Hey Nerds 👋,
This company can’t stop having news. Whether it’s raising at a $60bn valuation, applying for an OCC Trust Charter, the full launch of their harness, or doing skits with Tony Hawk and Bam Margera. Ramp is everywhere.
And unlike nearly everything else in finance. They’re winning the marginal VC dollar in the age of AI. How is this even possible? I had to ask.
I sat down with Eric Glyman, Co-Founder and Co-CEO of Ramp, at its New York office to get into exactly that.
When Eric and I spoke two years ago, we talked about why banks would struggle to build products the way Ramp does. This time, we’re talking about why this “spend card” has routing tokens, and the answer was surprisingly simple and rational. Eric’s argument is that once software has a variable cost, helping customers get more from every token becomes part of the same mission.
Eric has one of the most fascinatingly simple ways of thinking about business. To him, his business can be reduced to a simple equation. AI merely changed the inputs.
There’s a lot for the Fintech Nerds in this one 👇
🧠 How on earth can someone underwrite a $60bn valuation? 100% YoY growth, having accelerated 25%, multiple 9-figure products, and 70% of Ramp’s R&D team is working on products it hasn’t released yet
🧠 Are they a Fintech trying to be AI? Eric says they never set out to be a Fintech company and are more of a productivity company. When tokens cost money, and AI saves time, they become core to the mission.
🧠 The internal tool-to-product pipeline. Eric says Inspect, Ramp’s background coding agent, now writes 85% of its pull requests. Glass, their harness, enabled new levels of productivity (and is partly why 70% of staff are working on new products, not core products).
🧠 What changes when non-engineers can ship code? Eric says they merged 7,000 pull requests last quarter, with customer support staff able to help fix the problems customers report.
🧠 What’s behind Ramp’s national trust bank charter application? And where stablecoins help it move money faster.
🧠 The big question. Why would you design your company in a way that’s not ready for AI? And how do you build a business an AI agent would choose, even with perfect information about every competitor?
Keeping up with Ramp could be a full-time job at this point. Cards, bill pay, procurement, accounting, AI tools, a charter application. And somehow Tony Hawk.
But I wanted to understand how the company works from the inside. How does a support ticket become a code change? How do you measure whether AI is making people more productive? And how do you decide which of the tools you’ve built for yourself should become a business?
Eric gets into the mechanics.
This series tests my thesis that “fintech” stops being a useful label when the ambitions of its fastest-growing companies extend so far beyond finance. Ramp gives us plenty to work with.
Enjoy the conversation.
ST.
Meet the people building this stuff at NerdCon
If you want to get into how these tools are built, come to NerdCon.
The person who built Ramp’s internal Glass harness is giving a talk on how it was built. You’ll be able to hear from someone who made the decisions, then get into the details with other people trying to solve similar problems.
That’s the kind of person we’re bringing together: engineers shipping AI tools, product people changing how work gets done, risk and compliance teams making new products work in practice, and founders building the next generation of financial services.
Bring the problem you’re working on. Find someone who’s wrestled with it.
Join us in San Diego, November 19–20.
Full transcript
Simon Taylor · 00:00
Welcome to Brainfood. My guest today is Eric Glyman, co-CEO of Ramp, and this is a company built to stretch the definition of the word fintech, possibly even make it obsolete. Ramp started with corporate cards that just kind of work. Now it does payables, procurement, accounting, AI model routing, and Glass, a workspace for getting stuff done with AI.
They've also produced a Broadway show about guys called Bill who have too many bills, and put Tony Hawk to work with finance and a lot of ramps. This is a company that knows a little bit about marketing and standing out too.
Eric Glyman · 01:19
I actually think that these, like, very bizarre brand sponsorships are much more interesting.
Simon Taylor · 01:25
So much more. I mean, by the way, Tony [Hawk] and, oh yeah, yeah. I'm I'm a big fan of the [Tony Hawk video]games, as you know. So yeah, that was awesome, man.
Eric Glyman · 01:34
It's, I mean, for me too. So I've never had so many text messages from like high school classmates. Like it got on Saturday. It's crazy.
Simon Taylor · 01:43
Anyway. Eric, how the heck are you, man?
Eric Glyman · 01:51
I'm I'm doing great. You know, living's good out here in New York.
Simon Taylor · 01:54
It looks like it. Looks like it. Well, thank you for having me at your office. It's lovely. You know, I've got the whole thesis that like fintech is dead, but I think when there's one company that comes to mind that represents that, it's kind of you guys because like an AI harness and a router has almost nothing to do with fintech. So do you agree fintech is dead, or are you kind of against that idea?
Eric Glyman · 02:16
Was fintech ever really alive? You know, like what is fintech if you really think about it, I mean, let's take let's take some of the classics like the Finn, Wells Fargo, American Express. What was their first business?
Simon Taylor · 02:31
Stagecoaches and horses and moving piece of paper around.
