR from both of those customer cohorts has now accelerated sequentially for several quarters in a row. And the last reason is international momentum. So in Q1, 18% of our net new ACV came from non-U.S. geographies. That was tied for a quarterly record. So all the emerging products, large customers, international, driving accelerating net new ARR growth, which has allowed us to stabilize our overall ARR growth. And in addition to sustaining high growth as we've scaled the business, we've also delivered a lot of operating leverage. So on the left-hand side, over the last 2 years, free cash flow margins have improved by 15 percentage points. And on the right-hand side, we've now achieved GAAP EPS. We've been GAAP EPS positive, profitable for the last three quarters and cumulatively, over the last four quarters, which now makes us eligible for some of the larger stock indices like the S&P 500. And this combination of large scale and fast growth and profitability really put us in rarefied air. There are roughly 300 U.S.-listed software companies, of which roughly 60 are doing $2 billion or more of ARR. And then of those roughly 60, there's only three that are still growing north of 30% and are GAAP profitable, Samsara, Palantir and Datadog. Okay. So that's a quick snapshot of where things stand today and what we've accomplished over the last couple of years. We think that this growth rate can continue to be durable for several key reasons. The first is that we're selling into really large end markets that are growing quickly. The second is that our products address a really large TAM, and we still have a lot of white space opportunity in our core products and in our core geography. The third is that we are really purpose-built for large enterprise customers with complex physical operations, and that continues to be our fastest-growing customer cohort. Fourth, multiproduct adoption is increasing as we add more emerging products into our portfolio. And then lastly, we have this really nice balance of net new ACV contribution coming from landing new customers as well as driving expansions into our existing customers, and we think that we can maintain that to drive durable growth. So double-clicking on each of those points in a little bit more detail. Again, the first is that we're selling into the world of physical operations, which is massive. These end markets make up more than 40% of global GDP, and they span more than a dozen different end markets. And because our value proposition to customers is universal, we're helping them improve their safety, their efficiency and their sustainability, our ARR across the physical operations end markets is very diversified. On the right-hand side, this is a look at our ARR broken down by each of the end markets that we're selling into. And you can see there's not one industry that makes up more than 20% of our overall ARR. You have construction at 19%, transportation at 17%, wholesale and retail at 16% and field services at 13%. And beyond those top 4, there's another half dozen end markets that represent between 5% to 10% of our overall ARR. Not only are these end markets very large, but they're also growing very quickly. Over the last 10 years, U.S. GDP has grown 31%. And each of the physical operations end markets is growing faster than that, both individually, which is represented by each of the dotted lines on this chart, as well as in aggregate, which is represented by the black line. And you can see that overall, physical operations is growing at a rate that is more than 2x U.S. GDP over the last 10 years. And the companies in these industries require a lot of heavy assets and frontline labor to get their work done, which means they have very large operations budgets, which are not -- which are less discretionary. We took a look at our top 10 public customers and found that they spend roughly 80% of their revenue on their operations budget. That includes all of their frontline labor, their assets, their vehicles, their equipment, all of their maintenance, their fuel, their accident costs, their insurance premiums, all of those costs are encompassed in the operations budget. And then they spend only 11% of their revenue on things like their IT budget and other SG&A-related activities. Because the operations budget is so large, it means that we have a big opportunity to have a lot of customer impact and ultimately drive a lot of ROI for our customers, which you just heard from the customer panel. And because these industries are so asset and labor-intensive, they're also more AI resilient. This is a spider chart that Anthropic put out a few months ago. And around the chart, you see a bunch of different industries. And the ones highlighted in green are the physical operations end market. So you see transportation and installation and repair, construction, agriculture, those are all the industries that we're selling into. And the way this chart works is where they have highlighted in light blue, those are the industries that they think are the most likely to be disrupted by AI. And you see a lot of light blue shading tied to industries like business and finance and computer and math and legal. And you don't see very much, if any, light blue shading on the physical operations end markets in green. So again, world of physical operations, it's massive. These industries are growing very quickly. Within these industries, we're selling into the operations budget, which is very large and less discretionary, and these end markets are more AI resilient. Second durable growth point is that our products address a really large TAM. Top-down external research estimates that our products address a market opportunity that's $175 billion. Within that, our core products address a $45 billion market. That's our telematics, our AI dash cameras and our powered asset gateways. And beyond our core products, we have a $130 billion TAM for all of our emerging products that we've announced over the last 2 to 2.5 years, including everything that you've heard about today. Not only do our products address a very large TAM, but our penetration in our core products and in our core geography are really still in the early innings. On the left side of this chart is the North America telematics market that has been around for a few decades. Of the more than 35 million commercial vehicles in North America, only about 50% of them are connected today or using some sort of telematics solution. And that's fragmented across more than 35 different vendors, which means that the other 50% of commercial vehicles are still not yet connected. They have not gone through the Phase 1 transition that Sanjit talked about earlier. On the right-hand side, the North America AI dash camera market is only 15% penetrated across 10 vendors. And the other 85% of commercial vehicles are still not using an AI dash camera today. So while things like our emerging products and international, those are going to be really important drivers of durable growth, we still have a lot of growth opportunity in just our core products and in our core geography. Next durable growth point is that large customers continue to be our fastest growing. We now have 190 customers that pay more than $1 million of ARR. Those customers represent roughly 1/4 of our overall ARR or just under $500 million. And the ARR from the $1 million-plus customers has now accelerated sequentially for four consecutive quarters. Not only are our largest customers our fastest growing, but our enterprise customers continue to get larger over time. On the left side here, today, our 10th largest customer is paying $6.6 million of ARR. That's up 4.4x from 5 years ago when our 10th largest customer