Sai Shivareddy, Founder and CEO at Nyobolt: Building the critical power infrastructure for the AI world

0:00Hello, welcome back to the Scaling Europe show. I'm Sir Joltson. Today we have got on the show a very exciting guest, one of the UK's newest unicorn founders, Sai, co-founder and CEO of NyaBolt, the new deep tech unicorn coming out of Cambridge. Thank you for joining me. How are you doing today?

0:15Sai Shivareddy:Thank you for having me, Seb. It's great to be here. You guys just raised a$60 million Series C at a$1 billion valuation. Amazing news. How does it feel now that the news is out there and you've been, yeah, you've got it out? well it's another day in our growth uh story so but it's great to celebrate a milestone i'll tell you that um it's um not every day you have big announcements to make so um great great uh to be at this point absolutely amazing and for those who are not so familiar with the work that you do the company that you've built can you give a quick overview into to what naibolt is and what you do Yeah, so Nibold builds fast energy storage systems that enable dynamic power capability for the emerging AI infrastructure requirements.

1:07Sai Shivareddy:So everything from mobile robotics to data centers, AI data centers specifically, have this unique requirement in terms of very high power density needs where traditional energy storage systems do not meet the requirements. So things like batteries and capacitors don't do it. You've heard different flavors of capacitors, like super, null, try and whatever else you can, but they fall short of what is needed in terms of density. So we have the highest density that matters when it comes to power and we give you the fastest response. So what that means is extremely fast recharging times, factor of 10 over anything else out there, sometimes factor of 100, depending on what you're mentioning against.

1:48Sai Shivareddy:So we've been busy building Nibold Nibold through the discovery of a new material that allowed us to do this, but then we've got many inventions across the material space, the entire tech stack that is the hardware that the materials are embodied inside, and then there's the software stack that plucks and plays into any working environment. So we have a full stack solution to meet the emerging problems of uptime, as well as critical productivity metrics that matter in robotics and AI workloads. So I think it's number of coefficients per GPU per second, right? Like that's a metric that I always after these days.

2:25Sai Shivareddy:So if there's a fundamental advantage you have in enabling a higher uptime and a higher utilization rate, however the machines define, you'd have a long-term advantage over everybody else because you have something that's unique to you, patented, you manufacture it, you do all the hard things. It makes perfect sense and it must be a really exciting time we're building this business given the demand that we're seeing for the real hardware and infrastructure that underpin us ai companies um can you kind of maybe give some examples of like some of the most interesting or exciting partners that you're working with or where people are using your technology and the results that you see because you mentioned you know 10 100x faster be interesting to hear kind of real life examples absolutely yeah sure so well first of all um And if you look at where we've already deployed and our journey so far, so most of the battery technology companies or material science companies, you know, you start out as a new technology company.

3:25Sai Shivareddy:You figure out which problem do you solve because it's usually in deep science. You've got something new. It's a new scientific insight or a new material or new device. And you try to integrate or embed that into some ecosystem that you can enable something uniquely. And what we did in the first sort of four or five years of existence was to make sure that the material itself that we've discovered, we have a platform technology, essentially new materials on the anode for lithium-ion type battery system. So it's just one thing that's new that allows you to do something very different, which is accept an input power that is a factor of 10 higher than anything else out there without causing damage to the system or the cell.

4:13Sai Shivareddy:And what that means is you can use it forever or much longer than what you could with existing, say, batteries or capacitors. Capacitors have long life, but they don't store enough energy. Whereas batteries have shorter life and have more energy per charge. And they don't have those terms of like fast charge cycles that types of capacitors have. So we have a new device structure enabled by a new material that allows you to do both those things under the hood. And what that allows us to do is essentially enable things for customers in automated warehouse settings or essentially robot factories or AI factories these days.

4:57Sai Shivareddy:A high density that you get with power. So essentially kilowatt per unit volume, kilowatt per liter or kilowatt per kilogram. So that's essentially a metric that we stand out on that, which is a factor of 10x when it comes to input power density. So that input power density metric is what stands out combined with extremely long life. So when you have extremely long life, the cost per cycle of energy delivered, you end up with the lowest cost structure over the lifetime of the device. And a lot of these assets that are being deployed, they're being deployed forever, right? You don't want to change these things every year or every other year because phones and things like that, you might charge once a day, two years, 700 cycles, depending on how you charge it.

