The future of DSR and battery optimization with Michael Phelan (Co-Founder & Chief Executive @ GridBeyond)

Maybe we were all just ten years early, and now DSR will have its play.

We now have access in many, many markets, as you say, to the energy market, which is eventually the the bigger market or the one that needs most work in terms of balancing. Behind the meter, you can potentially access more revenue streams. Say, for for example, for a battery, we solve some constraints for a DNO that maybe you can't do with a very large battery front of the meter.

If we just reflect on the situation in Great Britain where about seventy five to eighty percent of all the megawatts installed batteries, good to go batteries, are operated by third parties. But in the US, I've been struck by how many asset owners do the optimization themselves. It's it's much more kind of sexy out here to be vertically integrated than it is in UK. What's the difference in approach and technology for Grid Beyond to optimize a battery in, say, ERCOT or or Kaiso or Kaiso and Great Britain.

Hello, and welcome back to Transmission. Today, Clinton is talking to Michael Phelan, chief executive and and cofounder at Grid Beyond. The conversation covers the past, present, and future of demand side response, how AI is changing battery optimization, and a look at Grid Beyond's story and portfolio.

Don't forget to enter our one hundredth episode exclusive Michael, thanks for coming on the podcast.

Yeah. Thanks for having me.

And I can't believe we're doing this both in Austin, Texas, but remotely. And you're in a hotel room somewhere, and And I'm in the studio, and we didn't we didn't sort this out so we could do this in person.

Yeah. I guess it was a little bit last minute even in terms of myself as to whether I get in today or or later tonight or on a in terms of the flights.

Well, we can imagine we're in the same room. Michael, it's it's a delight to have you on the the podcast. You and your team have been doing flexibility stuff for an awfully long time, perhaps the first or one of the first companies doing demand side response, and now the company has evolved into something much bigger and bolder. Before we get started, can you just tell us about how you began with Grid Beyond and what the journey was over the last decade and a half?

Yeah, so we eventually started out doing, demand response and energy management to industrial customers and some commercial customers, mainly in Ireland and then in the UK where the market started to develop around that time.

We had a an idea that we would automate the system and make sure that the loads, generators, etcetera, that we were controlling would always be there and would be dispatchable in a in a known manner so that if we said we had a hundred megawatts, we could deliver a hundred megawatts to the grid and, make demand response, especially load based demand response, as reliable as possible.

So that was pretty much the founding of the company or the original idea of the company.

And when did you start the company?

It started around twenty ten.

So the markets were starting to become available in, as I said, in UK and Ireland.

And we could see that, especially with intermittent renewables, that the grids were going to need things like flexibility, so load flexibility generation, and eventually storage like batteries.

Grid Beyond started in Ireland and the UK, but you're now all over the place. Can we just do a quick rundown of what Grid Beyond is now, what the company does, who the customers are, and, who who do you serve, and how many how many people people are in the company, where are you in the world? What does the company look like? The top trumps card of Grid Beyond.

Yeah, so Grid Beyond is fundamentally a distributed asset management company. So we're managing everything from things like solar and wind, managing and balance and putting them into energy markets. We also manage hydro and anaerobic digestion.

We then have, on the other side, a lot of industrial load that we put into the market, a lot of generation, pea plants, and the likes, and, also batteries. So we put all these systems into the markets. And we're present now in, UK, Ireland, but also in the deregulated markets of the US, Australia, and Japan.

Wow.

We've about a hundred and fifty peep people working with us at the moment.

And who owns Good Beyond? I mean, congratulations. I saw you you guys did your series c not that long ago. Yeah. Who owns Good Beyond?

So it's a bunch of, strategic, investors and venture capitalists along with the management or the ownership of Grid Beyond and the employees.

The Grid Beyond went from a a DSR focus to, a battery focus to now lots of different asset classes. Does that mean that you're serving lots of different types of customers? You know, is it asset managers, or is it load, you know, commercial industrial loads? And and how do you manage having so many different customer types?