was paying $1.5 million of ARR. And you can see it's up almost 1.5x in just the last year. You can see that really large step-up from FY '25 to FY '26. In the middle, our 25th largest customer today is paying $4.2 million of ARR. That's up 5.3x over the last 5 years. And on the right-hand side, our 100th largest customer today is paying $1.5 million of ARR, which is the same size that our 10th largest customer was just 5 years ago. The next durable growth point is that multiproduct adoption is increasing as we're adding more emerging products into our portfolio. On the left, 96% of our large customers and 92% of our core customers subscribe to two or more products. And that has increased slightly over the last couple of years, but we've effectively reached full saturation. Most of our customers at this point are using at least two products on the Samsara platform. In the middle, 70% of large customers and 55% of core customers subscribe to three or more products, and that continues to increase over the last couple of years. And on the right-hand side, 20% of large customers and 10% of core subscribe to four or more products, which is roughly 2x higher than it was just 2 years ago. And the increase in multiproduct adoption is really being driven by all of these emerging products that we've announced over the last couple of years. So outside of our core three products, telematics, AI dash cameras and powered asset gateways, we've launched another 12 emerging products or product families or categories over the last couple of years, including everything today. These emerging products are now approaching roughly $150 million of ARR. And as I said earlier, emerging product net new ACV contribution was more than 20% for the last two quarters. And the final durable growth point is that we have this really nice balance of net new ACV contribution coming from landing new logos as well as driving expansions through our existing customers. As you can see on this chart, very consistently over the last three fiscal years, roughly 40% of our net new ACV has come from new logos and the remaining 60% has come from expansions. Within expansions, more of it has come from upsells or license increases. But you can see in FY '26, cross-sells, the light blue, is picking up as we've rolled out more of these emerging products. Double-clicking on new logos, we still have a lot of opportunity to continue to land more new logos over time. When we look at our CRM, we've identified more than 200,000 core customers that would pay more than $25,000 of ARR. Today, we have roughly 13,000 of them or 6% customer penetration. And even recently, new logos continues to drive a lot of our net new ACV. If I look at just the last three quarters, all three of those quarters are in the top 5 of most new core customers ever added in a given quarter. We also have a significant opportunity on the expansion side with our existing customers. So if we took all of our existing customers and all of the ARR that they pay us today as 1x, and we were to go wall-to-wall across all of our core products, we were to monetize all of their commercial vehicles with AI dash cameras and telematics, and we were to monetize all of their powered equipment with gateways, and went wall-to-wall across all of the emerging products that we've rolled out, we estimate that the uplift would be more than 8x. And even just within our two core vehicle-based applications, the telematics and AI dash cameras, we estimate that today, we're only roughly 35% saturated in terms of the commercial vehicles that we've monetized within our customers. And we've seen very consistent expansion within our customers over time. This is a table of our top 20 customers by ARR. And in each row, the light blue rectangle represents the quarter in which these customers first landed with Samsara and the subsequent dark blue rectangles represent each quarter that these customers did an expansion with us, whether that was an upsell or a cross-sell. And you can see how consistent these expansions are over time, even to the customers that have been on the platform for the better part of 5 or 6 years. All 20 of our top 20 customers have contributed some amount of net new ACV over the last four quarters. So this consistency gives us a lot of confidence that expansions are going to continue to drive durable growth for us. Okay. Those are all the reasons that we think the growth rates can continue to be durable as we scale. We also think we can operate more profitably over time for a couple of reasons. The first, we've demonstrated significant operating leverage as we scale the business, and we think we can continue to do that, most notably in operating expenses. And then secondly, we are hyper focused on managing SBC as a percentage of revenue. So on that first point, we've demonstrated a lot of leverage as we scale the business. On the left, operating margins have improved by 17 percentage points just over the last 2 years. Again, free cash flow margins have improved by 15 percentage points over the last 2 years, and GAAP EPS has improved by $0.52, and we are now GAAP profitable for three quarters in a row. All of that leverage has really come across all of the functions on the P&L. As I look forward, I don't expect as much near- to medium-term leverage in areas like COGS and in R&D as we have a very ambitious product road map, and we're investing in things like AI and hardware to drive durable growth. But I do expect more near- to medium-term leverage to come from areas like sales and marketing and G&A as we improve productivity and we benefit from a lower cost of sale on a renewed dollar of revenue compared to when we first landed that dollar of revenue. We're also very focused on continuing to manage SBC leverage as we scale the business. Our biggest expense by far as a company is headcount, and that gets translated into compensation, which can be bifurcated into cash and equity. As we've scaled and started to generate more cash, we've made a number of structural changes to our equity program. So the mix of cash versus equity that we give as part of compensation and the level of employee which participates in the equity program. Our overall headcount hiring plan, who we're hiring and where we're hiring and what are the equity implications of that. And then the vesting length of the equity grants that we're giving out. And all of those structural changes to the equity program have resulted in a lot of SBC leverage over the last few years. So in FY '24, SBC as a percentage of revenue was 25%. FY '25, we lowered it to 22%. Last year, we got it down to 20%. And this year, we're on track to hit 18%. We really manage equity, and SBC is a real cost of the business, and we expect these metrics will continue to improve as we scale. Okay. So last slide from me, really just summarizing everything that we've talked about. We think the growth rates can continue to be durable and profitable as we scale, again, because we're selling into really large end markets that are growing quickly. Our products address a really large TAM, but we still have a lot of white space opportunity just in core products and in our core geography. Large customers continue to be our fastest growing, and we think that will sustain. Multiproduct adoption is only increasing as we layer in more of these emerging products into the portfolio. And we think that we can maintain this really nice balance of landing new logos and expanding our existing customers. And underpinning all of those durable growth points is the ability to continue to drive more operating leverage as we scale the business. So with that, I'll welcome everyone back up on stage, and we'll get into Q&A.