Eric Glyman · 02:35
That's
Eric Glyman · 02:37
It's the Pony Express, right? And like, what the heck? Like these labels or categories have to do with the global financial conglomerates that they become, and I think companies evolve. This happens all all the time in all sorts of industries. You know, when I think about how Ramp even showed up in the fintech category to begin with, we started with a product which was, you know, a card that wanted you to spend less, that help kind of reduce your expenses, and you know we move money. But in some sense, I actually always thought of Ramp as a productivity company, not how do you move funds from A to B, but how do you move fewer? How do you waste fewer dollars?
Simon Taylor · 03:13
And in that sense, a router that helps you waste fewer dollars or do things faster, and even an AI harness like productivity. You can make an argument for that, but like, what is there some extreme level where you wouldn't go? Like, how are you thinking about the limitations of that?
Eric Glyman · 03:32
Look, like, I don't think we're going to be selling tacos anytime. But look, maybe
Simon Taylor · 03:38
on the marketing side, who knows? But
Eric Glyman · 03:41
don't give my team any ideas. Be careful. But I think that just for a comment for a minute on Glass and Router.com, something very subtle but important happened a few years ago when tokens started to have a cost. I mean, it used to be such that if you were using fell in love with Excel and you know you were using it for 10 hours a day instead of one, might need a new hobby. But it wasn't a big. You wouldn't get a larger bill from Microsoft. There weren't really larger costs, or even Salesforce, like on the software side, they didn't really have larger costs that they would experience or that they needed to pass along back to you. With AI and tokens specifically, the more tokens you use, the more it costs because their electricity costs go up. They have to allocate demand or their their limited supply, and so, the more you use, if you go from one hour to 10 hours in AI, you might go from $1,000 to a $10,000 bill. And once software started to have marginal costs and variable costs, I think software in some sense became money. It actually became a financial product.
Simon Taylor · 04:58
Yeah,
Eric Glyman · 04:59
And I think people don't talk about it enough. But the minute that suddenly large portions of companies spend actually became software that could could ride up and ride down, and were hard to predict, it actually became almost necessary for Ramp to start looking at and saying like, why should companies be spending this much? Can they spend less? How do you get more for every dollar, hour, or token? And so we actually view that as just really entirely consistent with the original thesis of trying to help make companies more productive. But it is interesting, right? It looks very
Simon Taylor · 05:33
different to like a spend management card, or procurement, or to sort of all of the things you've been doing around it. Even stacks, like for accounting, feels like an adjacency. It's very much not like the traditional financial services product. It's adjacent to it. But this one feels like there are entire companies dedicated to doing Router.com or to doing Glass. So, did you have to build this yourself, which made it easier to externalize? And and is that a common product trajectory for you?
Eric Glyman · 06:04
It is. Yeah. I would say in some sense, like Inspect is maybe one of the best examples of this, right? Where I mean, even go back to the name. Why is it called Inspect? Where did it come from? And you know, if there's any front-end developers listening to this, they might say, like, "Oh yeah, when I right-click something on Chrome, it'll say Inspect Element."
Simon Taylor · 06:27
Yeah.
Eric Glyman · 06:27
In the origin, it was a little over a year ago. There was someone on the front-end engineering team, brilliant engineer, was experiencing everything that AI could do and what it was changing, and he said, "I right-click what's going on, see the bug, see the issue, and just have AI go and you know patch the code, maybe fix the front-end code, change the color of the button, change the dimensions, whatever it is, and go from there. And he kept pulling on it. This is great, but you know it'll come up with a code change. But when it actually merges, it doesn't render the way that I assume. So let's give it a virtual machine. Let's give it access to our codebase. Let's go and build controls. Let's go and actually take it further And further, and now, 85% of all the pull requests written on Ramp are written by Inspect. It's a background coding agent. Last quarter, we had 7000 pull requests written by non-engineers. The majority of them, people on customer experience, read the above report, and it goes. And I guess what I'm getting at is, I think this desire, we happened to bang against it much earlier because we had a very forward-leaning organization, very agile, who wanted to use the latest tools, and I can give a couple of different examples. And the products weren't well suited, and so it became an internal product. And you know, we want other people to experience this. Glass is a great example too, and even, you know, expense management. If you take kind of the original innovation, we were the first again to the fintech to say Why do you know, maybe we should do expense management? Maybe we should get rid of it. You know other fintechs like would integrate with it, but they never actually did
Simon Taylor · 08:21
it. Yeah, and you're trying to almost get rid of as much of the experience as possible. I always describe it to you guys as being an aggressively efficient user experience in that, like, hey, wait, I got it. Let me try and make that happen a little faster. And the name ramp, the days. Ramp. Com. There's this like very obvious attempt to create velocity here that's going on. I guess AI has accelerated that to some extent. But if you think two three years back, what does daily life look like now that is different than two three years ago for you and for folks on the team? How would you say that's changed?
