Ep14: Powering AI Data Centers With Long-Duration Energy Storage

(upbeat music) Joining us today is Giovanni DiMato, president of CM Blue Energy, where he leads US commercialization of long duration energy storage technologies, designed to support grid resilience and high load infrastructure. He brings more than 15 years of experience across energy storage, distributed energy resources, and grid-connected infrastructure with prior leadership roles at EPRI and strategy and consulting. Giovanni has been closely involved in shaping early storage market frameworks in California, and now focuses on scaling non-lethium solutions for industrial and mission critical demand. Giovanni, walk to the show. - Thank you, Samir. It's a pleasure to be here. - Great to have you on. Thanks for spending some time with us today. So for people who don't know you, can you tell us a little bit about your path? How did you get to be here? How did you get to be the president of CM Blue Energy? - Yeah, so I've been in the energy storage industry for about two decades now. So around since before lithium ion was really a thing and watching it grow and before coming to CM Blue, I had a long history of working with lots of different players in the market. So at EPRI, I spent about eight years working with utilities to help them understand the value proposition of energy storage for the grid and for utilities specifically, as well as before that helping lots of different energy storage and solar companies commercialize and go to market in the United States in management consulting and also helping in California bring about the first energy storage procurement target, which was AB 2514 in California back in 2010, which was very successful and a lot of others have replicated that since. - So you spent a lot of time in this technology, I'd like to acknowledge you, what is it about batteries that fascinated you? - I think the most interesting thing to me is that they are a swiss army knife. I think a lot of people have used that analogy in the past. Also, it's kind of like bacon. Bacon goes good with everything. So does energy storage. There's a lot of different use cases for energy storage because it's dispatchable and it can be both a load and generation depending on how you're dispatching it. So that really opens the door for all sorts of use cases from the very edge of the grid to where maybe it's not even connected to the grid like a microgrid or off grid, all the way up to pairing it with nuclear power plants with conventional fossil generation or with large renewables and everything in between. So that's what makes it very interesting to me. - Very cool. And can you tell us a little bit about CM Blue? What does it do? What is it doing differently from other battery companies out there? - Yeah, so I would say that our unique value proposition at CM Blue is our solid flow technology and that really has two parts to it. So one is we are not using critical materials rare earths, other things that can't be locally sourced and don't have any geopolitical risk in terms of sourcing from all the way from the active chemistry all the way up to all the components. And that's kind of been our strategy from the beginning to be able to manufacture locally. And the other piece is our longer duration. We're really focused on 10 plus hours, where we think that the market is going and needs and there's a gap there and we're ready to fill that gap. - That's really interesting. Most of the projects that we work on are four hour batteries. You know, 10 hours changes the economics dramatically, I'd assume. Can you tell us about the scale of these batteries? Like how big are we talking about what's the typical project size? - Yeah, so right now our first version of the product, we are looking at mostly utility scale, also large commercial industrial and commercial industrial includes like data centers, AI data centers, as well as, you know, like I mentioned earlier, pairing it with large generation, either renewables or conventional generation. So kind of our main building block, we call the powerhouse is a 25 to 50 megawatt size plant at 10 hours, so that's 250 to 500 megawatt hours. And they can be scaled, you know, up or down depending on the specific use case and the needs of the site and the customer. - Yeah, those are really big projects and I'm like now, just also just curious about the cost, how does the cost compare to batteries that are being sourced? Either domestically or internationally? - Yeah, I mean, economics is key and at CM Blue we're our solid flow system. We see it as, you know, competitive today with lithium ion in the over the five to six hour duration and as we go to use cases and customers that need, you know, longer than that and need that 10 plus hour duration, we're more and more competitive than lithium ion, which is the, you know, the benchmark today, 95% of the deployments out there. So we see that as the hurdle to beat and we're driven to keep our economics, you know, at par and better depending on the application. - You mentioned that the rare earth mineral risk isn't there. Is everything sourced domestically? Is there a tariff risk associated with your product? A big part of the, I guess the question that we're solving for with our clients right now is just around the tariff risk and it's, you know, as you know, it's completely unpredictable at this point and a lot of people are struggling with it. - Yeah, and, you know, from the beginning, our strategy has been to produce locally, including the supply chain from, you know, our active chemistry. The feedstock for that is something right now, you know, our headquarters is in Germany, just outside of Frankfurt and that's where our first manufacturing facility is. We are sourcing all of