Aaron Levie: The Restless Founder Plotting the Future of Cloud Storage and AI

0:00I would probably start with the premise that most incumbents are awake, they're alive, they're not going to miss some kind of window that's happening right now. Even though I said that incumbents are kind of awake, alive and going to move quickly, 30 % of the companies will not fully capture the potential opportunity. Startups will find a clever wedge. And then, boom, you have the new startup emerge that can kind of move fast enough that it hits escape velocity. That's the voice of Aaron Levy, co-founder and CEO of Box. What started in a college dorm room became a cornerstone of the cloud era, and Levy helped shape it.

0:37Nearly 20 years in, with Box crossing a billion in revenue, he's not coasting. He's charging forward, reinventing the company around AI, rethinking everything. While others might defend the past, Aaron stays in the future, curious, restless, and moving fast, so the rest of us can too with him. This is Mike Maples Jr. of Floodgate, and it's go time with Aaron Levy. This is Mike Maples Jr. and welcome to the Pattern Breakers podcast, where we explore why some founders radically change the future and how they stand apart. Together, we'll learn about the counterintuitive mindsets and actions behind their remarkable success.

1:18Brace yourself for a world where chaos is welcome, naysayers are often a positive signal, and movements galvanize misfits who transform the impossible to the inevitable.

1:40In 2005, inside a college dorm room, Aaron Levy co-founded Box with a hunch that the way businesses stored and collaborated around information was about to change. At the time, cloud storage wasn't a given. It was a point of view about the future that few had yet embraced. Levy didn't just build for where the world was. He built for where it was going, and Box won big. But here's the twist. Founders often fall into patterns after they succeed. The urgency fades. But not with Levy. Now facing the AI frontier, he's not delegating Box's reinvention. He's leading it, personally, deep in the product details, like it's day one.

2:21What makes Aaron Levy fascinating isn't just what he saw back then. It's that he's still looking forward, still experimenting, still starting over, not because he needs to, but because that's how he's wired. It's not just ambition. It's a kind of enduring curiosity, a trait that has kept him in box forever young. Let's talk to him.

2:47Aaron, welcome to the podcast. Hey, Mike. Thanks for having me, man. You bet. I've been looking forward to this. Why don't we start from even before the beginning? I think it was 2005 when you started Box at USC. What was going on then and what got you to thinking you wanted to start this company? So before the beginning, so in 2004, I was a sophomore in college at USC and in high school had done a bunch of web internet startups, tried a bunch of different ideas. None really took off. Obviously, you're just kind of high school or building websites. And then in college, they kind of got a little bit more serious, a little bit more professional, but still nothing really worked.

3:30And I was doing an undergraduate business course where one of the projects was to essentially identify a market and do a SWOT analysis on some of the players in the market. And then simultaneously, I had an internship at Paramount Pictures, where my main job was basically like faxing documents and like, you know, scanning documents and just moving around data, you know. And so for some reason, you know, it's always hard to know, like, what was the very first spark that kind of comes together in one of these things. But for some reason, I kind of just latched on to this idea of internet storage would just make life a lot more efficient.

4:05And then in this class, I got to study the existing market of internet storage, which kind of was born out of the 90s. And most of the companies that had launched these online data storage products, they just stopped innovating also. And so kind of over a few weeks, maybe a month or two, you kind of build this sort of deep conviction and epiphany, which is, wait a second, like storage is getting cheaper. The Internet's getting faster and more people are going to be online. And yet everywhere I look, nobody has actually created a modern solution to this problem. Wait a second, like this could actually be a pretty cool startup opportunity.

4:37And so that was kind of the early ingredients that led to coming up with Box. Okay, but how did you act on that? Did you start writing the code yourself? Did you find co-conspirators pretty early? What was your minimum viable product to kind of test the idea? Yeah, so another kind of confluence of things, my background was in sort of front-end design. I was one of these script kiddies in high school. So you put together Perl scripts and you're automating things, but nothing that nothing that would have really been able to scale out as a file system that we were trying to build. So so I had a couple of friends, ultimately the co-founders, but they joined more officially about a year later that were helping on some of the technology and the ideation on the technology side.