Eric Glyman · 08:59
Get a lot. We were joking about this before we started recording. I get more tired. It's like over the past week. Like let's talk about it. We did. There was the announcement of Glass and the launch of that, and there are 1,000s of companies that have signed up, and we're onboarding people as quickly as we can to the fundraising announcement, which is fun. We collaborated with Tony Hawk and Bam Margera, Jared, and the New York Stock Exchange. We submitted an application to the OCC. Like, the speed of everything is so much faster when you really internalize. What does it mean if every employee has access to tools that can accelerate cycle times, ship features faster, do research faster. It means what normally would have been months of work can get compressed down into a week. We've functionally had a major product launch every month for the past five months, and so it's it's fun, but. It's a lot. It's
Simon Taylor · 10:01
exhausting. And look, as somebody whose full-time job is to try and keep up with what fintech companies do, I could make the full-time job you guys at this point. It's kind of ridiculous, especially after the last week. And we'll come to each of those pieces individually because I think they all warrant their own space and time. But I do think it's something about being inside the singularity is is really really tough and really interesting. But on your AI products specifically, do you see them? And is a goal for those to compete with the dedicated companies who do that thing? There are companies selling internal harnesses for companies, you know, like like Glass. There are companies selling routing software for tokens. Is that the goal, or is the goal just, "Hey, I've got a client book; they might find that useful," and it worked for us?
Eric Glyman · 10:51
It's it's almost something a little bit different, right? Like if the if the overarching mission of the company is to help every organization get more from every dollar an hour, and a major like I think AI is becoming this third major input to how companies do business. Maybe just to give like our internal framework, like classically, if you were converting capital into things, you could sort of convert that into two uses instead of firm. You could go hire people,
Simon Taylor · 11:21
yeah,
Eric Glyman · 11:22
or you could go hire a vendor. That was about it. Yeah, those are your two choices.
Simon Taylor · 11:28
And you called it that third pillar of spend, right? And I think when you did the the last blog for the evaluation announcement, however many months that was,
Eric Glyman · 11:36
Yeah, it's uh, and it's even gotten for me like sharper as as we think through it, which is, you know, as we've seen spend even get larger, we're now past 2% of all white collar labor spend is the equivalent of what's spent on AI. In some ways, it behaves like people. It's like most spend is agentic. You ask a request, and then an agent hires an agent, which could have another agent, and so you actually don't know how much you're going to spend. It's also rented intelligence. It's not really your own employee, and the companies that are spending effectively on this are growing much faster, and they are hiring faster. And if our mission is to help our customers succeed and have aligned incentives, we should want them to be able to use this well, and at the same time, it is this major category of spend which organizations are very ill prepared for. And then it becomes the question of what's the right form factor. We started that question. We said, well, actually, if you want to help companies control their spend back in 2019, corporate cards are a really important way managing bill payments, and so I think that if you kind of look at this, this was the surface area, and so it forced us into the form factors of where these types of spend are occurring or getting controlled, and we needed to be there, is what pulled us really into it.
I think that the other, you'd ask a little bit of like, who are we competing against and what makes us really different? I think a lot of companies get stuck in the trappings of they sell you X product and that's their competitive set and that's what they're trying to do. What makes maybe financial and maybe fintech business is really interesting and I think maybe rampant, interesting case study is, if you thought about like our original business model, we would monetize in some sense the way that a large card issuer would, but we would use that to compete against Concur, or against Expense Buy, or against Bill, or against Coupa, or name a soft a software product, and that allows us to tap into different profit pools. In some cases, they're greater to go invest in making a great product in a market that is under investing in the in these areas. And I think that when you have aligned incentives, like I don't sell tokens, we buy tokens. You think really differently about what companies should need.
Simon Taylor · 14:00
I think that's logical that you've already sort of proven one revenue stream can fund a product that goes somewhere else and competes something else, and that's kind of an interesting business model. Heard a few folks say of the companies that have a real shot at financing intelligence, Ramp is Ramp is probably one of them. But what I find so interesting is that there are companies out there that might, in your view, not be spending enough on AI, right? Because you you say yourself, the companies that are growing the fastest are the ones spending on the most. It's correlation. Your data team is is quite good at pointing that out. Would you ever point that out to a company that could grow? Because that is an aligned incentive.
Eric Glyman · 14:42
It's super fascinating. The short is, yes, we would very much like to, and ultimately intend to. In some sense, like every company is engaged in. No matter what you sell, the goal is to turn $1. Into more than that,
Simon Taylor · 15:01
yeah,
Eric Glyman · 15:02
and you can take your strategies, you can design your departments or your products or you know whatever it may be, but those are functionally means to an end. And part of where Ramp sits, which is very different than most companies, is we see dollars going out. We're increasingly helping companies store funds, and with the launch of receivables, bring them in, and we connect things into your store systems of record, and so we can actually start to run this analysis of, okay, you spent more on AI, or you spent more on ads. Did you grow faster?
Simon Taylor · 15:39
Yeah, interesting.
Eric Glyman · 15:40
Did you not like you know and and you can start to actually see these qualitative differences between firms and as time goes on, you know just as we're trying to maybe build an Iron Man suit for every business to run a little bit easier to that they're wasting fewer dollars, fewer hours, everything gets easier to do. I do think as time goes on, that naturally extends into can you help them allocate capital?