All right. Thank you, everyone. So now we'll move over to the Q&A session. And just to keep the schedule, I'm going to prioritize the sell-side analyst first, and then we'll kind of go from there. So let's start here with Michael Turrin from Wells Fargo.

Thanks, Mike. Appreciate that, and thank you all for the time. It was a really good informative day of material. I guess I want to focus on, first of all, the tracking label seems like a bit of a no-brainer. And so what I'm curious about is we can't see your road map for emerging products and your ability to replenish those and continue to add to the durable growth profile. So I'd love to spend a bit more time on discovery. How much of it comes from unlocking new verticals, finding things like the cargo theft ROI case? And then how should we think about where the budget comes from when you're presenting these new products to customers and we're kind of gathering all of that input.

Sure. So I'll start, and David, if you want to add anything. At Samsara, we run customer feedback loops all the time. So events like this are great. We spend a lot of time out in the field, and that's where we hear about those operational challenges. Cargo theft is an interesting one, in that it's not something that a lot of companies report on, but it is epidemic proportions, like it's happening all over the place. So as we spend time on site, you'll hear, hey, we had an entire trailer stolen. We had a whole load stolen and that kind of thing. And that's where products like AT11, which is the asset tag that we introduced a few years ago, came from. As we introduce those products, that really catalyzes the discussion for things like the tracking label, right? I don't think we would have gone straight to the tracking label, and customers wouldn't have even thought to ask us for it. But once you show them the tag, they say, this is great, can you make it smaller, right? That same thing is happening on the safety front as we think about the ride-along, like that's a whole new concept. I don't know anyone else in the industry that offers it. But once we started deploying AI after the edge, once we started talking about looking at the entire drive, the whole driving experience, that's where the conversation organically led us to, you know what, we spent a lot of time and money and effort on ride-along. So we think that this kind of concentric circles approach is the way to go. And if you think about it, the perimeter is just growing and growing. So we don't have any shortage of ideas. But we're always thinking about timing, sequencing, applicability, like which industries is this going to affect. So we've got lots more ideas, I guess, is the short answer.