Well, I suppose for us the top level is customers that are on a path to zero carbon. So they're looking to decarbonise their energy use and, therefore need renewables from people with wind and solar and hydro and anaerobic digestion on ones on that side. They also then want to maybe twenty four seven power match, so they need things like flexible load and batteries to allow that to happen. So I suppose at the top level, we see ourselves as serve serving primarily the CNI customers, commercial industrial customers, but also more laterally working with SME and residential, maybe through API and and partners to provide flexibility services. But a lot of it is around, path to zero carbon and helping companies on the path to zero carbon.

And, of course, we're gonna talk a lot about batteries on this podcast. That's what we're sort of obsessed with. But before we get there, could you just reflect on the state of demand side response? And this is particularly it kinda hits a nerve with me.

I used to work at Kiwi Power. We were a demand side response company in in the UK and Europe. And I suppose the twenty tens was all about that the hype cycle was all about demand side response and flexibility and this idea that markets were gonna open up to c and I, and DSR was was the next big thing. And in my experience, it didn't really materialize as fast as we needed it to.

And there was a there was a ton of reasons for that, but I just from your perspective and Grid Beyond who have been doing this for no. It's your it's your fifteen year birthday, by the sounds of it, next year. What's the state of DSR right now? And what what reflections do you have on the last ten, fifteen years?

So DSR is still very important in that if you get control of the assets and they're you can forecast them and know what capacity you have at any point in time, it's I wouldn't say it's not far off equivalent to a battery in terms of what you can do. So you can store energy in some of the assets. You can, obviously, turn off the assets at other point at points and use them to balance the grid. You're right in what you say that the journey was a bit slower than people expected.

But now DSR normally is capable of getting into the reserve market, certainly into the energy markets, will become incredibly important, I suppose, as we go through the electrification of, of transport and and heating in the to balance the DNOO networks. We will have to use the the very loads that we're using for the electrification. We'll have to be flexible and we'll have to be controllable, or otherwise, you're gonna have terrible problems on the DNO network. So we think it's kind of, an equally important part of the journey as, say, batteries and other flexible assets will continue to be important, but requires a lot of, I suppose, intelligence to make it useful.

So you need to be able to forecast load as well as prices and various things. You need to have dispatch control on the assets as well, especially if you want to put them into energy markets, avoid peaks, and various things like that. So we see it as having a a good future for companies who have the kind of systems that people are using today to control batteries. If you use that those same systems to control loads, you can make quite intelligent.

Yeah. I think what for for me, the reflection was the the entry point for DSR in the UK, in many cases, was reserved services like store. Just think about the UK right now in the twenty tens, and we ended up with, you know, many gigawatts of non b m store across the system that was kind of acting as DSR and then some non dynamic frequency response.

But beyond the reserve services and ancillary services, we never really got the market access bit right to fully access market prices, and it just took so long. But it seems like with things like virtual lead party and some more sophistication in market access now in Great Britain, that's starting to happen. And especially, I mean, in in ERCOT since working here in the last year, it is amazing how large flexible loads contribute to the system over here. So DSR I think this the twenty twenties could be the time. Maybe we're all just ten years early, and now DSR will have its play.

Yeah. I think that's true. We now have access in many, many markets, as you say, to the energy market, which is eventually the the bigger market or the one that needs, most work in terms of balancing.

And, we have that under four p four one five in the UK. We have it, as you say, with the suppliers and through peaks and that in ERCOT. We have it directly in in California, in CAISO. We have it directly in PJM. We have it directly in Australia. So, yeah, I think PSR and and and indeed behind the meter batteries have full access to all markets just like front of the meter assets. And I think that could be a bit of a game changer as the as the markets develop in that, behind the meter, you can potentially access more revenue streams, say, for for example, for a battery, maybe solve some constraints for a DNO that maybe you can't do with a very large battery front of the meter.