5:46Sai Shivareddy:If you have 1 ,000 cycles in a battery, that's great. These applications, you're charging 20 times a second or 20 times a minute or 20 times an hour. It's depending on the application. So now you need cycle life that's like many orders of magnitude that's better than a traditional battery. So it's not, we're talking about like 20%, but we're talking about like many orders of magnitude better. Same way with the power density as well. So if you want to fit inside GB200 and VL72 rack, the latest, or GB300, like the Rubin architecture, you're talking about a factor of 10 to 50 times the power density of compute within a rack that was just five years ago.

6:25Yeah.

6:26Sai Shivareddy:So the densities increased many fold. and the technologies that exist to support that maxed out about three years ago. And we've been building a few years before that. So new things need to come to support this. This happens at the semiconductor level. This happens at the power delivery, at the interfaces. This happens on the bus inside the rack. So anyway, I wouldn't go into too many details here, but the extremely dynamic nature of the power requirements, dynamic means not steady, not stable. So it's constantly fluctuating up and down, up and down, many times a second, as I said, or many times a minute, depending on what the workload is.

7:07Sai Shivareddy:And in robotic settings, you want the bots to be working as hard as they can, like ideally with 99 or higher uptime. Today, the bots have less than a 30 % uptime because they're charging most of the time and working far lesser. so when we first launched into this space so this is with our first announced customer symbotic symbotic is actually by market value the largest pure play robotics company in the world they've got these large contracts in distribution centers such as those in the u.s of walmart these are public stories where all of the distribution centers is a huge multi-million square foot warehouses that are completely automated by symbolic solutions and we are sort of so supplying to this sort of an environment where it's huge installations and we're in for as long as a warehouse would exist you know in terms of that architecture and this is a highly custom architecture and you know we enable these things so we go at the ground zero in terms of the clean sheet and either plug and play into what they have or co-invent new ways of keeping fleets running with very high uptime so you know we're selling uptime in a way rather than selling hardware got it and the the symbiotic um i thought was really interesting that like leading the series c the series c was only up like just over a year after um the the round you raised before so i wanted to like yeah where where did the round come from was it was it a result of the the results that you're driving in real environments was it due to like overwhelming commercial demand and what was the thinking with raising i think it was around that was double the size as your previous one just a year after the last yes yeah yeah that's right so well some of these things they line deep tech space a lot of things happen ahead of um the usual metrics people are after like it's very difficult to like say what exactly happened but what tends to happen is when you've got your product market fit or you know clearly that your technology is indispensable and um you got to a point where it's indispensable.

9:30Sai Shivareddy:Generally, technology is not indispensable because you've not found that fit. But once you're in it, it's existential, right? Like you need it. So what you do with that is figure out now how much of your opportunities that you have, you know, you tie with one partner versus you try and split. And then we decided it's better to go long term with one partner and be in it together in terms of sharing their growth. because for them, their vision is to move every box in the world, right? Like when I talk to my customers, like, well, I want to power that move. So how big is that opportunity? And it's, you know, in their view, a trillion plus.

10:10Sai Shivareddy:And, you know, even if you like 2 % of that, just in like one deal, you know, you're in the multi-billion outcome in terms of your addressable opportunity, which is already logged in, right? So in this sector, it's transformational. Like an account can be transformational to the company. valuation is the arbitrary i'll say that but our revenue is grown 5-4 we're hitting 50 million a year run rate right now but it was less than 10 like six months ago um it could be it's a great nice could be 10x in the next 12 to 18 months or it could be 20x i mean so the outlook is very exciting um and partially it's that um and partially it's also your completely de-risk not just the technology but your ability as a team to execute build your supply chains build your manufacturing plants you know ship product around the world certify it the hard things that you can imagine when you have like complex supply chains with trades and tariffs and despite all those things we've been able to deliver and keep the customer happy so when they've seen that then you know you're in it for the long run because you can do things not any not just a new technology company can do you're doing you're doing things that none of their suppliers will do even if they will have to change a new change that existing product and tweak it a little bit because all of these companies take a long time to change things and you can change way faster because you know for you you you go with the speed that's also your your edge right as a startup and have you had to change much in the business you talk about the complexities of shipping and logistics and doing all this kind of complicated stuff involved when you've got a real hardware product growing at the insane rate that you have what have you had to change internally to enable you to stay on top of you know all the logistics while rapidly scaling well let's say we spent good part of 100 million dollars long before i had one dollar of revenue i mean that's that's what's different in in in our world where you know a lot of upfront capital gets started you know you You spend way ahead of revenue.