5:24And then we ended up having some contract engineers go and build the first version of the product. I did all the kind of front end design and experience. And then we got that built out. We launched that as an initial, you know, now I would look at it like a prototype. It was fully functioning. It worked, you know, it wasn't like leaking your data. And we launched that. And we launched that in maybe January, February of 05. And basically, you know, relative to other startup ideas that I'd had, this is the first one where just like regular people on the internet were like happily using the product.

5:59and giving us feedback. And I was like, oh, wow, this is actually what people talk about when they talk about like doing a startup. Like all the other ones have been a website, like the traffic kind of comes and then they leave and no one ever sticks around to do anything on your product. It kind of sucks. But this was the first one where like somebody signed up, they like uploaded their files, their shared files. They would email us saying, hey, I wish you had this feature. We had some early success with some of like the very primitive bloggers you had like Engadget and Gizmodo. And, you know, these were like kind of the first internet blogs, they'd review technology.

6:32So we had we had some reviews on those sites, we did some giveaway promotions, we had, you know, different kind of referral things. And, you know, each day, you know, maybe it started with with three signups, and then five signups and 10 signups. But each day, it was kind of growing, nothing that felt like an overnight success. And there's an important asterisk, we launched as a paid only service, mostly because we just didn't, you know, think there was another way you did software, like, okay, you have a software application, you're going to charge for it. And basically, you know, just to just to show the kind of, you know, complexity or tension here, I had applied for a couple internships that spring for the summer and I got rejected from the internships.

7:09If I had gotten them, we could have been on a very different timeline because what happened was by getting rejected, you know, it meant that my co-founder and I, we could spend the entire summer working on the on Box the entire summer. So we built basically Box out of first our kind of dorm college, you know, apartments, and then out of my co-founder's house in Seattle, you know, 15 hours a day, you know, growth hacking, building more features, launching new things. That's when it really started to feel like a startup. And then ultimately, you know, actually like going out and pitching investors that kind of made it even more real.

7:44When did you decide to drop out of USC and pursue it full time? Like, when did you decide, yeah, this is a thing for sure. So in the summer, we pitched every VC we could find in Seattle. I found Bill Gates's fax number. We faxed him a prospectus. I think we got the equivalent of a 404 from a fax machine. So that was automatically rejected. We dropped off a printed prospectus at Paul Allen's house, mostly rejected by everybody except for a couple. We found this group of angel investors in Seattle that were willing to take a shot on us. So we raised about$80 ,000 and we sold a quarter of the company for$80 ,000 that summer.

8:25And then later in the summer, because I guess we were just like drunk on needing dilution, we also pitched Mark Cuban and he then decided to invest in us later that fall. So then we took on a few hundred thousand dollars, which again was another kind of 20, 25 % of dilution. That gave us increased conviction that, wow, like if Mark Cuban is saying this is a real business. That's pretty serious. We should take that more seriously. And then what was happening is throughout the fall, it was very distinctly obvious that it was almost impossible to keep doing classwork and run the business. I'd be in the middle of classes that I just knew I was going to fail out of and just answering customer support email inquiries on a BlackBerry.

9:08So that kind of ramped up through the fall and that made us realize that we had to make a decision at some point. And how did you get Mark Cuban's attention? Mark's incredible because to this day, he's just like the most online person probably on the planet. If you found Mark Cuban's email address in 2005 and you emailed him, I actually think there's like a high degree of likelihood that he would have responded. I am utterly amazed at his email responsiveness. Especially because I think now it's like a known thing that he does it. And you'd think that at some point, if everybody knew that you were so good at email, you would have to be so inundated by email 20 years into that.

9:46And yet to this day, he's just still incredible at triaging it all. So we sent him an email and we were just like, hi, we're two kids that are building this online storage product. Any interest in blogging about it or partnering with us on anything you're doing? And then that kind of snowballed into an investment case on his end. But all of that kind of reached this crescendo in, I believe it was Thanksgiving break of 05, my co-founder and I, we were just like, you know, I think this is reaching a point where we have to kind of, we have maybe a once in a lifetime opportunity to jump on this. And we got lucky because the zeitgeist was starting to emerge.