Simon Taylor · 16:05
I like that idea of profit pools because I think about SAP and Oracle and QuickBooks and Intuit. Like you're already in the QuickBooks Intuit space, but then there's entire categories of companies that look at your ad spend and entire country and trying to pull. Then there are companies that pull that into a dashboard. There are there's this like patchwork of little SaaS and AI companies that you're kind of competing with in terms of profit pools. And I imagine for investors, that kind of makes kind of a grounded part of a of a valuation that could also be more thesis driven because there aren't many companies in finance that look like they've got a story in AI right now on a 60 billion raise with 2 billion of revenues, 30x. That's a lot more than Block. It's a lot more than PayPal. And Block had a decent growth year last year, albeit not nearly as as as much as you guys. So is is that how you're grounding it? That like, hey, the ramp has got a lot of room to run because we're going after all of these other profit pools, and our growth continues to accelerate because we are focusing on this. Or do you think there's there's something else that the investors are seeing in you guys that maybe some of the folks that look at that price of the company and get sticker shock might miss?
Eric Glyman · 17:19
So It depends over what time horizon you're looking at things on. Right, like if you are looking at what is the last 12 months, or you're looking kind of behind you, everything that grows fast looks incredibly expensive, and everything that grows slow looks very cheap.
Simon Taylor · 17:42
Yeah.
Eric Glyman · 17:43
If you look ahead and you think about the durability of that growth, sometimes you might look okay. If we were to look at the expected revenue or contribution profit, gross profit, take your favorite your metric, and you look at the multiples that point in time, this is part of why people really like fast-growing companies. It turns out they might actually be very underpriced because people tend to under underprice what it means for companies to grow really quickly and to continue growing really quickly, and in Ramp's case, accelerate at scale. And so there was sort of two major theses that were at play in this most recent round. One set of investors, I would say, come from a more classic lineage of investing, which is it's dollars and cents, market size, and you can do calculations in a spreadsheet. And for them, what they saw was an incredibly unique financial profile. A business that's growing more than 100% per year, growing more than 25% faster than it was this time last year at much larger scale, and is generating cash. And the TAM, by the way, is enormous. Like it's like when you think about corporate cards in the U.S., which is a subdivision of it's it's not even the majority of the world for sure. Like we're maybe three to 4%, and so and you think about general payments were not yet 1% in the U.S. And the physics of it suggests that this could keep growing for a really long time, and so I just think the number of businesses on the planet with that financial profile are you can certainly count on one hand and debatably like there just aren't other other ones like that, and when you do the calculation of if you look at what is the growth adjusted growth growth profit? If you try to back out, what does it mean if one company is growing 20% versus growing 100% or 90? And you try to figure out
Simon Taylor · 19:52
three to five years exactly.
Eric Glyman · 19:54
You don't even need to go that far. Like when when you start looking like one year out, actually. Many investors believed they may be right, they may be wrong, that ramp was just on the financial basis, even if we didn't do new products at all. And then one of the things that Kareem will sometimes get at in conversations with investors is something along the lines of the following: of Look, like the majority of people here on this floor, 70% of the R and D team, are working on products that we have not released yet.
Simon Taylor · 20:29
Yeah.
Eric Glyman · 20:31
And if you follow the logic tree, like either we are incredibly bad allocators of capital and we're just burning money, Or we might be working on things that are hard to model, but might lead to new S curves, new types of products that will grow explosively. And I think if you had looked at Ramp in 2020, you would have underwritten. You might have underwritten as a card business. If you look at it in 2022, you might say, "Okay, well, there's cards and there's bill payments." In 2024, there's like a large and growing software business line. You look at it now; it is a it is a comprehensive platform that is building more and more. And so, you you have another class of investors that I would say we're often more associated with earlier stage investing, investing very heavily into this round in part because they they see the potential to help companies operate much more efficiently in this AI era, and so it's a bit of both. Yeah,
Simon Taylor · 21:44
I guess there's a dynamic that that class of investor really recognizes, especially when cash flow positive, which is that's an efficient team, right? Like, yes, you can go on a million side quests, and they all don't need to succeed, but they can be if they're affordable to do. I mean it's an interesting AI dividend that you guys have got, having been able to become masters of using it yourselves to be more productive. You can then turn those tools external. But like, what's the rough split across your product lines today? You know, so you were mentioning card and bill pay and that sort of stuff. What does that kind of look like today?
Eric Glyman · 22:20
Yeah. Look, like the way we're pacing in the next 12 months, the majority of the business won't be card, which is very shocking to people who think about ramp and
Simon Taylor · 22:32
the card.