Great. Next is Alex with Wolfe.

Thanks for a wonderful session. Maybe I'll -- I appreciate the high-level takes from the financial portion. If I push one layer deeper on particularly the shipping label product, which seems like it's getting a lot of excitement, both on the floor and the room. Maybe just at a high level, talk about pricing and TAM. I mean I asked Claude what the TAM could be, and it estimates somewhere in the $30 billion to $50 billion. And then I look at the TAM slide, and I actually think the TAM this year was less than the TAM last year. So help us understand a little bit of how to think about that. And then given some of the products are now graduating into this consumption motion as you're -- the power of the platform underneath, how does that layer in and factor into the financial profile of the company?

Maybe start with David just on some the tracking label stuff.

Sure. I mean it's early days with the tracking label. We think the market is, obviously, big. And for us, one of the exciting things is that we have customers today that can adopt the product and make use of it right now, and that's what they're doing. But there's also all these other organizations out there that we wouldn't necessarily talk to today. We kind of talked about NVIDIA and some of these other folks. And we think there's going to be broad applicability there. So we're excited to see what the TAM expansion looks like. And then I think the additional exciting thing is that expanding perimeters as Sanjit just talked about, we get to talk to new customers, that we get to get new ideas and kind of see how the technology tailwinds meet them.

Do you want to cover pricing a little bit, too. I think he had ask a question. So how does consumption pricing work?

Yes. So it is a consumption-based model. Customers commit to buying a certain number of labels a year for the next few years. List price on the label is, per shipment, is $15 per shipment. And then it's enterprise-based discounting like everything else we do.

Okay. Kirk?

Kirk Materne of Evercore. I had a question on sort of on the agentic automation layer, and two, for you all. One, does that compress the phases, meaning as someone starts thinking about AI, do they -- they essentially still need the data. So does that help compress sort of the phasing of -- it might not compress the rollout, but assuming the buying decisions might get faster together? And then secondly, from a go-to-market perspective, when you think about waste intelligence and things like that, those are becoming a much more verticalized offerings. And is it sort of a natural gravitation for you all to think about more verticalized sales overlay to go after those specific sort of opportunities, whether it's waste intelligence or potholes or actually, you're just selling into government? So just kind of curious about the compression and then sort of the go-to-market around that.

So the phases are really interesting to think about because the kind of initial rollout at Phase 1 ends up often being the bottleneck. Amit talked a little bit about the adoption journey that our customers go on. They're not going to be able to take their whole operation offline. So they have to do it in pieces. They might have 40,000 employees, like you heard from Tom. And so how are we going to get this out there is a real big challenge. So I do think once they get through that hurdle, the Phase 2 part happens pretty quickly. And then Phase 3 is brand new. We just started talking about agents today really on mainstage. So I'll let you know how it goes. But we think that it will be faster because once you get through that initial adoption hurdle or hump, now you've got the data. That's this huge, huge unlock. And frankly, that is where the most anxiety is, which is we're going to have to deploy and touch 40,000 pieces of equipment, retrain 12,000 drivers, introduce this into the system. So we are very thoughtful about that with our customers. And then do you want to talk a little bit about verticalization and some of the other operational AI?

Yes.

No. I think the question was more around like how we sell...

From a sales perspective, the only verticalization we've done is with public sector, and we're continuing to monitor and eventually at scale, we will explore it.

Okay. Jason?

Jason Celino with KeyBanc Capital Markets. Congrats on the shipping label. It sounds like a great product, but I'll ask a different question. So the -- I found it interesting, the new go-to-market team on the emerging product upsell. Coincidentally or not coincidentally, we've seen kind of the emerging bucket, and net new ARR, also, kind of uptick. How big is that sales organization today? And is that something that could be getting bigger? Are you putting more resources behind it?

So we started this product specialist beginning of the year in Q1. So it's relatively new, and we're seeing early signs of success, but it's something I'm consistently monitoring to see how we continue to invest and at what time as we release future products. And I'd also imagine that certain products will graduate from that.

Andrew?