12:16Sai Shivareddy:And people who are not familiar with this, it's just insane. You can never predict when that inflection point comes. And as optimistic as you are, more often than not, the first part of the revenue is very unpredictable. But once you're in it, it's just a question of how do you ramp up production faster or how many partners you bring in. so the basic systems you set up long before that you have to set up long before that otherwise it'll break yeah okay that makes sense so you build the and it's quite different from the software world yeah so you you build for the structure to execute at that level so for instance i invested um about 20 million dollars in a production facility in in the uk long time ago long before we had any demand um and then we had to invest another like 20 million dollars in the in a boss and cell engineering facility and the design center in boss long before we had a single contract sign.

13:17Sai Shivareddy:And then same into a production facility to iterate all these designs to make sure that if there was the customer in that window coming in the next 12 to 18 months, your supply is up and ready before they need it. Or if you're making a change, that change can work in 12 months. So the things that you need to do to deliver hardware, especially when you have something new behind it, like a new material, like you've got to plan way before things happen and when it happens you know these are pretty common like first year of revenue you know tens of millions second year hundreds third billions i mean that's that's pretty common in our world like to get to that point to get to that first point is usually the hardest thing and then you have believers who believe in it and usually people believe in the team people believe in your strategy people believe in your vision and you know all of those things come together over a period of maybe two to three years just before you get to that point and so now you're at this point where you are experiencing rapid growth you're looking out at the size of this huge opportunity you have ahead of you what are the key bottlenecks is it you know supply chain supplies is that what it is you can't build things fast enough i mean just think about it right like your working capital goes from like 20 million to like 200 million and your revenue has to go.

14:30Sai Shivareddy:But it's not unheard of, like even in the SaaS world, right? Like your CAC and all those things, and you've got those metrics worked out, so you just get that equity capital. Whereas in our world, things like physical assets and inventory, you have other ways of financing it. And there's like debt and trade finance and supply chain finance and like upfront payments for materials. You know, you can split these things accordingly, but you cannot set up factories overnight. they take at least nine months or six months or and that's when you're doing extremely well in our part of the world it's usually years right so we had to be extremely nimble about or let's say yeah well nimble as well as um strategic about how we navigate all the global challenges you have and make sure that no matter what you're thrown with tariffs or blockages to supply, you have another way.

15:30Sai Shivareddy:But this is pretty common in manufacturing world. So you have the experienced team that understands some of these things. I've been in that trenches for at least two to three times before this. So I knew what I was signing up for. And I think part of that is really coming from the mistakes I made as a founder for over 10 years before this. And you mentioned that like revenue in the short term is quite unpredictable because it could be quite a big, one big partner, one big deal could massively transform revenue. How do you balance that then with trying to think medium term, long term about building out factories, building out the supply?

16:08I try to balance that with actually not always having a clear view of how much demand you're going to have next year.

16:13Sai Shivareddy:You know, it's a good question. It's actually something that makes or breaks companies. Like if you looked at the first EV boom, let's say the boom in 21, electrification, electric vehicles, we were in that space in a way because we put out an EV that charges in four minutes, you know, four and a half minutes to 80%. And that was the world's first inside an EV like a few years before any of the Chinese companies who are leaders today, long before even Tesla, you know, like that you can charge in four minutes. And what that allowed us to showcase was the team's capabilities and essentially what we could do with a tiny budget.

16:48Sai Shivareddy:And we did that for under$2 million. It takes companies like$100 million plus to get to that point, especially in Europe, even more. So you can do things on a shoestring budget. You can do things very quickly. So you can make plans that don't cost so much. That's number one. Essentially, you build trust with your supply chain. and if you're a multiple time founder usually you have that credibility with or you have your core team that you work with who have who have essentially all these relationships and with vendors supply chain partners and everybody else who will come with you in your journey and allow you to create that world that is needed because when you're winning they all win it's essentially getting the supply chain along with you like you would say in your investors along with you because you need that too but in hardware when you're shipping things a lot of the deals you do are very strategic in nature because these mature over many years they know that it's going to take time so if you have those parties with you so in fact part of our previous rounds we had with our suppliers who invested and now it's customers and these parties don't go away overnight right like for them exit doesn't matter like they're in it for whatever decades right so you're having different conversations with people like this so So it's not for the average VC.