What do you think about behind the meter batteries then? Because, I mean, it's just so hard to do. Right? It's that developing and installing these batteries, particularly on commercial industrial of the kind of half a megawatt to two megawatt size. It's just so hard to do because of all the transaction costs and the the the business case is very complicated.

Is that starting to get easier now?

I think it is and I I I think the thing that would be quite important is in certain markets, say if you look at the UK, say under one megawatt, you can get a connection more easily and more quickly than you can for anything above a megawatt. A lot of the grid is is constrained above certain sizes. This this is also the case in California, in parts of PJM, in New York, and various places. So behind the meter, especially if there is a exit available capacity, can be a a quicker path if you like to placing a battery in revenue than, say, front of the meter assets, especially as the, connection, capacity becomes less and a lot of the grid becomes constrained because of people putting in lots of renewables, lots of batteries, and lots of load indeed and, causing all sorts of constraints on the grid.

I've got to ask you. I mean, we're doing we're following in your footsteps a little bit. So we are a UK based company now in ERCOT launching in California soon and then East Coast. And you guys are a couple years ahead of us in expansion.

How has it been for you as someone coming from a GB and Irish grid perspective trying to learn all the ins and outs of all these different markets? Because to run a business like Grid Beyond, you do need to have some technical knowledge and understanding of all these different market structures and what what's the nuances between them. So what's that been like for you, Michael, in your personal journey, learning about all these different types of energy markets around the world? Any top tips for me?

Yeah. So we kinda top leveled it. So we were kinda looking at, look, they all have energy markets. And they'll either have day ahead, intraday, continuous intraday, real time balancing, Or otherwise, they'll have day ahead and real time like you see in the US.

Most markets have reserve, and most of them have frequency response, and some of them have capacity. So at the transmission level and energy level, you're always gonna find some combination of those, markets. We also saw a pattern. So US and, Canada and and, say, Australia are fairly similar.

They're day ahead real time type markets.

Whereas UK, Ireland, and Japan are fairly similar. They have day ahead, intra day, continuous intra day, BM, reserve frequency, and potentially capacity markets. So there are similarities in certain groups, and there seems to be two groups of markets as well, like the European style market, which Japan has adopted, or the American style market that Australia has adopted. But there seem we haven't really come across anything else. So they all fall into those categories. And then in the DNO, then you're starting to see some services coming available there around capacity and voltage balancing and things like that that you might see in Ireland or some of the markets.

Could you just expand on that for a minute then? The the difference between European style markets and US style markets. What what do you mean there?

Well it's just that the way they balance is somewhat different. So a lot of the work in the US is done day ahead and then it it's rebalanced again if you in the fifteen minute market or the scared market. Whereas in in, Europe, there's a lot of rebalancing done intraday through various auctions to make sure that everything is, is okay as you go along and there's even a continuous intraday, which is more like the real time markets. So just more markets, I suppose, to bid into more markets to do things within in Europe including UK and Ireland and Japan then maybe in the US.

And you guys are operational in Japan now?

We are. Yeah. We have some customers, going live as we speak.

Wow. I find that so daunting, as a a CEO of Moto Energy. One day, we will be in Japan. But the the language barrier, I find a little bit scary. How do you guys go about expanding as a European company to Japan?

Well, I guess with language, you have to use local people who obviously speak the language and can interface then with the grid operators, energy markets, and and various things. So, certainly, it becomes much more important that it's local people who have, obviously, good knowledge of the electrical markets. And, you could have problems then on the on your own side in terms of things being translated from Japanese to to English. But, we found that, you know, if we can get good people on the ground that know the markets, know kind of how the markets are changing, then other than that piece of translation, it's fairly similar to other markets.

Good luck out there. It's an incredible market. I mean, watching from afar, Japan is going through one of the fastest, most major overhauls of its energy system. It's incredible to watch. They really, really wanna make net zero happen quickly.

Yeah. They do. And, they're quite active, opening up the market.