10:24Zuck and team had dropped out, moved to Palo Alto. YouTube was these ex-PayPal younger people building this thing. You know, this is probably our shot. Like We have a website. It's working. It's scaling. So we dropped out that winter break and moved to the Bay Area. In the early innings, did you ever have any near-death experiences where it just wasn't going your way at all? Or was it sort of gradual up into the right for the most part? I think everything was gradual up into the right so far in the story. The near-death starts to happen over the next couple of years where we had kind of more multiple near-death experiences.

11:04So maybe just to accelerate, you know, we drop out. We had two other friends, the ones I mentioned earlier, also drop out a few months later. And this freemium thing is this, you know, real opportunity, which is if we had a free version of the product, more people would sign up. We could then upsell a percentage of them. So we launched, you know, Box as a freemium product. And that starts to take off where now we're seeing like thousands of signups a day. We raised a Series A. We'd probably be raising like$300 million if it was back then. And our Series A was 5 million posts. So I'm very jealous of all founders right now.

11:36So we raised this$5 million post from DFJ, Josh Stein, and Emily Milton. And we kind of, at this point, were a little bit in wandering mode because what was happening was we were kind of a consumer slash prosumer product. So we saw the writing on the wall that Google had to enter the game, Apple had to enter the game, Microsoft, Yahoo, Apple, I mean, sorry, Facebook, et cetera. So that kind of scared the living shit out of us because we were like, okay, you've got these massive incumbents. All are going to pick up on this idea that you could store files in the cloud and make them available for people.

12:11And all we're getting is we're collecting$2.99 at a time. It's kind of hard to make math work. And then at the same time, we had another kind of part of the customer cohort, which was these businesses or teams or professionals where we were still charging them$2.99. And maybe we had a pro version like$7.99. sense. But you could see that there was a lot more functionality they needed, and they'd be willing to pay a lot more if they had that functionality. And so we were in this kind of wilderness period in 06 to early 07, which was we might have to kind of decide one path. Are we going to do this consumer thing head on with the big tech companies?

12:47Or do you go after this enterprise market? Are you trying to split the baby? The team was like, I think we have to do enterprise. We have to be a B2B product. And I was actually the first to resist it because there was nothing cool about doing enterprise software in 07 or 06. And so I resisted it, but eventually the conviction built that, okay, we probably can't survive as a consumer company. So we did a relatively hard pivot. We said, we're going to go enterprise. And where it gets a little bit more near death through that pivot, unfortunately, the economic recession kicked in. So that was a really complicated exercise.

13:18But we were lucky that we got Mamoun from now Kleiner back then, USVP led our Series B. And he kind of went head to head in some respects with the partnership and said, I want to bet on this team. And we looked nothing like what they were used to, a freemium viral disruption in the enterprise led by 20-year-olds. So that was a near-death experience because we could have gone out of business simply because we were in a business model transition, bleeding a lot of cash because of our subsidization on the freemium side. And then we had a couple of similar experiences later on as we scaled. Even the Series C, we were not yet fully crushing it on the enterprise front.

13:59And Royo Driscoll kind of really bailed us out on that one from scale. So we've had a few of these near-death experiences, usually related to the business model still needing a little bit more time to bake, running low on capital because we're grinding out the growth, which is why at this point I always instruct founders to be very thoughtful about burn, cash flow. But it's still getting the engine going was certainly hard back then. And what is it that caused enterprises to buy? Was there one key reason or one key use case where you were just 10x better than anybody else and that gave you the momentum?

14:38The killer app for us was secure file sharing collaboration in an enterprise context, which is Sally, Bob, Timmy, whatever. You can just sign up on Box. It was Box.net and then became Box.com. You can just sign up for an account, start using it, have a secure way to share and collaborate. And then we'd be able to go and talk to the IT department about, hey, there is a better way than using these on-prem legacy systems. Can we introduce you to Box? But we had this viral engine that was going and kind of getting us in the conversation. And then there were a couple moments that sort of added more inflection to that.