18:09Sai Shivareddy:It's not for the average crossover funnel. It's not like, it's very difficult. So usually it's easier when, let's say it's at the boom time or like a hype cycle is the peak. Like it's quite easily money flows from non-strategic sources. But in, I mean, also you've seen like when it gets to an upper level, you have companies financing rounds, right? Like even in the AI world, as you see. So it's, these are extremely strategic in nature. and when hardware is involved, that's the only way. And did you go out and use traditional first? Exposure in terms of risk to supply chains. So there people will take risk with you because you're enabling that market that they don't have today, but then they're betting on you in the long run, right?

18:56Sai Shivareddy:So it's the same as an investment bet that you would make as an investor, your play investor. Yeah, that's really interesting. Did you go out to more traditional VCs? Yes, of course. I mean, I go to everybody. It's pretty common. When things are easy, it's so easy. When things are tough, market sort of retreated 22, 23, 24. I mean, it's still difficult for the traditional electrification segments and battery companies. Name one battery company that's actually got to this point in the last three years. Not just in the UK, even in the US. They've all had down rounds. So we had to navigate all the hype cycle because we were counter-intuitive or it was essentially like...

19:55Sai Shivareddy:What's the word for? The opposite of what is hyped up in terms of a technology. So anyway, so it's, so what we're doing, it was not on anybody's roadmap, right? It was just totally at the opposite of the spectrum. So we have to convince people that this is worth funding because it's not in anybody's roadmap, right? So it's totally left field from a pure play VC point of view, because when they go talk to the market, market will say, here's the roadmap for all EV players. And all EV players are running after energy. we were running after power and lifetime and when the ai infrastructure boom happened we were well positioned way well way more positioned than anybody else in the market like globally to fulfill on this as as this technology that was waiting for this uh problem set to come together so it's sort of you you need to have a bit of that foresight you need to believe in some of these things coming together within the five to seven year window because when you start out as a founder you tell your team, look, stay with me for the next half a decade in that sense, like five to six years.

21:05Sai Shivareddy:And we'll start seeing the success that we want to see. Otherwise, you know, nobody pursues anything forever, right? Like you have like a four-year resting cycle for options. People run after that or leave or things of that happen. So, you know, it's common on one level there. but people who stick with you for longer, like some of these booms, booms, yeah, events, you don't know when they come, but when it comes, you're ready for it. It's been similar to all the memory, you know, like the plan with the HBM side, like Micron, other companies, like they invested way more than their peers a couple of years before them.

21:43Sai Shivareddy:Right. So it's about timing. It's about taking that risk. It's about working within what the supply chain can allow you to make it possible and committing to that partnership or set of partnerships across and then yeah it's a different world but you know once you do it it works like magic because those are huge barriers like huge moats yeah it's amazing and i've got one final question because i think you are probably one of the most exciting or definitely up there in terms of companies have come out of cambridge in the last few years right it seems like the cambridge ecosystem is doing particularly well what's it been like being founded they're building from then scaling from cambridge well it's very hard there while there's a brand uh there's no industry in cambridge i mean yes there's like design industry like you see for arm and things that and yes there are like big design houses consultants cambridge consultants like all these like traditional like ttp all the all the you know the brains behind a lot of some of the fundamental technological transformations, obviously a lot of that commercialized elsewhere through hardware and other things.

22:53But it essentially told me that

22:58Sai Shivareddy:that shouldn't be a limiting factor because your customer's outside, your suppliers are outside. You start there, but you don't stay there from an operation point of view. So I'm on the road like 90 % of the time. But it's a great platform to start, I'll tell you that. It's maturing. Obviously, for AI and the AI talent density, that's unbeatable. Great. It's actually one of the best places in the world for that. But if you're trying to build hardware and trying to build factories, you've got to go further out. But that's part of the way the city was built. For 800 years. Thank you so much for joining me.

23:45like you absolutely fly a really cool business and a really cool deep tech company coming out of the UK so thank you so much for joining me Sai

23:52Sai Shivareddy:thank you Sted thanks for doing what you do it's great to have cheerleaders I love doing it