Andrew DeGasperi from BNP Paribas. Just wanted to ask a question in terms of changes in the regulatory environment. I mean last month was the Supreme Court decision. I think earlier this month, you had the executive order in terms of extending some of the enforcement beyond the freight brokers to also warehouse operators. So I just wanted to understand if maybe this -- have you been hearing more activity from these other customers versus last month? Could this be an incremental opportunity for you in terms of extending your product within these customers?

Maybe Sanjit, do you want to take that one?

Yes, sure. So I think in general, there was already awareness, especially among enterprise customers, that they need to invest in safety because if you heard this term nuclear verdict, these basically are payouts that involve $30 million, $40 million or $50 million kind of payouts when there's these accidents. So that was already sort of out there. A lot of these more recent decisions relate to third parties. So if you're contracting with -- if you're a freight brokerage or you're contracting with a subcontractor, does the risk sort of extend to them? This is all kind of new area. But I would say, in general, a lot of these prospects are already coming to the table to invest in safety, so the awareness is there. So it hasn't felt like a step change. I don't know, Amit, if you sense anything different among the enterprise customers. But again, the awareness is very high. And then the question is, again, going back to change management, how do we introduce cameras and how do we introduce safety programs at scale into these large frontline workforces?

Chris?

Chris Quintero from Morgan Stanley. I wanted to ask about the infrastructure you've built around the models for your AI intelligence and agents. Are you doing multimodal kind of routing, deterministic elements, probabilistic elements? And it seems like voice is going to be a big way of how your customers are going to interact with your AI and software. So curious kind of how you're thinking about that angle, too. I know you have the investment in HappyRobot. So curious if that's it or you're looking to build something else there?

Yes. I can start.

The three nerds are excited.

Yes. Obviously, yes. So there's a set of things here to our stack that's important to remember. Like many and most of our models run on edge. It's an important element. Those we train ourselves, they're small models, they're efficient, they can run there. It's a key element when it comes to cost as well because that means that we can detect on edge, and we don't have to upload data and whatnot. It scales in a much better way. So that's the first thing to remember. But then we have models that also run in the back end. And there, we use everything, like we train our own models, deterministic, non-deterministic, fine-tune them and whatnot. So like -- and we do tap into all the new models from all the labs as well, including the open source.

Yes. And Chris, you asked about voice. So that's a modality we're excited about. But to the broader point, we are excited about multimodal data in general, right? So images, voice, GPS waypoints, all of these different things coming together. And sometimes different models perform differently, right? And so what's great is there's a diversity of models out there. To Johan's point, we're fine-tuning, but we're also reevaluating constantly because the landscape is changing so fast. So we tend not to be wedded to a single model. We have routers and we can try different things. We're running evals constantly. And I think by the end of the year, we might shift our model mix, again, based on what's right for the job that our customers trying to run.

Matt?

I wanted to ask a question about agents. Sanjit, you had an interesting slide up there about the ROI potential of the maintenance agent. When you read the IDC study about the core product ROI, there's a lot of really tangible sources of ROI, whether it's insurance premium reduction, fuel savings. How do you think about presenting the ROI of agents to customers? And what are the components we should be thinking about?

Yes. Well, with agents, the key is task automation, right? So you can think of it as -- for ride-along, for example, traditionally, you'd have to put someone in the cab with the driver, you're double paying, right? You've got the driver and the other person. So you can start quantifying what is the value of that and then how many ride-along do they do and what's the savings. Similarly, for warranty claims, you might say, well, what was your claims rate before? How much warranty recovery you're getting, that's where your ROI is going to come from. So I think it actually is going to vary agent by agent in terms of use case and also, kind of customer by customer, like what's the kind of frequency of use and so on. We're very optimistic there because we know that there's a lot of task automation opportunity here. Many of our customers, they just say, hey, we are bottlenecked right now. We don't have enough labor to do this. And so we are leaving warranty dollars on the floor or on the table, knowing that if we fill out all the paperwork, we'd get more, but we just don't have time right now. So I think we will be able to quantify more as we roll these out at scale, but we know the opportunity is there.

Let's go with Dan.

Dan Jester, Bank of Montreal. So I really appreciate the update on the network density and the ability to use that to build new products. As you think about growing internationally, that's been a place of focus recently. Has the ability to drive density in new international markets, has that sort of changed your thought process about entering new markets? Maybe to ask it a different way, if you're launching all these new products, do you need to expand in new markets with the same urgency that you would have had maybe a few years ago?