15:12So the iPhone, while it was launched when we were kind of more on the consumer side, it started taking hold in the enterprise. So all of a sudden you had mobile devices that couldn't tap into the on-prem systems. And then iPad was just one more version of that, which is you would have like an executive in a boardroom pull up an iPad and they'd be like, how do I get the board deck? And the company was like, oh, yeah, that's not going to work with our on-premises, you know, document management software. So they needed Box to go in and actually enable that experience. So we kind of had, you know, a bunch of these big technological shift tailwinds that we could ride where we were only a piece of that tailwind, but that tailwind was causing an architecture change in the enterprise that we could ride alongside, which you could easily extrapolate that as an important lesson for any startup and especially an AI as an example of like making sure you're on the right side of those tailwinds.

16:01And then maybe the only other final point was we were just simply cheaper, faster, easier than the incumbent. And so we could come in as this disruptive, you know, kind of easy land and expand product where it's all upside for us, all downside for the incumbent. And that let us sort of have this, you know, nice business model, you know, asymmetric advantage. Were some subset of the IT folks just hostile to you? Were they just like, get out of here? I don't even want your stuff in my company. I view you as the problem, not the answer. And where some just, where have you been all my life? This is amazing.

16:36I'm just curious, the spectrum of those early people, how many of them were early believers? How many of them were neutral to hostile even? One of the reasons why I have so much conviction on any kind of freemium or commercial open source type model in an enterprise context is it lets customers kind of select in to your funnel as opposed to you. You know, if you just imagine two completely different parallel universes in one universe, all you have is a bunch of, you know, account names and you have to go call each one and figure out where they are in their, you know, in their journey and, you know, how much do they like the cloud or whatever.

17:15And now alternative universe number two, you have a freemium product and all of a sudden you just see a heat map of like, oh, this company is taking off, this company is taking off that company is. Obviously, the signal to noise ratio is 100 times better in that latter scenario. We, by default, had the benefit of we could target and go talk to companies that were already using the product. That meant they were already a little bit bought in. Now, some said, okay, I don't want my users doing this. And so we actually are uncomfortable about you guys. But we tried to be very serious and professional and cognizant of the natural organ rejection that would happen in an enterprise context.

17:52So we were not a, you know, it's our way or the highway type company. It was very much like, hey, we'd love to hear what features would you want? You know, what enterprise products and capabilities can we build? You know, you're almost in this flow state between sales and product and go to market when it's working because you're not just like handing a sheet of paper to the salesperson and saying, just sell this is all you can sell. You're giving that salesperson the ability to kind of bring us feedback and we're going to go and build that. And so and that that became this really nice flywheel.

18:20And so it's funny, though, because, you know, I have to I have to kind of, you know, call this out as a blind spot. Like, you know, I think a lot of founders have to do that all themselves. And this is where you have a lot of these founder led sales kind of dynamics. I actually didn't. But but but only by virtue of the freemium model was generating leads. And I kind of saw my saw my role as as being the kind of, you know, arbiter of, you know, what are what are what's the sales team hearing from the customer? What are we building on the product? How do we get that flywheel going? So I certainly talked to customers, but really the sales was really all happening from a classic sales team.

18:53Yeah, to me, the sales learning curve is really interesting as a notion. One of the reasons I find it so interesting is I think it's easy for startups to get fooled about whether they really have product market fit. And so I've got, you know, everybody's conspiring to tell the founder that he or she has it. because the seed investors want to raise an A round, so they want to pretend that you have it. The entrepreneur wants to believe that they're making progress, and so you're looking for any positive signs of life that you can find. One of the things that I came to realize was that progress down the sales learning curve is a very good heuristic for product market fit, because you can ask yourself, when I add an incremental salesperson, can they pay for their entire burden cost across the entire company and how soon can they do that, right?