Eric Glyman · 22:34
Yeah, you know, even on volumes, you know, On non-card payments that are monetizable, this isn't just transfers. Or I think a lot of fintechs are pretty sloppy. Like they'll do a bank transfer from A to B. There's no monetization, and they count it as TPV. Like it's it's it's it's not like excluding all that stuff. Like move over quarter trillion a year, growing well over 100% on non-card volumes. Software is a you know multiple nine-figure business line growing extraordinarily quickly. The new businesses, router, stack, glass, growing extraordinarily quickly. Substantial enough that even investors seeing kind of the growth curves on it are attributing significant value. Like, like one way to think about it is like the majority of our customers are using multiple of our products. I think over a quarter, nearly a third, are using three or more, and so what's happening? And I think part of what's fascinating, even back to your earlier question about financial profile, you have this this this triple effect where we are growing logos in many months faster than we were, whether it's like a year or two years ago on a percentage basis. So there's more companies; they're using more products, which each generate their own
Simon Taylor · 24:11
cash. Yeah,
Eric Glyman · 24:12
and companies are getting larger. Like we we were so thrilled. I think Uber gave a fantastic quote as a part of at least for the latest round. To we have leading companies at enormous global scale, and so it's leading to the business itself is getting just radically more efficient, and it's allowing us to continue to invest at a time when I just think most competitors, certainly in our industry, are just not. It's not even just a pace thing; they're just not putting real dollars to
Simon Taylor · 24:52
it. Yeah, you can see cents and dollars, and I think the efficiency side of what you guys do and having actually harnessed AI, I spend a lot of time with. CEOs and and you know folks like how do I use AI and I'm just going go read the Ramps blogs because I think there are companies like yourselves like Stripe and a few others who've who've been able to effectively bend the curve in their feature velocity and put Robinhood in that category as well. But I'm interested in what it looks like for you and your dashboard of the team success, you know, Geoff put out an interesting blog post about how you're measuring people and how all of that's changed since AI came along. Can you talk to me about what that looks like from your seat?
Eric Glyman · 25:31
So some are really basic of in our sales and go-to-market organization. The real test is: Are we selling more? Is the productivity you know going up. Really? Yeah. You know, it's it's it's funny. Like people sometimes ask for these like very exotic measures. Like I talk to their founders. It's like, you know, are you looking at like PRs or stuff like that? And it's like, no. Like, your are your salespeople more productive? But I did like the point
Simon Taylor · 25:56
about non-engineers that have pushed PRs. I think that's probably an interesting one.
Eric Glyman · 26:00
Yeah.
Simon Taylor · 26:00
And I think in Jeff's blog post as well, it was like levels of AI. I think yes, capability.
Eric Glyman · 26:07
I I think it's an incredible post to kind of just evaluate of how are people really using AI. I think one of the sand traps that many organizations find themselves in right now is one: everyone you hire are paid to be professionals. They like to appear as if they know exactly what they're doing, and so it's very structurally hard for someone to say, "Boss, I have no idea what I'm doing. Like, I don't know how to use this. Like, you know, in our industry, like, I meet bankers all the time who used to use Excel in Windows, and then you give them a Mac, and they're like, I can't do it. It breaks my brain. It breaks my brain, and it's like it's the same program with some different mappings of keys, and you know that's hard. And you look at people that you're asking them to just like just do all your work through AI. It is so it is such a big ask, and so I do think that Jeff's framework is a wonderful one to kind of get a sense of how AI native your organization is. But I also just think looking at the speed to do anything is is interesting. And the reason a lot of organizations try to manage towards outputs, we're much more obsessed with understanding the inputs and speeding them up. And so maybe an example in like the product development world to unpack this is you know we talked about last quarter 7000 pull requests were merged by non-engineers which is a lot right two three years ago it was zero like if you were not an engineer you weren't shipping code what does that really mean so in the old world, if you were on the customer experience team and you got a bug report from a customer, at most organizations it's well SLA to say sorry. Here's our help. At a better organization, they would they would sort of categorize the pain points, and then once it hit a certain threshold, someone would go take that. Could be from support. Could be from product operations, and categorize and write like a user story of, we have this bug. Here's the issue. Here's the proof. Then it goes into some Scrum board or agile method, right? Where it goes onto someone's list to go prioritize. Someone decides is this important or not? That exercise could take a week or a month. Who knows? Hopefully, it makes it onto someone's list. Then some, you know, product or design person figures out what's the right way to solve this problem. They design it in. They send it to the engineer. This sounds so painful, but then like two, three months later, maybe the bug is patched, and the customers who are annoyed have moved on. Have probably never heard or were informed. What something like inspect means is that someone in support can see that, or AI can be reading all of the bugs,
Simon Taylor · 28:56
yeah,
Eric Glyman · 28:57
all these emails, and generate a design, and ship it, and patch it, and test it, and email the customer back in 15 minutes. It's fixed, and so everything gets dramatically faster. And so sometimes, what I'm trying to get at is there are certain types of adoption of how do you get people to culturally get over where you're trying to make the person faster, but there's I think, the more interesting question for organizations out there is, why have you designed your organization the way you have in a world where computers can think?