I suppose developing the markets in the same way as probably UK, Ireland are the ones they're looking towards because they realize their markets are fairly similar to those. So they're interested in companies from those markets and looking at best practice from those markets, realising that UK and Ireland are probably a number of years ahead in terms of the transition.

Going back to batteries now, can we just do a stock take on you guys' portfolio? So what kind of batteries do you operate? How many megawatts do you have? Where are they?

And do you control them with your own software? Are you using third parties? Do you have traders? Is it all automated?

Could you just give us the the lay of the land for GridBeyond managing batteries?

So at this stage, we have batteries of, various types and sizes probably in all the markets that we're active in. So while we start normally with demand response, we fairly quickly get into batteries.

And we put batteries really of every size or most sizes into all available markets. So that can be anything from, the, energy markets, obviously, continuous trading markets in intraday, balancing markets, reserves, if they make sense in places like Texas, frequency markets, obviously placing batteries in all of those. We also manage batteries of all sizes. So we would probably have a lot more of the smaller batteries because we have a fairly distributed system and use a lot of AI based forecasting and robotic trading, which allows us to obviously manage smaller batteries where you do wouldn't necessarily have a person per battery as as some of the others do.

How how big is a smaller battery? Can we put some numbers around this?

What we think in terms of, let's just say people like us doing the trading, in in most markets that goes to twenty, maybe to fifty megawatts. What we're finding above that in particular in the US, and to a fair extent in Australia is that it's either utilities are large, developers are putting the batteries in place, and they tend to have their own trading teams. So they're much more interested in just buying, say, forecasting software and optimization software, potentially market access software, and even an EMS from people like us, but doing the trading themselves.

Whereas below a certain size, the developers tend not to have the trading competence and tend to want to buy it all as a service.

More like what we've seen in the UK where most of the developers didn't bring the trading competence in house and, used other parties like us and and a few competitors to provide that service in the in the UK.

How do you think that market's gonna change? You you guys must do a lot of thinking about this. So if we just reflect on the situation in Great Britain where about seventy five to eighty percent of all the megawatts installed batteries, grid scale batteries, are operated by third parties. So there are optimizers and trading houses and software companies like Grid Beyond and others that provide the optimization services. And so the asset owner doesn't tend to do that themselves.

But in the US, you're absolutely right. I've been struck by how many asset owners do the optimization themselves. It's it's it's much more kind of sexy out here to be vertically integrated than it is in the UK.

Do you think that's a timing thing, or is that a scale thing? So is is the fact that the u that that ERCOT built their assets just a couple of years later means that those asset owners are a bit more informed, and they wanna do it themselves? Or is it the fact that assets are generally bigger, kind of fifty, hundred, two hundred megawatts rather than starting out at ten, twenty, thirty? Or is it something else? What what what's causing that change, and where do you see the next, say, five years?

Are we on the cusp of loads of optimization services in ERCOT, or is actually the vertically integrated route the the best way for those customers to figure their own assets out?

I think it's probably going to go fairly strongly in the vertically integrated area. So you're going to see a lot of the solar developers, for example, putting batteries with solar in, especially in areas with duck curve. You're also going to see them wanting to provide the generation side of path to zero carbon to their customers, and probably have some confidence in terms of trading themselves. So I I I think in certainly in US and Australia, I think it's going that way.

Also in Australia and and indeed in US, you're seeing the utilities a lot more involved in in terms of building battery systems, in terms of in building solar systems, wind systems, etcetera. So I think it's very much that way in the those markets. I think in the UK, it's probably starting to go in that direction as well. I think as batteries become more mainstream and attached to, say, wind farms, solar farms, etcetera, I think you're probably just gonna see large developers and utilities wanting to use them and maybe using them as well to balance their portfolios, say, firming wind or firming solar and various things to make that available then to their customers.

So I think it's certainly, the larger systems are probably going in that direction in terms of the it's just going to be large utilities and large developers providing those systems and maybe looking for really good software systems that allow them forecast everything, trade everything, and then they can put their proprietary algorithms, if you like, into those systems rather than developing everything from scratch.