I'm happy to start. So we have been focused very deliberately on a couple of the core geographic markets. So in North America and Western Europe, this is where the majority of the physical operations TAM and value lies. And I think we've achieved some really significant network density. David, you had the slides on the network map. You look in Mexico, you look in the U.K., we're there. So we feel really good about that. And that means that we can introduce products like the tracking label into those markets as well. Over time, we're probably going to expand with our customers into some additional geographies and the density in the network will grow organically there. But right now, we're not bottlenecked on market opportunity by any means.

Let's go with Dylan.

I appreciate everything today. Maybe a dual question for Johan and Amit. It was kind of focused on the product innovation question as well, too. But I think it's interesting with Waste Intelligence and maybe tracking to some extent as well, too, but how you can start to compound that ROI with a revenue-generative kind of value proposition, I guess. Is that a fair characterization? And how do you think about kind of the existing capabilities, but also future capabilities to expand the revenue side of the equation? For Johan. And then, Amit, as you sell that, right, you have more vectors, more surface area and kind of explaining that ROI across departments, how that helps as you kind of move up into executive level conversations as well?

Yes, I can start off. So first of all, like it's still early days. I can say these are definitive answers, but in both, this example of potholes and waste, we have [ $7-figure ] deals in the pipeline there. And the revenue generation component of it is a part of the story with the customer. So there's certainly something new here because most of what we've done to date has been cost avoidance.

David, maybe talk about revenue generation for tracking labels to that.

Yes, sure. So we partnered with early carriers and DCL is on stage with us today, and they're now operating and offering that label to some of their customers, and they certainly view it as a revenue-generating opportunity. This is an upsell that they can provide a preferred customer experience and monetize it. And so I think, again, early days, but certainly seeing signal and appetite for it.

And then Amit, do you want to cover the go-to market?

Sure. From a go-to-market perspective, we love new products, right? I mean it just makes it so much wider. It makes it stickier with the customer. If you remember that graph I showed you, they buy initial, whatever it is, right, whichever product, safety, telematics, and then they continue to purchase, but now my sales team can talk about all these other products, and it just makes it stickier, and we're able to go wide into accounts.

Alex?

Maybe for you, Amit. This is Alex Sklar with Raymond James. Just following up on that go-to-market. So product breadth as a demand driver, just talk about how that's impacted win rates as the product breadth has really increased over the last 3 years for your core products in that initial land standpoint?

Sure. So we are continuing to see great growth with our core products. And this emerging products, it is helping us just continue, as I mentioned, over 20% of net new ACV in Q1 came from emerging products. The specialists are just able to go in and co-sell really well because they have certain knowledge on that product and can sell it. And then after -- over time, we'll see these graduate, and then there'll be new products that we'll assign specialists to if needed.

Alexei?

[indiscernible] from Alexei's team at JPMorgan. So given the impressive scaling in data points that you've collected and just comparing that figure that you disclosed year-over-year, which specific products or perhaps customer adoption behaviors do you most credit with helping you accelerate that scale? And as you think about that moat going forward, what do you see as the most pivotal method of continuing to scale that data capture going forward?

Let's just start with Sanjit.

You've seen that kind of acceleration in that chart. As we deploy our products, we often come up with creative new ideas for data capture. When we first released the cameras, we weren't taking pictures of garbage bins. But now, we see that there's a lot of value there. The same thing with potholes. So the cameras have now become a general purpose sensor for us. Same thing with diagnostics. Initially, we were just looking at fault codes. We're looking at things like miles per gallon and fuel consumption. Now, we're looking at engine performance. So oftentimes, what we find is there's a data opportunity, and there's a question of do we want to gather it, move it into the cloud and so on. When we start to sense there's value in that data, we bring it in. And so that's where a lot of that compounding is coming from. It's multimodal, it's video. It's now going to be voice. It's workflows. We have hundreds of millions of workflows that people are running through our system. And now, it's also data integration. Think about work orders and maintenance kind of work and costing. So a lot of this, again, comes back to the product development philosophy of concentric circles. And what tends to happen is the data follows those ideas. We say, wow, this is a really interesting idea. Let's start gathering the data, and that's what's driving some of that compounding.

Mark?

Mark Schappel with Loop Capital. Sanjit, I wonder if you could just give us an update of your view of the competitive environment and also whether you're seeing some of the platform companies kind of move into industrial operations.