our components from Europe at this point and that's something that we want to kind of copy and paste and one of my jobs is to bring that manufacturing to the United States as well and that supply chain from, you know, the feedstock up to the components. And so I think that in Europe, you know, we're doing it now, we're proving that out and I think it'll be even easier in the US to do that as well. You know, the feedstock is all abundant materials. It's non-lithium and the other components, such as, you know, we have some pumps. Those can either be automotive fuel pumps or pumps used in cooling systems today. So they're very well known, very deep, you know, supply chain that can be done locally, you know, through the automotive industry or other industries. So, you know, those types of things make us, I think, unique and put us in an advantage for those fiat concerns in the US market. Also just in general geopolitical risk of not having your energy solutions and your resilience to different things that might change in the politics and geography to have that local as well as just the economics too. These are infrastructure projects. So, you know, building these things close to our, or manufacturing these things closer where you're going to build the projects makes a lot of sense from logistics standpoint. - I want to talk about something that probably, a lot of people out there don't think about on a daily basis and that's a load shape and demand. And you've pointed out in the past that electricity demand is being reshaped by AI and data centers and industrial electrification. Can you help us understand what that means? - Yeah, so, you know, before joining CNBLUE, I mentioned I was at EPRI for about eight years focused in the energy storage program. And I understand, you know, how the utilities, their mindset and also how they operate. And data centers present just this once in a lifetime challenge, I think, or maybe the lifetime of the entire grid, right? That certain data centers they talk about, you know, being 10% or more of an entire utility service or a total load. That's just unheard of for one particular source to be that big. Also the dynamics. So we're looking at not just data centers, but now AI data centers that have these spikes and training loads as opposed to when they're just being called, you know, in chats and things like that by regular users, those things really change dynamically. And that's really stressful for the grid. It's also difficult from an interconnection standpoint. So a data center and AI data center might need a 200 megawatt load interconnection. And the utility says, well, we can give you 50 megawatts in a reasonable time. So that's kind of like bring your own generation to make that happen to get to the speed to power kind of concept there for AI data centers that need to go it online. They want to go as fast as possible. But there's only so much speed that a utility can provide and the grid to make that interconnection happen. So if they can bring energy storage to that equation and bridge that gap, I mentioned between the 50 and 200 and also bring conventional generation, bring you renewables. solar and wind to that project, then you can very quickly have that full 200 megawatts for the interconnection for that project that you need and do it in a way that's also 24/7 reliability, resilience, all those other things that AI data centers need as critical infrastructure and to have that really high uptime. I was at a data center conference just earlier this week and we were talking about this exact issue and I'm curious with kind of where you sit in the industry. Are you seeing a lot of hyperscalers out there who are looking for this solution, the sort of bridge power solution with batteries? Yes, so our engagement is across the board with all sorts of hyperscalers and data center providers, either third party or direct to the usual suspects that the big hyperscalers as well that they are very interested in getting online as quickly as possible because they're all in a race to get their data centers online and they're fighting against a real physical problem of being able to build the infrastructure for those size loads and I think a 10 plus hour energy storage duration really gives them a toolkit to be able to pair whatever type of generation rather that be you know solar wind or conventional gas to be able to get that you know the scale that they need as quickly as possible and do that with the flexibility that they need as well. So I was on a panel earlier this week where we were talking about how volatile AI powered demand is which can impact on side grid and generation equipment for example by creating torsional stress on generator shafts and the topic of best came up and how it can help to maybe solve some of those problems. Can you talk a little bit about that? Yes, so energy storage can be a buffer between that volatile load or the dynamics of an AI data center ramping up extremely quickly for like the learning and not put that stress onto the grid. So it buffers you can you know in seconds or sub cycles so you know like less than a hurt one hurts or less than the 60 hurt cycles you know milliseconds be able to bring down the load and absorb that in the energy storage system. What's also great about energy storage in general that it's multi-hour so it's not just looking it can handle those very small changes but it can also help at the longer hour multi-hour scale to also bring down the loads you don't so a utility or grid operator doesn't have to ramp up another peaker plant or another other generation to be able to serve that load. So I think that the you know there's just the operational piece of it on the grid then you mentioned you know where and tear on equipment. So I think