So I I think it seems to be going in that direction as far as we can see in a lot of the markets.

So if I've got you right, so your take is that you think that just to to say what you said back at you just to make sure I understand it, was that the the the optimization model of kinda giving the keys to a third party might swing back towards asset owners taking more control. And does if that's if that's right, does that mean your business model is changing? So rather than operating the assets, you're, you know, building the tools to help asset owners operate themselves. Is that is that what's happening?

Yeah. Pretty much. So above a certain level, we're more of a SaaS player. So say, maybe pick a figure of fifty thirty to fifty megawatts.

We see it, you know, that the battery has been developed and that scale up are developed by fairly large developers who are maybe adding solar, wind, various things as as we go forward, and they're just looking for software. So they're just looking for a software based system that they could use as with their software developers and traders maybe to add extra functions into it and and trade in the markets themselves and those assets themselves. Whereas below that size, say, and for certainly for behind the meter and smaller front of the meter, there may be an opportunity for people like us to provide those services.

And I suppose on the other side of it, we would be helping with the path to zero carbon. So on path to zero carbon, as you know, you're looking to power match against the load. So you might be looking at the data center and twenty four seven load, and you're trying to, make sure that as much as possible of that load is provided by renewable, which is obviously coming from the batteries maybe and the and the solar and wind that the developer has in the in the remote sites. And we might also then be controlling some solar and and batteries, for example, at the data center site.

Just to zoom in on that, so you've said path to zero carbon a couple of times. What what does that actually mean?

So it's trying to have a situation, say, whereby a data center is mainly provided for by solar, wind, hydro, anaerobic digestion or things that are non fossil fuel. And they're basically getting their power from those sources as opposed to from the grid and from fossil fuel sources.

And that we're power, we're matching that. So we're looking on the other side that the load. We're looking at where the power is coming from. We maybe have, obviously, information on the wind farms and solar farms and indeed batteries and other systems that are providing the data center. And we can give put a number on how much of the power is actually coming from those sources and how much is coming from the grid. And then we can look at the grid intensity at that time to see how much is actually coming from fossil fuel on a twenty four seven per match basis.

Alright. Can we talk about AI for a second? Because Grid Beyond, you guys have done a really good job of marketing yourselves as an AI company. There's a AI is all over all over your website when I was doing my research and your materials.

How are you guys using AI in your business?

So we do a lot of, forecasting. So we would typically look to forecast things like load. We would look to forecast renewables, things like wind and solar, anything that's intermittent, and we look to forecast price. So we will typically look to, forecast the prices in any given market, say, the day ahead price, the the real time price, the reserve prices, frequency response prices. Any market that we can get access to, we want a price on it. We also wanna see if the if the prices are gonna spike or the the grid is going to constrain, and we're gonna see spikes safe as you might see in a place like Texas or you might see in, Australia or or Ireland or or indeed UK, where you where you where sometimes the wind can't get out, it's gonna cause a spike. Or maybe here in Texas, it's very, very harsh and just hardly enough really to supply the grid, and there's gonna be price spikes then in certain places.

And then we use that to feed into an optimizer. So the optimizer then is working out what's the best thing to do with a flexible asset like a battery or indeed a demand response asset at the time of that price spike to make sure that, if it has spare capacity, that capacity can then be used to to bid into the market and make itself available at at those very high prices.

And and which bit uses AI? I mean, what what which part of that process is rich in AI, and and what what part of the process means that AI is, an important, factor of the technology stack, if you like, rather than just a typical, you know, typical modeling approaches or yeah. Which bit is is using AI and why AI?

It's really the forecasting. So in the forecasting, obviously, we're using ensemble of all sorts of algorithms to give us really good forecasts of the things like the the renewables, like the price, like the load. So heavily using AI in that area. We sometimes use deep learning as well alongside solvers in the actual solving of the problem as well. So in in really in those two instances of, forecasting everything and indeed optimizing things, we would look at all the time, really, I suppose using AI and forecasting and sometimes using it in in in the solving area.