I would say, generally speaking, the competitive set has been consistent with who we've seen in the market for many years. It's the same names, and you should feel free to chime in. And by the way, in different geographies, we'll see different names because they're more regional players. What we've seen though is our differentiation is this platform play. You've seen the multiproduct attach. I think we've done a great job simplifying the operation for the customer. That's playing really well. And then I think your other question is around the larger platform players. I'm guessing you're referring to Googles and Amazons and so on. Yes. They definitely have a footprint among our customers. They might have Azure for VI software or for ERP or something like that. What we're doing is the kind of hardware, software cloud combination that's quite unique and not something that those players tend to offer. So I would say it's very rare for us to compete against them in a deal.

Any other questions? Sure. Okay.

This is great. I'm just curious how you think about the long-term risk from AVs because you talked in the keynote about robots and others, but one of the automations that are likely to come into a core of your customer base would be autonomous vehicles. How do you think about fitting into that world?

Yes. So AVs are something we've been excited about for a few years, and we're headquartered in San Francisco, which is like where all the AVs seem to be born, right? Like you see them all over the street. So for us, it's really been about expanding the opportunity. We see a way that if you think back to what our customers have talked about on stage, many of them have aspects of their operation where they may be labor or even driver limited. So AVs present an opportunity to move some of that freight from point A to point B. And so we view it as an and. They're going to want to see it in their operations. We also think there are going to be many different types of AVs. Eric talked a little bit about how they have autonomous diggers and material movers and things like that. Those tend to come from different OEMs. And he also mentioned how looking for a single pane of glass to orchestrate his operations. So we think that's going to be fundamental as well as connecting all the workflows together, driving automation using agents, making decisions and then maybe dispatching AV or dispatching a person or some combination. So that's the opportunity we see. And it's still, I have to say, very early days on the AV front. Most of our customers have been experimenting a little bit with it, and the counts are still in the thousands. Eventually, it will get to tens, hundreds of thousands, but there's 90 million commercial vehicles in operation between North America and Western Europe. So that's why we see it as an and for many, many years.

[indiscernible]?

Not AV related as much because if you kind of think about more connected devices as in vehicles, whether it's through -- I mean, right now, it's mainly Bluetooth, right, like in terms of what a lot has work on. But if you have direct-to-device connectivity through LEOs and other things over the next few years. Does that open up a surface area for you guys to have sort of more products and capabilities?

Do you want to talk a little bit about Hubble?

Sure. Yes. So we're early investors in Hubble, start-up company that's working on the possibility of Bluetooth to space, Bluetooth in air quotes. And we're excited about what that could open up for our customers and what that can mean for visibility. It means that over the ocean is all of a sudden a possibility. It could be that things get tracked everywhere. It's really early days on it. We're, I think, probably first people experimenting with the certain technology and seeing water can even run through the pipes. We'll see what happens. We're excited about the concept.

Yes. And then expanding that concept, there are a number of new constellations coming up. I think the promise is there, that we will be able to connect directly to the device. Many of our customers operate in really remote areas. They're literally building the roadways or the energy pipelines. So we know there's interest in operational visibility, connectivity and eventually, things like video. Technologies like Hubble would provide us like basic location, things like that. Eventually, we'd love to be able to get streaming video for operational AI and intelligence. So I do think that, that will expand the market opportunity, but we are kind of behind the connectivity coming online.

Alex?

Now it's round 2, so we can get a little spicier here. The other two companies you put on that slide of size, scale and growth, Palantir and Datadog, they have customers that pay them multiple times what some of your largest customers are paying them. Again, given whether it's the shift towards consumption, whether it's the fact that 8x, if you sold everything to everyone at all attached that you can realize in your customer base. I guess maybe, Sanjit, for you, as you think about the density of absorption within your largest accounts, where you go from 5 million to 10 million to 20 million to 30 million to 40 million, like what does that look like? What has to happen to unlock that opportunity that allows you to kind of take advantage of this much larger pie?

I think Dominic had a really interesting slide that showed our customer profile and how it's been continuing to shift up over time. It's worth noting, our company is 11 years old. And this enterprise opportunity, it takes a while to nurture. You have to have the right sales efforts, and then they have to adopt the technology. So I think over the next decade, you are going to see our largest customers get larger and larger. And then they have to be ready to adopt. A lot of those -- if you think about a Datadog or Palantir, those are fundamentally IT-oriented products, and the market is primed to consume a bit faster because they don't have to put hardware devices on the bulldozers and that kind of thing. So that's the practical side that we also think about, is in the fullness of time, we know that these customers can get very large. You heard about the scale that Performance Food Group operates at. Like there are so many things we can do for them, but they need to put us on each of those trailers. They need to get this out to the front line. So that's why we think of it as the opportunity is absolutely there. We're going to continue to build out the platform, but it's not going to be all at once. It's very rare for someone to land at that scale.

