Oana Jinga, Co-founder of Dexory: Dexory’s robots save warehouses millions

0:00Hello and welcome back to the Scaling Europe show. I'm Seb Johnson. Today I'm joined by Oana Jinga. Oana is one of the co-founders of Dexery. Dexery is one of the coolest UK deep tech robotics companies that we have here. Raised hundreds of millions of dollars building some fantastic bits of machinery. How are you doing today? Thank you for joining me.

0:16Oana Jinga:Thank you so much for having me. And like, yeah, now I have to live up to that intro. Yeah, well, I'm sure you will. For those who don't know, can you talk a bit about the robots that you're building? They're quite unusual in a few regards. So maybe just explain a bit about what you've built and what the robots are out there doing. Sure. So at Dexory, what we say is we provide a data intelligence platform for logistics. But the way we do that is we have these realtor robots that are 18 meters in height that go around warehouses and digitize that space, pretty much creating what we call a digital twin of the warehouse.

0:49Oana Jinga:And then the second part of the product is actually that twin where our users actually interact with the data. And there's a lot of really, really cool features on what they could do with that data afterwards. So it all starts from the robotics side. And we did need something that can capture a lot of information at speed from these spaces that I've seen people that are not close to warehousing and logistics probably don't realize how big warehouses are. You're talking like millions of square feet, right? So being able to get that ground truth and eyes on the ground, it's usually very, very hard because it takes a lot of resources and people to capture anywhere close of that information.

1:21Oana Jinga:So the robots are there to do that data capture, but then the magic happens in what we do with the data afterwards. And can you talk a bit about what the real value is to the companies that you work with? I know you've got some amazing clients, but to try and, I guess, ground it in something that other people who are, I guess, not so close to the warehouse world can find a bit more tangible. tool? Sure, it starts with that visibility element. So because we capture so much information, we're pretty much giving them the exact idea of what they have in the warehouse space-wise, goods-wise, like machinery-wise, exactly what you would call like a digital representation of that reality on a continuous basis.

1:57Oana Jinga:So the robots are scanning 24-7, they keep updating this digital to what I was talking about, to have exactly what is in a warehouse at a particular time. And then having that knowledge, then they could do a lot more on the back of that, but also ensuring that none of the processes that they normally have can be derailed because something is not where they're expecting it at the right time in the right quantity to fulfill an order. It's a very basic term with kind of our eyes inside the site. And because that site is, again, very, very big at scale, this is something that never, never existed before.

2:28Oana Jinga:And we are very unique to say, actually, that we're the only ones in the world with this technology and hopefully continue to be to be there and be number one for a while. yeah amazing so i guess it's linked to sort of supply chain optimization it's giving your customers the eyes the information the data so that they know exactly what they have and where across all their warehouses so that they can then sell move ship things in the most efficient way is that right exactly yeah you put it very very very very well so nice can you talk about maybe some of the clients that you have and how they're using it some of the value that they've seen so we work with um a lot of the big third-party logistics providers like your dhl your maersk db shanker like anyone that you see on the back of a truck is you're driving up and down the motorway basically um but also with um what we call end brands so that could be manufacturers like a lot of car manufacturers retailers pharma companies um so anyone that has warehouses and has some stock in that warehouse um obviously would be a very good customer for us um and i mean it goes from like from being particular numbers like we have customers that have been using us for a couple of years now um and on the back of the information we're giving them they managed to pretty much save literally tens of millions of pounds um and initially when you work with your typical kind of corporate especially from a traditional sector you put together a business case with them and they're being very conservative and their cfo kind of cuts half of the value that they think you're going to bring uh then you kind of come in and the fact that like yeah 12 24 months after they actually go back to this like okay you know what you're absolutely right like we should have like expected all this value i think it's great great kind of testament that we are building something quite unique um but the biggest thing for me is like these big players it's not that just they're working with us you know one warehouse here one warehouse there is like they're they're expanding with us at scale and they're actually taking us into new markets we're now live in about 16 different countries all organically driven by by some of these big customers we have which again is testament that we must be doing something right because they just want more and more of it So, yeah, you're talking about really big cost savings, but also massive opportunity for them to get more revenue in because we help them optimize their space so they can bring more stock in.