Let's go to optimizing batteries in different regions. So what's the difference in approach and technology for Grid Beyond to optimize a battery in, say, ERCOT or Kaiso or Kaiso and Great Britain? What's the how does the approach change?

The approach, in other words, the core system is the same. It's the same types of forecasters we're using. The data feeds would probably be the same, but the relative weights of the different data feeds would change depending on the market. So we're taking data feeds from lots of places.

We could be taking hundreds of data feeds to forecast, nodes, say, in a in a NERCCOT or in a in a in a Californian market. But the relative importance, if you like, of all of those things will will vary, and we might use slightly different algorithms or blended algorithms then to forecast price spikes. And, obviously, price spikes are very important in in ERCOT, maybe a little bit less important in some of the other markets like UK, Ireland, and those. So the blend of things we're using in the forecasting is the same.

Our it it it the blend would be different, but the the fundamental algorithms we're using are are are kinda based on the same fundamentals. It's normally the same feeds you're using. You're just looking at different things, and you're understanding that markets are different. So if a market like Kaiso has, say, things that are more like capacity markets, like base interruptible program, and that that's gonna affect the pricing if they offer subsidies for batteries and things like that. That's gonna cause more batteries to come into the market because the economics are different. And therefore, it's gonna cause that market to be run somewhat differently and maybe be less spiky than an ERCOT where none of those things exist. In other words, there are no subsidies.

There is no capacity market. So there's nothing other than the pure price signals to pull people into the aircraft market. Hence, you see much more spiky behavior.

And is your software capable of literally bidding into all of these different markets, submitting submitting data to the RTO, submitting, bids to auctions for ancillary services and reserve products. Now are you are you really doing all this automatically?

We are. Yeah. No. It's a robotic system. So it's, uses probabilistic forecasting.

So it's looking at all the forecasts we're producing, the errors on the forecast, and it's placing then different quantities of battery, if you like, into different markets based on the errors that we're seeing to maximize the revenue really from the, batteries. So it's not the type of thing a human could do. So if you if a human was looking at the kind of forecasting we're doing and even how our solvers work, they'd end up with a headache after a a number of fifteen minute intervals.

So it it is a fairly sophisticated system.

I'll just think about, like, for example, you guys are in I think you named seven or eight markets with different, and and Europe as well. So let's say, how many markets are you in?

Well, at the moment, we're in, the main US markets, which are ERCOT and CAISO. Some work in PJM. So three main markets here. We do smaller stuff in SPP, MISO, and a little bit in in New York.

But then we're also in Ireland, UK. Then we're in Australia, which is obviously a lot of nodal markets, and we're in Japan. They're our main ones. And we would have a kind of a watchful eye on Europe and negative prices and various things.

So say say there's eight markets there and there's a hundred and fifty staff, you know, per it it feels incredibly light on headcount compared to some other optimizers that have maybe eighty to a hundred staff and are only in one or two markets. So something you guys are doing is very different to other optimizers because somehow you're managing to build fully automated software with all these nuances between eight different markets with way less people than other optimisers are. Is that right?

Yeah. We're much more technology heavy, and we would be providing the technology, say, in the US to trading houses. So the trading houses have the traders. Whereas most of our, say, UK competitors are the other way around. They're more traders with some software. So we're heavy duty software with some traders, mindful of the fact that we expect the market to go more in the direction of, the large developers having their own trading capability and looking more for a software, certainly in the US market, Australian market, and potentially in Japanese market as we go forward.

So for us, it's all about the software.