Simon Taylor · 29:32
I had this wonderful conversation with David from Nubank, and he said the best thing about AI is it's allowed us to look at every single process and question it, and a lot of the benefit was just changing the process. It might not even be from AI. Some of it comes from AI, obviously. But I observe about Ramp is I think a helpful thing is you've always been obsessed with clock speed. Like that's sort of very core to the name of the company and everything else. So if you can be obsessed with clock speed. Then something that helps you with that is culturally sort of sensible for you. I think what's interesting on the talent side as well is you said in an interview I think with David Sandra that you see yourself competing with the big AI labs now for talent. So what's your pitch for like winning somebody that could go there and to to come here instead? I mean, the New York office is nice.
Eric Glyman · 30:22
New York office is nice. No, look, it's it is a fascinating time right now in the world, and people have very different skill sets, and they're, you know, maybe 1020 people every year who their brain is just wired in a beautiful way, touched by God. Right where if if if you want to design extraordinary algorithms and think about the breakthroughs required in order to go and there's design novel architectures for for learning or reinforcement or there's a number of challenges that you may only encounter at the labs and perhaps the the shape of your brain is is suited to that. You should go there, but there are a lot of people who are extraordinarily brilliant, who brains whose brains thinks a little bit differently, and I think appeals and are attracted by what Ramp is able to do. In some sense, what Ramp is trying to do is be a lever on the world. If you can go and make every organization 257 and a half percent more efficient every year. What happens? Well, the median customer on ramp grew their revenue last year by 16%.
Simon Taylor · 31:46
The
Eric Glyman · 31:46
median company in the U.S. grew their business by about three to 4%. We are helping organizations go faster, and if you were the kind of person who was interested in not just beautiful math, but actually understanding how we can write code, software, build products that will make regular people's lives better, Ramp is an incredibly fast way, high torque way to do that. You know, we have organizations that are trying to make sure kids are somewhere safe to go after school. We serve over 10,000 nonprofits. We have organizations that are trying to fix the healthcare system. I think that when you actually can go and build tools that can demonstrably help people get more from every dollar an hour, I think that the mission attracts a lot of people. Of it's not just interesting, you know, abstractly, but you can see the results of your work. I think is a big driver. I think that the other really strong pitch is this is a place where founders come and people are building new parts of the product all the time.
I think people aren't interested in just maintaining always. Some are, but like maintaining 10-year-old software. And you look at at at Ramp. You know, every month we're launching major new products. You get to work on small team. You get to use the best tools. We were Claude Code's first customer in their alpha. They have a great blog about that. Cognition in Devon filmed their launch video out of our office. You know, they're an extraordinary company. Cursor. We were one of their first 10 customers at Ramp because we believe in love of the craft, building extraordinary tools, being on the edge. Like you are going to use world-changing technologies often before they've changed the world, and sometimes that leads to better products that are helping customers. Sometimes it leads to great ideas. Maybe you'll you'll take and build with.
Simon Taylor · 36:09
Another thing we did talk about is like one of your kind of approaches to velocity is you have a handful of partners like Stripe and Marqueta and many others. Is that still very much the model for you, or are you going to go further down the stack and eat
Eric Glyman · 36:27
it? We love partnering with organizations, and I also think too, in general, I think people are too zero sum about this. Like one of you've become a great mentor, an investor in Ramp is Satya Nadella, and you know I kind of think that if Microsoft and Apple could bury the hatchet and become great partners, and they certainly are, like there's hope for the rest of us. You know, you think about businesses like Netflix. I think at one point Netflix may have been the largest customer of Amazon Web Services, which is amazing because they had Amazon Studios and Prime Video trying to go compete directly with with them, and I just think that the world is so big that the idea that someone has margin, you need to go and take it is, I think it's just a very old way of thinking. I I I think there are certain areas that you know capabilities that you know we've hired a vendor to do that perhaps at some points should get in sourced for for some reason, but I I think in many ways like the partnership with Stripe, it's a good example has never been stronger. You know, we also have, and I think sometimes that surprises people. They say, "Well, they you launched router.com on the same day they announced that they bought OpenRouter." Yeah, and it's like, well, one we didn't know, but two, it's also like it's fine. Like both of our businesses are very small in the grand scheme of the world, and so some other founders think very differently. I think some founders are just just want to wear your face and you know take all your stuff, and you know we could debate which founders are like that, but like that's not that's not me. It's not Kareem.
Simon Taylor · 38:15
And I appreciate that. I think one of the things I've observed is it tends to be companies that end up smaller that think that way, and it's also one where I've come from TradFi, where everybody's your frenemy. It's just it's just a thing that happens with scale. But I did notice you applied for an OCC trust charter, which potentially is somewhat stablecoin aligned. You guys have done a lot in the stablecoin space, maybe even a bit of a pioneer in that with financial accounts, talk to me about that charter and and kind of the the ambition for it. It's because it's not a traditional FDIC insured like we're going to be a bank that does lending looking charter to me. Sorry to cut across you, but
Eric Glyman · 38:54
yeah, it it is for a national trust bank charter, and given it's an active application, there's probably some limitations to to what I can get into, but I think that look one. I actually I think one of the things that this administration is doing that's very interesting is you know I think there's a recognition that you know 2008, 2009. All the chatter was, you know, there's too few banks, and the ones that are around are getting too big to fail.