4:38Oana Jinga:We make sure that they could get orders out the door in time and even like more orders that they would normally do. So all of that, again, taps into the revenue side as well. That's really interesting. So that sort of market expansion side is being led by your current customers who want to bring you into their spaces, whether it's, I don't know. I know you're very big in North America right now, but what other markets are you in? Is it Asia? Yeah, we're live in Australia, New Zealand, South East Asia, Middle East, all across Europe, I would say. And then, yeah, the US, but also Canada and Mexico, so North America all together.

5:13Oana Jinga:And it's exactly as you said, and actually I was talking to somebody the other day, like what's the easiest way to expand? I mean, obviously, usually people ask from Europe into the US and I keep telling them, like find a customer that could take you with them. And to me, that's kind of the best way to enter any new market, because then you will have a very strong case study. You'd have an advocate there rather than you putting a team on the ground and working obviously like from the ground up. Already starting with a local player is fantastic to have. How challenging is it given, I guess, like the physical nature of some of the work that you do?

5:46You've got to get these 18 meter robots into their factories. So if you don't have a local presence there, what does that entail? Yeah, and it's a great question because that was kind of one of the first things that we considered when we launched the product.

6:00Oana Jinga:Like it has to be something we could get anywhere very, very quickly and it can install very easily. We don't actually need anything infrastructure-wise from our customers. We don't have to make changes to the warehouses or any of that. We just need a power plug, which obviously everyone has. So we wanted to make it as seamless as possible to get in and install. So the robots actually, they divide into two normal boxes that just go in air cargo or on a ship. And the moment they get to the customer side, we do actually send two of our deployment engineers, but you don't need more than two people.

6:29Oana Jinga:And within two to five day maximum, they get everything up and running on site and the customer already starts getting value. So it's a very, very quick process versus a lot of like your typical automation processes that take months and sometimes even years to obviously put together. So we were very intentional, like it's got to be very easy to install, very easy to ship anywhere in the world. and then the customer should have that value very, very quickly, which then obviously kind of triggers the conversation. Okay, we want more of this. Like, where else can you go, right? So, but you have to be very intentional about it because otherwise it can become a massive pain as you scale and you realize, okay, now we have to do a thousand of these.

7:03Oana Jinga:Shoot, that's going to be complicated, right? Yeah. What type of ongoing support or maintenance support do you provide? Is that something that you have to, you know, every few months you're flying people out to maintain, check on them? Is that like as they need it? How does that model work? Yeah, it's a bit of both. I mean, we do have regular check-ins on site like every six months or so. Again, going back to that intentionality, we wanted to make sure that the robots are as robust as possible and also like overly engineered so that the more kind of features you put onto the software side, you have the hardware, so you don't have to always kind of go and upgrade and change and all of that stuff.

7:38Oana Jinga:So like the future proof has to come from the beginning, which means that we don't really have to touch them or do anything to them for pretty long times. And unless something, I mean, like, obviously, they run 24-7 continuously. There's, like, miles and miles and miles of running every single day in these warehouses. So you would have some wear and tear sometimes, like, I don't know, something happens to a wheel or it starts collecting, you know, plastic from the wrappings around pallets or stuff like that. So in that case, I would do what we call predictive maintenance. Like, there's a lot of sensors that would tell you on the robot something is wrong here.

8:08Oana Jinga:Remotely, we can kind of see what that is. If we can't deal with it remotely, then an engineer would go inside. But that's quite rare. It's usually like, yeah, every six months, somebody would go in, check everything's fine. Yeah, change the oil on the car, things like that, and then pretty much continue running. And so how was this process managed before? Like how are other companies doing who are not using your robots? Yeah, so our biggest competitor is pen and paper. Because to collect data in any shape or form, people normally start with inventory data. So they need to know, like, is the right pilot in the right location?