I just think about even just one market getting all of the nuances of timing and bidding and you know, just think about the UK, I mean, and Great Britain. So just the different dynamic services and the state of charge management forecasting and the way that you can combine that with, reserve products and then all of the forecasting for day ahead and intraday. Just that, even the algorithm to cut to forecast all that and to model it, and then all of the interfaces with the third parties for National Grid, to Epic, Norport, and all of that. There's just such a huge technology load to build there for one market. And then to do that with Eights and have it fully automated is remarkable. Let's talk about what's happening next. Michael, how long have you been at the helm?

So for fourteen years.

And what when you go and pitch GridBeyond to investors, you guys just did your series c, and you're talking about the business in five or ten years' time. What's the big future that you have in mind for Grid Beyond?

So we think that, say if we take front of the meter, we expect that the developer if you like, the developers will have their own trading capability and and will need software. Then as we look for behind the meter smaller assets, and in particular, twenty four seven power matching for as part of path to zero carbon, we think the CNI customers and eventual SME and residential customers are going to need a good bit of help in that area. So we think we're going to be building out into that area to help that transition. In other words, for our customers to go from where they are today, maybe not necessarily knowing what their carbon intensity is, to being able to, get close to twenty four seven, carbon free energy on one side.

We also think, though, as as we go forward as well, we're going to see a lot of problems for the DNO, that they're going to, obviously, the consumers will decide to electrify things like transport, and, heating and cooling. And as that happens, then you're going to see a lot of load or increased load on the distribution network. So, again, we think there's going to be a lot of need for flexibility management in those areas to to help, solve those problems, ideally using those very loads. So using the those EV flexibility loads and in the thermal storage and thermal, cooling system load or heating system loads that are inherently flexible, to make sure that we you don't have problems in the distribution network.

So we think there's a lot of it a lot of innovation needed, a lot of opportunity in front of us that we need to serve over the next period of time.

Alright. And now we get to my favorite questions. First one, we wanna hear what you'd like to plug. Is there anything that you wanna announce to the global battery community that listen to this, or if there's some news or something else, then this is your chance to plug, and then we'll get to the contrarian view.

So yeah. I think we talked about it a bit already. We think there's a need for and and a great opportunity in having very sophisticated forecasting systems for all the prices that are available in the market. In particular, trying to forecast constraints and price spikes, because obviously in places like Texas, you need to get the price spikes right, and you need to be able to forecast them, or otherwise, your battery is not going to have a very good year.

So we think there's an awful lot of really smart things that can be done in AI and through software to allow you forecast all those things. And then there's a lot of optimization that can be done using those probabilistic forecasts because there's gonna be errors on the forecast and feeding them into solvers or deep learning type systems to figure out what to do with the battery at at any given point in time based on those errors to to maximize its values. So we think there's huge value in that area from AI, and we think it's a really, really good use of AI. And, obviously, we have the software to do this kind of work, and think it'll be really valuable to people in some of the more volatile markets over over the next period of time.

Alright. And now onto your contrarian view. Michael, what do you believe that not a lot of the world believes?

I I think one of the bigger problems the market is going to face is if we see an acceleration of the of the electrification of transport and heating, we think that could cause a lot of problem in the in particular in the distribution network.

And, we think that this the systems are not in place to manage that today, and we think that could accelerate. So say people decide, hey. We need to do something about, the decarbonization of the world and, you know, to chain make sure that the places like Texas don't continue to get hotter and hotter every year and become uninhabitable at some stage in the future. So to to stop that happening, we probably need to electrify various things, go down the the path of of renewables.

In that, I we are gonna cause a lot of potential problems on the on the distribution network. And, obviously, having smart systems and derms and various things to manage that, that would be very, very important. But we're not seeing much investment in that area yet. But then you could argue maybe that that that heavy duty transition hasn't happened yet.

But that could happen very quickly. So, say, the consumers get a hold of that in, say, Britain, in Ireland, in California, maybe East Coast US. It's already kinda happening in Australia. Huge pressure put on the distribution network.

You then need very smart systems to manage all that flexibility, if you like, that would be on the grid to make sure that you don't overload the grid.