Maybe Amit, like as Johan and David continue to build more products is awesome. Kind of what's your view from a customer perspective? Do you think about opportunities as you assign them out to accounts? Like does it get bigger, smaller? Like how do you think about that?

I think I would echo what Sanjit said. It will continue to be a bigger opportunity, but the change management on the customer side is a lot. So that's where they'll take time to adopt it.

Michael?

Let me give Dominic a question.

[indiscernible]. I was trying to block him actually.

I noticed.

So one of the questions we get about Samsara is just rising or just across technology, but just rising component costs and impacts to financial profiles. We know you have a sophisticated procurement team. We know there have been periods previously where we've seen some free cash flow margin impacts from ramps in costs. So can you just speak to your ability to manage gross margins and free cash flow margin in a market that's seeing cost escalation?

Yes. I mean -- so for this year, again, I think we're trying to have gross margins be flat to where they were last year. And we've guided to still in this increased price environment, 100 basis points of free cash flow margin. And so I think that we're doing a pretty good job of managing it. And again, we've operated through these cycles before. If you go back to 2022 coming out of COVID, supply chain shut down and couldn't turn on fast enough to kind of meet all of the demand. And we were in a much different financial position at that time. Now, we've got $1 billion, $1.5 billion of cash. We're generating positive free cash flow. We're well capitalized to be able to handle this. And the supply chain, our team is more sophisticated and the relationships that we have are stronger. So we actually view it as -- we feel very well equipped to handle this period of time, but also as like a strategic advantage in that, we are the biggest company. We're the most well capitalized. We have the best strategic relationships. You can hear the customer demand just from the customers kind of on the panel and their kind of desire to go through digital transformation. And so it's really an opportunity for us to drive more market share gain and really take advantage of our -- how well capitalized we are as like a strategic advantage.

Kirk?

I don't know if this is for Dom again, I won't put you on the spot, or maybe Amit. But just how do you think about scaling internationally right now? You're seeing really good product market fit. There's always a concern that you scale too fast, you have too many people in the field. But how do you think about that right now? Because it would seem all the same opportunities in Western Europe, Mexico exist that they did in the U.S. Just -- I'm just trying to think about the pacing on that, I guess.

I think we really just look at the data and like how productive we are and what we're hearing from kind of customers, and we're very dynamic in the way that we're kind of rolling out how we're allocating capital. We allocate it across different -- obviously, the product features for different geographies have kind of different needs and obviously, on the go-to-market side and how much kind of capacity we're ultimately adding. And so we really just look at the data and respond. we have the advantage of not being in a situation where these international markets are as fragmented as they are in the U.S. So we don't have these like really large competitive incumbents where we have to kind of race out with really bad unit economics to get into these markets to win. We can kind of take our time and pace the investments in the capital based on what we're seeing and making sure that we're not kind of chasing bad unit economics. So I'd say it's very dynamic. International is a huge important part of the durable growth strategy, and we feel good about the investments we're making this year.

Matt?

This is Mike Richards for Matt Hedberg at RBC. Maybe just something we didn't talk a lot about today was Agent Studio. And so I was just curious your thoughts around the ability to productize bespoke agents that these companies are making and how that will accelerate your product road map.

Maybe Sanjit or Johan, kick it off.

Yes. Yes, sure. So in this Agent Studio, that is exactly what it is for. Customers can create bespoke agents, and they can specify them in plain English and what they want them to do. We don't have all capabilities in there, but we see this expanding such that they can truly integrate into their operation and do a multitude of tasks. So as such, we see this as a very exciting path forward, especially as part of the platform and the larger connectivity that we have as part of that.

And the way we think about productization, we've talked about operational intelligence in Johan's demo. So you saw like Waste Intelligence, for example. Some of those, we may see customers be able to get towards with Agent Studio, but maybe to really go big, to really make it amazing, we're going to need the product teams to get involved and build some more capabilities. So this is what we mean by running feedback loops. We're going to keep an eye on, well, which are the breakout templates that people are really attaching to. We're going to spend time with those customers to understand what else could we build. Are they set with Agent Studio and the consumption model there? Or would they like a full-blown product? And that's where we have to be dynamic.

Great. All right. Any other questions? Okay. Let me wrap up then. Okay. Well, thanks for joining us today and learning more about the Samsara story. Goodbye, everyone, and we'll see you again soon.