conventional generation there was never the expectation that you were going to have these types of data center loads like ramp up like one gigawatt in a few minutes or a few seconds even and energy storage can handle that and seeing blues solid flow system can handle that with no problem in sub-second intervals you know it's inverter based it's just very quick at being able to respond to that and not have the operational or wear and tear impact that mechanical systems would have with those types of dynamics. Yeah hearing that and having I guess my like lure head on thinking about the contracts that we enter into with suppliers for other types of technologies for example natural gas power plants you know they they're there are always carve outs to warranty terms around wear and tear and if you're doing something and operating projects in a way that's resulting in a lot of wear and tear it seems like that's just that's just a recipe for disaster because you're not going to get that covered under warranty claim and you're you're risking the the life for the project and an uptime and all that kind of stuff. It seems to me hearing like what you're saying like battery is not like a possible solution it's almost like a must have for these large data centers. Yeah maybe biased in my response here but I think that going back to my Swiss Army knife analogy and why I've been in the energy storage business for almost two decades now is that it can solve those problems in a multitude of ways and with a multitude of technologies to address it in the energy storage space and you know I think it's it's just a very powerful tool from down to milliseconds up to 10 plus hours applications and do it all at once at the same time is an amazing tool that the grid didn't have you know 20 years ago you know everything was just in time and we paired exact the demand exactly with the generation or the supply where you know that's not the case for just about every other industry in the in the history of the world that we had storage and now we do have storage and utilities really love that tool and I think that it they do see it as a must have and it's just you know there's more and more evidence as well as data and you know actual projects live to show how energy storage can you know really help the grid operators the utilities and the customers solve really difficult challenges yeah I feel like this is this is all happening live and rather quickly but I'd love to see someone put out a study that demonstrates the impact on the longevity of the project overall uptime guarantees all that kind of stuff so lithium ion has kind of dominated for the past I don't know 10 years or so accounting for over 95 percent of installed us grid scale storage what are the trade-offs that people should understand as they evaluate lithium ion versus flow batteries and even thermo and hydrogen systems yeah so I think one of the key topics that's you know fresh especially in the US with some you know fires and things that have happened in recent times is that there's this risk of communities and regulators and politicians you know slowing pushing pumping the brakes on you know putting huge lithium ion systems you know next to schools or hospitals or just neighborhoods in general because you know there is a toxicity risk of the catch on fire and then the fire itself of course you know those things are all need to be considered I think they're technical challenges that we can address but you know the risk is still always going to be there so you know things like we talked about AI data centers and other like critical infrastructure schools hospitals you probably you know in general you want multiple options so I think you know when when utilities when customers go out to the market they don't want to put all their eggs in one basket on one particular technology I think it's great to have a lot of different options and that's where CM Blue Solid Flow technology comes in where we can provide some of those other options in terms of you know we can cite our system you know in a data center hall in a generation hall and not have that thermal runaway risk there's also some other things like the geopolitical risks of we're seeing in the US you know the fiat constraints for being able to take advantage of the tax credits there's there's some real questions about you know can we do this quickly enough in the US to avoid you know not get capturing those incentives yes you can do the manufacturing maybe can't but as the levels continue to ratchet up as a year to year they go up as well as who owns it so divesting is also an issue that you know what a lot of the companies are Chinese that are lithium ion are they actually going to sell their shares in the company to be able to produce in the US I don't know that's a tall order compared to just the manufacturing piece you know I think that customers utilities and and the industry want options and CM Blue is you know from the beginning our strategy has been to build local from feedstock to you know all the component manufacturing so that's a we see as a real competitive advantage for everyone to have you know those options the flexibility to you know address this huge unprecedented demand growth and also you know how you address that with potentially with energy storage and with longer duration energy storage like CM Blue you you kind of alluded to this earlier when you talked about batteries as operational assets but meta announced in April that they would be deploying ultra long duration storage to support AI data centers as an active operational asset for load shaping and 24/7 you know clean power can you talk a little bit about batteries as an operational asset rather than an insurance policy yes so I think traditionally or