19:42If you can do that, you can hire salespeople at will, which I'm going to assert means you have product market fit and there's no debate. If you can't do that, then the question becomes, okay, where can you and where can't you? That's the friction between you and strong product market fit. But it's a way of saying, hey, look, I'm not being a distant critic here. It's just there's a set of salespeople and they're at a certain level of progress down the learning curve. And we can assess, right, if we're a one out of 10 or 10 out of 10 based on that. Yeah, that's super interesting. That brings to mind kind of an adjacent thing that Rory, when he joined the board, he had this mantra of just like hire a sales rep every month, like clockwork.

20:24That'll show you what's breaking in the business and or what's working. By the time you have six sales reps, if you have leads for four of them, it's going to cause the marketing team to kind of have to step up. Obviously, you get to a point where like, if none of them are performing, then you don't have product market fit. But it's kind of causing this ability in the business to test what's the upper bound versus I've seen a lot of startups with product market fit hire too slowly, because they're trying to kind of like hyper optimize, you know, each reps productivity rate or whatever. And it's just like, no, no, just like keep hiring sales reps, let the business, you know, kind of grow into, the amount of capacity you have by forcing these other parts of the business to step up.

21:04I am convinced there's been no important company that's gone out of business because they hired too much go-to-market capacity. But there's been lots of companies that could have won their market by having more go-to-market capacity. So I think it's more fatal to have a window of opportunity and you sort of didn't build that rhythm fast enough than, okay, I built it up, the market It wasn't there, you know, because that means you wouldn't have worked anyway. So, like, who cares? You know, what was it like as the company started doing better and, you know, you started to become a real company? You're not a startup anymore.

21:36It's no longer zero to one MacGyver juking and jiving, going from a jazz band to sheet music, you know, where you're trying to hire execution-oriented people and trying to manage that. What was it like from your perspective? Because it probably felt pretty different than the early days. Yeah, although that's a great analogy. But to be totally fair, I think we've probably only done the sheet music thing for like a year in any stable way. So that's about as long as it lasts in this industry. But there was a period where from 2010 to let's say 14, maybe it's a nice period, which was we were doubling or tripling each year.

22:17that was causing us to put more fuel on the fire, relatively capital inefficient. But what we had conviction of is that we were acquiring these long-term value kind of assets in the form of customer cohorts that would pay off within two years, let's say, of the acquisition. And so you just want as many of those as possible. And so we did that about until maybe 200 million in revenue. And then obviously just large numbers kind of kick in. You're scaling a little bit more slowly. But I'd say that was the sheet music period. And then we sort of had to figure out that, okay, we can't just be a one-act company.

22:52We're going to have to sell more into the customer base. We have to build out more platform capabilities. We have to have a whole sort of series of security and governance and data protection solutions. So that was a journey where we had to then go from a single product to a multi-product act. That's kind of like back to, let's say, jazz band again. And then you kind of get your sheet music for a couple of years. but then boom, COVID happens. Now you're back to kind of figuring out what's changed, what's new. So I'd say, you know, there's, at best, I've been able to have a calm demeanor for like six to 12 months in this business.

23:26And speaking of, you know, now we've got AI. So, you know, we're back in, we're back in, we're back in jazz band. Yeah. And you've talked about AI's potential for years, but more recently, you know, I've seen you talk about it even more. I'm just, you know, could you walk us through how you're thinking about AI's impact on content management collaboration and how's it evolved from Box's early days? Yeah. So for us, this is almost a complete, you know, sort of reinvention of the company. And if Box feels more today, like it did 17 years ago than any other point in the history, in the sense of the just sheer amount of change that's happening in the business and outside the business and the kind of clear feeling that you have to jump on it with actually the added stress slash excitement of the speed of innovation happening from the AI model providers is insane.

24:25The change of behaviors is happening faster than I've ever seen. The wired in and connectedness of everybody participating in it is just incredible because we're all in the same platform talking to each other. But it very much feels like, to me, we are in hardcore startup mode, reinventing the business because of AI. And the reason why it's elevated to that is in our business and industry, we store unstructured data, financial documents, contracts, invoices, marketing assets, resumes, project plans. All of that data is exactly the data that was relatively opaque to your computer before AI. You created a file, you stored it somewhere, maybe you'd remember to go back to it, maybe you'd find it at like, you could find the file name and search for it or some words in the document.