Simon Taylor · 39:31
Yeah.
Eric Glyman · 39:33
And you flash forward 1718 years later, and there's fewer banks, and the banks that are around have gotten bigger,
Simon Taylor · 39:41
yeah,
Eric Glyman · 39:42
and I think that they've taken this idea very seriously. Of perhaps there should be more people who they want to be regulated and be audited and and go through all the work. That perhaps there should sure should be room for more. Types of competition that could be better for end consumers and businesses, and so one. I think it's a really interesting environment, and I personally believe it's it's actually structurally a very good thing. And when you're exactly right, the there are limited powers of a national trust bank, but I think one of the things that's very interesting that we see businesses looking for is in an agentic world, businesses often want support in doing what's right for them. For example, I think many businesses would love to have a bank or a partner that, well, if there's, they can kind of set. Here's how our investment strategies. Here's how we'd like funds to be stored. Could you just move things to higher yield on my behalf, and I'm going to give you authority to go and do that. Yeah, I think that
Simon Taylor · 41:03
ability to be in the flow of funds is is is one of the major benefits. I think the other thing I noticed about Ramp in particular is you were early to stablecoins. I think you've been somewhat of a pioneer in it. Talk to me about where that fits into your overall, you know, save time, save money thesis, and and why you guys are leaning
Eric Glyman · 41:22
it? It's um. I think the whole industry is so fascinating. I I think that they. I think the crypto industry, in some sense, was so taken with how beautiful the math was, that instead of explaining all the ways it could make your life better, often the pitch for Bitcoin is: Have you read the paper? The math is amazing. It's it's functionally better, and I think that I mean
Simon Taylor · 41:48
the math is amazing. It's beautiful. It is beautiful,
Eric Glyman · 41:51
and so we can make podcasts about that. But like I I think in some ways sort of lost the the plot of there were certain ways where the infrastructure. Maybe the pitch should never have been about. Here's how the infrastructure was designed. And an analogy is, it would be deeply weird if Ramp went out and said, you know, here's how Nacha works.
Simon Taylor · 42:15
Yeah, it's
Eric Glyman · 42:15
fantastic. Yeah, exactly. We can talk about this, but like here's an R03 error code and what it implies and what to do with that. And I think what attracted Ramp to it was beyond being great technology. There are so many types of payments where, you know, like a classic example of this is let's say it's Friday night, you have a vendor whose payroll maybe depends on funds getting there, and they're based abroad. In today's ecosystem, if you miss a certain cutoff, there is no way to get the funds there by Monday morning. Zero stablecoins, you can, and it can cost a lot less. And so, one of what's the products that's been growing extraordinarily quickly is this idea of like a stablecoin sandwich, where it's fiat on one end, fiat on the other, stablecoin in between, and you can actually move it extraordinarily quickly. You can do things for lower cost, and I think that there there are also problems too you can solve through just storage of stablecoins as well in countries that have very volatile currencies. We see businesses all the time that want their employees to have a card that, when it swipes in local currency, you know the the merchant gets the currency they expect, but the business doesn't want to hold that currency. Instead, you can hold it in USDC and we instantly go and convert it. You can swipe the card. We have this in a lot of regions in Latin America, and so I think it's just an extraordinary primitive to build on. I expect it will grow very quickly.
Simon Taylor · 43:49
I think for companies as well that are growth companies that are yours in the U.S. to be able to that have some LATAM interest that that can make a ton of sense for sure. You guys are private. My good friend Alex Johnson wrote a piece about perpetually private machines. He put Stripe in that category. Is there any pressure? Do you feel any pressure to one day go public? Is it an ambition, or is is there sort of so much room in front of you? You're not thinking about
Eric Glyman · 44:16
it. I, I mean, really, we're just trying to build a business that stands the test of time. That's durable. That outlives me, and you know, delivers extraordinary impact over many, many, many decades. And I think most of these businesses do go public. I'm not like. I think for some people it's a deep philosophical issue where they prefer being private. I don't have those views. I think one day Ramp will, you know, will very will very likely be a public company, and then you have kind of the rest of your life. And I think again back to Microsoft, like. You know they've been public for on the order of 50 years. I think 90% plus, 95% plus of all the value they've ever created has happened in like the past 40. Yeah. You know, maybe even the past decade. Actually, when you think about the appreciation, yeah, from like 300 billion to almost 3 trillion. Yeah. So yeah, last 10 years. As you were
Simon Taylor · 45:19
smart, Satya did a good job there for sure.