conventionally energy storage at data centers has just been UPS or uninterrupted power supply where it's just bridging a small gap when the power goes out and when they're on site diesel generators can ramp up which is like 10 minutes maybe or less I think that meta announcement that you alluded to, they really see the value of the full Swiss Army knife potential of energy storage. So yes, it can do the UPS, that conventional application, but it can do much more if it's longer duration. So when you make it longer duration, it could be used every day in the cycling of the load profile of the data center so that you can bring in renewables, you can bring in other generation and optimize its operation and do that on a daily basis where you're bringing down the cost to serve the data center, as well as continuing to bring up the reliability, the resilience, and all those other aspects that the data center operators really care about. In addition, the economics too. So if you're lowering your load, you're lowering your power cost from the grid, you're also maybe even making it possible to bring your data center online, because it takes so long to build up the infrastructure, transmission lines, large scale generation. If you can do that quicker, companies like hyperscalers like meta really value that, and I think that the industry values that as well. - So I have a question for you. You hear the data centers are required to bring their own power. Should they be required to bring their own batteries? - I think those kind of go hand in hand. So I think bringing the whole ecosystem probably is the way to think about it. So I think that they are thinking that way, 'cause if you're bringing your own generation, you need to have the whole distribution grid piece in mind that's serving that load. And energy storage is just a natural fit there, especially when you need the flexibility, 'cause you're not a full grid operator, you're kind of hyper focus, you need flexibility. And energy storage is flexibility. - What do you think is one thing that regulators are getting wrong today when it comes to batteries? I know there's a lot of proposed changes that are kind of happening around the country, but I think there's anything in particular that they're getting wrong or that they really need to work on. - I would say from the non-Lithian longer duration perspective, we're newer to the market as a class in general. So it's really important that regulators don't have their blinders on to think just because 95% of the installed market today is lithium ion. Don't put regulations in place that limit us to continuing down that path. Open it up to make sure that all technologies are thought about and allowed to participate. So a very specific example would be when you come into a small town, let's say, and you're talking with the fire department, and you say you wanna put in an energy storage project. Let's not jump to it being lithium ion when you say that. It should be that you're thinking about all the options and the regulations, the laws allow for that type of flexibility as well that you can have lots of different technologies when you say energy storage. - We've had a small number of incidents, but like any new story to kind of get magnified where batteries are, you know, the explode or they set on fire, is that a real issue? You know, like you said, it's a small number of cases. It is a real issue. It's a real risk that has to be mitigated. I think there's a lot of technology out there to help mitigate that risk. And there's definitely a place where you can put those lithium ion energy storage systems that have thermal runaway risk or explosion risk. But there are also places that you can't do it. So, you know, having all the tools in your toolkit in terms of energy storage technologies can help you address those places where it's absolutely critical that you don't have those risks. And also where you just don't want those risks. - Makes sense. Is there anything you think that data center owners or end operators are completely overlooking when it comes to batteries? - I would say, you know, we talked about that meta example. I think there are a lot of people in the data center industry who are still, when I talked to them at first, thinking of energy storage as just UPS. And it's so much more than that. It's that Swiss Army knife that you can apply to lots of different use cases within applications within the data center, especially when you get up to, you know, the longer duration, the 10 hours where you're literally using it almost every hour of the day, either charging it or discharging it. It becomes a key tool or a key part of the system at the data center just like the server X. - So the last question I have for you is, what's next for CM Blue? - Well, I think that the biggest thing that I'm focused on and I think that we're focused on in the US in general is to get projects, you know, realized with, of course, data centers, AI with utilities, with IPPs and others as well as in parallel bringing our manufacturing capabilities that we've proven out in Germany today. We have a gigawatt hour facility there, a pilot facility and bring, you know, that gigafactory to the US and be able to really show our, I think, competitive advantage of being able to build locally and have the supply chain local as well. - That's great. Giovanni, I wanna thank you so much for being on the show today. I really enjoyed the conversation. - Yes, thank you, Samir, this was great. - Thank you for listening to UpTime now. If this episode sparked an idea, pass it along to someone new network and leave a comment. I'd love to hear your thoughts, ideas, and recommendations for guests. Till next time, thank you and remember to subscribe. (upbeat music) (upbeat music)