8:42Oana Jinga:Do I have the right quantity of items as my systems are telling me to do? And all of our customers usually have a team that does that manually, going location by location every day, counting boxes going up and down or scissor lifts that are shaky and wobbly and everything else. But of course, humans can only go as far as like you count about like number of boxes. You put that on a piece of paper. You cannot take high resolution images of these things like yet. Maybe we're going to get there or you blink and then that kind of gets recorded. and also yeah you can only do like I don't know even like the best kind of person will probably do a couple of look a hundred locations an hour like we do about 10 to 12 000 it takes us a few hours to do a full wall to wall check of a site so it's first of all I get the amount of data we collect but also the quantity of it it's not something you can replicate right now.

9:30And so I guess these warehouses were built initially for that manual human-led process do you see a future that you know once we see more robots like yourselves doing the data collection but also more the automation moving around do you see warehouses fundamentally changing the way that they're laid out or structured to now be optimized for this world where it's not humans with pen and paper it's actually you know entirely robotic it's a great question and i think you have two

9:58Oana Jinga:sides of it right like there are i mean over i don't know um like 250 000 warehouses worldwide or something like that based on the latest numbers, which are built. And like a lot of them have been kind of built over the past, let's say, 10 years that have tenants that are in there for like 20 to 30 years. So within the next decade, at least, there has to be more ways. Like any kind of robotic kind of comes in and operates within this kind of brownfield environment. Because if you kind of come in and you ask a customer to build around you, that's a very big investment, a big commitment. And we're seeing a lot of like the technologies that do that actually kind of drop a little bit in adoption because of that massive commitment up front.

10:38Oana Jinga:So you kind of have this timeframe where you will kind of have to operate within whatever exists in order to be able to scale very fast. Of course, like otherwise, you could kind of like do one by one and slowly kind of take it. So I don't think that's going to change significantly in the short time. But also like I hear a lot, and obviously like I'm surrounded by people talking robotics, right? And obviously humanoids is a very big factor now. And everyone's like, oh, the world has been designed for humans. And I always counter that. like no the warehouse has been designed for the forklift um it's very very different the human is inside the forklift but the whole layout like the racking the space that you have in between everything has been designed for a machine uh and even like the floor is i mean yeah you might have a few bits and bobs and like a like a little pothole here and there right but it is pretty like even and straight and and uh quite quite um nicely maintained because these forklifts have to move a lot of goods so the warehouse itself i don't think it's going to change significantly in terms of that structure and actually robotics fits very well within the way it's it is right now because of that kind of design for the for the forklift side if you go into manufacturing lines and production lines of course it's a little bit different um but um in the warehouse space i don't foresee massive massive changes oh that's super interesting okay yeah it makes a lot of sense because yeah yeah the forklift point is yeah that that's what that's how people operate and that's people get around right and so it's easy to apply a robotic player to that then try and exactly exactly like you would you put a humanoid in a forklift might as well just put an autonomous forklift there right exactly exactly yeah yeah that makes a lot more sense um and i guess like you you must be cutting millions and billions of data points right and we're in this moment now where everyone's talking about the next stage of ai's physical ai we're talking about world models we're talking about you know really ai progressing from llms to something a bit more robust and three-dimensional what role do you see you know Dexry playing in that you know you you must have this insane data layer do you use that just to kind of optimize your own you know data robotics or do you see well well actually we've got this amazing data can we build a model with it can we do something exciting with it I don't know I'm just curious that's probably the most exciting thing about what we do for sure and and especially over the past I would say like six months to a year when we realize not how much data we have, but also with all the other existing tools out there, like what we could do with it.