Eric Glyman · 45:21
He did a great job, but but my point is, look, I think there are a lot of interesting attributes of being public, as well as some pain that is worth considering. Right now, business is growing so fast; we're we're really just trying to focus on keeping up, and and you know, I still think as as far forward in many ways as Ramp is in adopting AI, we have a lot of work to do, and that requires a lot of change. And going through an event from private to public at some point maybe may be right, but it's it's it's an incredible amount of effort, and it's the question of what are you not doing? Yeah, it feels
Simon Taylor · 46:00
like a distraction, I can certainly imagine with the speed and the pace of change, life inside the singularity. But is there some point at which the AI labs and agents get so good at managing money, I don't need Ramp anymore, or do you think there's a world in which actually Ramp is the agent that moves my money? You know, like it just takes over more and more my finance ops. How do you think about that?
Eric Glyman · 46:23
I think all of these outcomes are possible. I really do, right? And and I think that one of the things we're trying to design on is is could we? I mean, do the thought experiment. Let's say an an agent with perfect information about you and all your competitors. The question is: Would that agent choose to work with you or not? And if the answer is no, in the long run, you might have a problem.
Simon Taylor · 46:55
Yeah.
Eric Glyman · 46:55
And so, when you think about like even the philosophy of like why why do we care about savings? Is I think it's a bit of a different game, right? If most organizations are engaged in, there's you know, the pie is so big and you can only slice it so many different ways. That is interchange. You know, you can't give more slices. You can't give more pie than there is pie, right? There's only so there's only so much there. And last year, the average organization, the median organization on on ramp, average higher, but the median organization on on ramp was able to cut their expenses, save about seven and a half percent.
Simon Taylor · 47:40
It's
Eric Glyman · 47:41
a lot larger than interchange, and so part of what we're focused on, whether it's in spend products, in storage of fund receipts, to even things we can do because we see parties on both sides of the transaction, is to be that most efficient method where even the most rational agent in the future chooses to use it. And I think it's important. I think agentic use on Ramp is growing. You know, we've had our MCP server up for over two years. I think most banks still haven't launched theirs. Maybe the modern ones are just getting around to it. I think it might be growing many weeks, 10% per week. It's growing extraordinarily quickly, and again, you really internalize like what that means. It's it's it's wild, and so part of what we're doing is designing a world where it might be that agents actually are running all the operations, and I think that was part of Cremano's background with Parabis was like agents doing stuff in 2015. So I think that is actually pretty reasonably likely outcome. I think there's another version where I think one of the things we experienced at Parabis that that sort of got us to want to go deeper to the metal and maybe get lumped into this this dead fintech category is it was one thing to see that a company that someone a consumer was wasting money. It was another thing to be able to do something about it, to be able to cancel the transaction, to go make the purchase, to go move the funds, to go store these things. And so, part of our work is centered around actually being close to the metal, or actually being at the operational layer where you can execute this, where you can translate a good idea into a great outcome.
Simon Taylor · 49:26
Yeah, the ability to move money for an agent is like the frontier at the moment. It seems like a lot of companies thought about it. Mercury does it in a limited way. A lot of banks are concerned about it. But there's a good study by I think AIVO that says just on the basis of like, if you ask an agent who should I use, you know, it's Ramp and Mercury and all those guys that are coming near the top, and the banks are way down the bottom. So, who's my agent going to choose, and can that agent move money? is is a pretty logical question. So, Eric, if we were to do this again in two years' time, what do you think is the same about Ramp, and what do you think's changed?
Eric Glyman · 50:01
I do think the mission is the same, and it's designed to be timeless. You know, I I think just as I imagine 100 years ago, I mean, what a different world we met her. You know, great grandparents, they were at the market. I think probably they would hope that you know you buy some groceries and you get an extra piece of bread or you know grains in the sack. I think that was true. It's true now. People always want a little bit more for less. I think that is certainly going to be true in two years.
I think it's going to be true in 100 years. And so, what I think will be utterly the same is Ramp is committed to help people and organizations get more for every dollar an hour, maybe token too. What's going to be different is I do. I think that the way that organizations operate is going to be different. I don't know if I'm all the way there, or there will be purely agentic corporations as sovereign economic actors, but it's possible. And I think that you're going to have organizations just as David and you were talking about.
We're part of the core inputs to how companies operate and make decisions. It's going to be agentic in nature in some sense. If I really listen to what kind of the leaders in that space, Jensen, you know, talks about, and again, I don't think he would describe himself as a fintech company or Nvidia as a fintech, but yet if you really listen, the future that he describes is you can kind of take a GPU, plug it in, pay for the electricity, and work is going to come out on the other end. In some sense, I think what he's describing is everything is going to become fintech. You know, all of software, all of work, and that actually one of the core inputs to many types of businesses, not all, but many, is actually can you go and convert dollars into productive work, and you know, and I hope that in two years' time we're successful in helping organizations continue to get more out of, you know, not just every dollar an hour, but token.
Simon Taylor · 52:29
I found it so interesting that I called the last episode we had the company equation, and you broke down the company as an equation. You just broke down. I think the future form of capital as an equation. Eric, I really, really enjoyed that. Thank you very much, my friend.
Eric Glyman · 52:43
Thanks, Simon. This is great.
Simon Taylor · 52:44
Thank you.
Eric Glyman · 52:44
It's fantastic.