13:03Oana Jinga:So, I mean, to give you some stats, we were probably like about 1.3 billion pallet locations scanned right now. And that's just a unit to kind of give a bit of an understanding of how much you're capturing. We're talking about 70 ,000 days of warehouse operations that we've kind of captured continuously, hundreds of sites around the world. And what's interesting in that is you kind of have different geographies, different type of product, different seasonalities. So there's so much in there that actually gets us as close as possible to actually having a proper real model of like how warehouses look and act on a continuous basis.

13:37Oana Jinga:Right. So none of this is synthetic data. This is the actual real data from inside the site. And then from here, of course, you can extrapolate a lot of it and start building synthetic models, all of it, because, again, it's all based on this kind of ground reality that we've been capturing. and we're continuing to capture. So our kind of, say our vision from even earlier days, like we started as a funding team, like over a decade ago, working on robots that would capture the world. Initially it was going to be in the house and then in retail stores. Like we had a lot of pivots until we got to the warehousing side, which was after the pandemic.

14:08Oana Jinga:But we always kind of had this idea that we could use robots to kind of digitize spaces. And it's exactly what we're doing right now. And all this data that we have is machine ready data, which then enables us actually to pass over to other machines, like we were saying earlier. It could be autonomous forklift. It could be picking robots. It could be humanoids at one point. And they could use that raw data almost instantly to actually understand the world themselves and know where to go, what to do, and how to operate, how to react. And that data set is 100 % unique. There's no one else that kind of has that in the world.

14:37Oana Jinga:So going back to your question, 1 ,000%, that's kind of the biggest focus that we have right now is kind of utilizing that data to turn it into something that we can make ourselves better and I'll see our customers as well and be able to be a lot more predictive and simulate scenarios for them in like very simple business value for like warehouse operations. But also when it goes to integrations with other pieces of technology, I mean, we're very much at that kind of first stage of physically out. It's like, here's the data. And then the actors that act on that data, it could be a very kind of wide type.

15:07Oana Jinga:It could even like be people in the first stage, right? Because you're kind of guiding them and orchestrating them to go to locations. But then of course, as you kind of go into other autonomous actors, as we call them, that gets even more exciting. And can you touch on the funding round that you did in the last year? I think it was your Series C. I think it was 100 million initially, and then it was a bit of debt as well. And then I saw the British Business Bank came in a few months ago. Can you talk a bit about what unlocked that round? Because it was a big round. And yeah, I'd be interested to hear how that fundraiser went.

15:39What are you going to do with it and what unlocked it?

15:41Oana Jinga:Yeah, it's interesting because, as you said, like it was last year, which in a startup's mind is like that was a decade ago. It feels like it's been forever. But we've been very fortunate to have some fantastic investors backing us and also partnering with us. Like we raised our kind of C, Series A, B and C, like pretty much like on a rolling basis. And the Series C was actually preempted. And it came at a fantastic time because, I mean, we got the term shit and the conversations were happening before the whole hype kind of started to take off. with physical AI. So we're one of the first to actually jump on that.

16:13Oana Jinga:And as you said, it's like 100 million equity. And then with the debt and everything goes to 165, which I mean, at the time was a pretty substantial amount for robotics in Europe. Since then, obviously we've got Neura that kind of crashed everyone out of the boat globally sneaking, but I'm not going to get there. So I think, yeah, going back to your question, it's been a great fuel. I mean, the number one thing we wanted to do is really kind of scale up the commercial side of the business. We were operating pretty lean on that and a lot of stuff is kind of coming to us because of the demand and the brand that we had.

16:43Oana Jinga:But now we have a very, very solid global go-to-market function. It's enabling us to go kind of 3x year on year continuously, which I mean, for a full stack company, it is quite, quite aggressive. And also, I think the one thing I always talk about, because there's so much noise now around like, we're 100 million, we're 200 million, 500 million, like all of that stuff. What I'm really proud about our revenue is that it is actually long term, kind of three to five year contracts with like these massive corporates, that it's a very kind of healthy revenue from our perspective. Like it's not a monthly kind of subscription.

17:15Oana Jinga:It's nothing like that. It's like a very, very long term relationship that enables us to be kind of quite solid on the future proofing of what we're doing. So if anything bursts in the bubble and so on, I mean, that's a very kind of solid foundation to build. And our customers are saying they're organically kind of expanding and growing, which again is all based on that solid relationship that we have with them. And I imagine it's very sticky, right? they you know even if even if you agree a three or five-year contract you know it's a lot easier to become dependent on a robot and data than it is you know a SaaS tool or something like that.

17:46Oana Jinga:Exactly you have to make some significant changes to revert to something else so then once you're in and everyone kind of really adopts that tool it becomes ingrained into the day-to-day workflows and I mean everyone so far like we launched about three years ago so we're having our first kind of renewal conversations of course and everyone's kind of renewed again for three years or more So I think, again, early days, and I mean, yeah, it's still young in the grand scheme of things. We're probably ancient, considering some of the newer things that are coming to market. But yeah, it's all the indicators are definitely there that we're on the right track.

18:19And given how hot, you know, world models are, robotics are, this deep tech, you know, this move towards robotics and hardware again, especially since you've raised your series C, are you getting more investor interest? Are you getting more inbound? Are you thinking about trying to capitalize again while the market's hot?

18:36Oana Jinga:I mean, never say never, right? I think, yeah, there's a lot of conversations going, but obviously we want to do it with the right partner and for the right things. There's so much more we could be doing at a faster pace. we have a very solid plan for like the next five years as everyone does but like if we could execute that in two and a half rather than five I mean that's always kind of been our ambition and it's quite interesting when you do talk to investors especially European investors and you kind of come in with that some people panic and you're like yeah but we're talking to the guys in the US last week and they're like that's actually quite slow right so it's like sometimes you have to do a bit of that education around like why you should go so aggressively especially when we're like comfortable obviously having the series C still in the bank and kind of giving us enough cushion to just continue like at the pace that we're at um but i think yeah it's if the right partner comes in with the the ambition that matches ours i think yeah we would yeah probably kind of consider that amazing and look last question looking forward for the next six or 12 months you've obviously got a lot of ambition what would you like to achieve where would you like to get to in the next six or 12 months to make you feel like okay it's been a good year yeah i think i mean of course the revenue side is a revenue side and we have some some super ambitious targets for the year.

19:45Oana Jinga:And I think the team is really pumped because they're now seeing things kind of coming through and them kind of obviously reaching their quotas and smashing that and so on. So that energy is just fantastic to see in the business. But for me, some of the stuff we were talking earlier about is like, what else can we do with this data? And we're barely kind of scratching the surface from a customer perspective as well, because we just launched what we call Dex Review Adapt, which is kind of our agentic play for Dexery with our customers. and they're using it to solve so many new use cases and going to so many directions.

20:13Oana Jinga:I'm like, okay, what else can we do? Because that then allogs more value and you could push the ACV higher and all that stuff, right? It kind of comes back. So I think for me, it's that product angle is the exciting thing. And then also, as we're discussing, how we integrate this with other players in our space is very unique because you do have a lot of people that are building models based on video feeds or like synthetic data and so on. But a lot of them kind of come to the table without actually being robotics providers themselves. Whereas like we were robotics providers first. So we understand how robots work, what they need, how they adapt to this reality, which kind of puts us in a very unique place because obviously that machine ready data is very different to the synthetic side of the data.

20:55Oana Jinga:So that angle is also kind of very exciting to us. And I think we're going to be in a very different place with it by the end of the year. Amazing. Well, it sounds exciting. It seems like you and the team are crushing it. And it's great to see, you know, like a robotics company coming out of the UK, scaling globally and doing cool things. So thank you very much for joining me. And let's catch up again at some point in the future when there's news to share, milestones to celebrate or something else exciting going on. Thanks so much, Seb.