And so just so I've got your contrarian view right on this, is it contrarian view that you think the impact for the the the negative impact for DNOs to manage this or distribution networks in general will be way bigger than we're currently expecting. Is that right?

I think so. We're kind of already seeing it in connection problems. So the DNOs are really struggling to to handle the connection problems that they're they're being presented with in terms of the electrification of various things. So putting in EV charges, fast EV charges, putting in batteries, various systems, solar on sites, etcetera.

So we think that I think that's only going to get worse as time goes on. And I think maybe that whole area needs a little bit of a look in terms of it's going to get very challenged as you electrify. And I'm not sure it's, you know, that the systems are there yet to help the DNOs manage all of that. Everything from connection management to, variable connection management to, you know, making sure that the systems are in marketplaces, etcetera, in place to allow flexible assets to be put back into the d and o markets and and make sure they're not overloading and not overloading transformers and various things in the local areas.

In my my contrarian view, no one ever asked no one really asked me my contrarian view, but my contrarian view on this is that I I actually agree with you wholeheartedly.

I I can only see one way out of this, which is where people opt out of the distribution network generally. And so you have loads of smaller microgrids where folks are putting generation on-site, batches on-site because nobody can wait a decade or so for the DNOs or the distribution networks to upgrade as fast as we're gonna need them to. So I think if you're a business operating in the UK, you will have to make a very difficult decision, which is, do I wait, which is perhaps not strategically the right thing to do, or do I have to invest big in avoiding avoiding the use of the distribution network as much as possible?

Yeah. No. That that's that's probably what you're seeing and probably the very beginnings of it as say the fast charger people.

Yes.

They would frequently put in a battery to manage the connection that they can't get. So say they can't get a connection above a certain level, they might put in a battery. Obviously, that battery will give them resilience as well from the grid if it went down at certain times. What they're actually act actioning in that way, putting in a a battery to to for the connection that they can't get.

And and and that is is the very beginnings of that. And you are seeing in, say, some of the fire zones in in, California and that something similar where people are putting in a system that will carry them through, say, the grid being turned off in in an instance of a fire or something in that, you know, where the PG and E or somebody can't run the grid at certain times or there's a danger of calling causing far forest fires. So the the local area then needs its own microgrid. So we are starting to see that in different parts of the world, and I think it is definitely part of the solution.

And we are seeing as well some of the industrials and commercials, you know, where they're very sensitive to voltage needing to put in resilient systems using batteries or voltage support from batteries at times of you know, they're saying, look, we've had lots of outages in the last year or two, maybe caused by local renewables or a weak grid or something, and they need to do something about it and need to do it now because they can't afford that to continue. And and they're there for putting in things like batteries to solve those problems.

And it could create a cycle, the second order effect of this is it could create a bit of a downward cycle for distribution networks because if you're a business owner or business operator and you go to distribution network and they're saying you need to wait five, ten years for you to get the size of the cable connected I mean, the size of the connection that you need for the things you want to do. And so you say, okay, I'm not gonna do that. I'm gonna spend the CapEx on on-site generation and on-site storage.

As that gathers steam as a, mindset and as a as a movement, there'll be a huge kickback on the charging methodologies of distribution networks. You know, so much of your distribution network cost is in a pence per kVA model. And are you really going to if you've got to spend all of that CapEx for on-site generation and storage, are you really gonna wanna pay the DNO in a pence per KVA as well? And this huge amount of your bill is just to keep the cable open if you like. So a lot a lot could change in that very quickly, but I think it's a good thing. I think, businesses and consumers using generating and using their own power, I think, can only be a good thing. The only downside is that lots of smaller systems don't get the economies of scale as, smaller a a a smaller number of bigger systems.

Yeah. That's true. And I mean, another example of this certainly, we're seeing it in places like Ireland, and that is where maybe data centers can't get a connection or can only get a connection to a certain level. And again, they're putting in, say, gas generation, potentially batteries and various things to help them out so that they can actually not build a data center today rather than waiting five to ten years because they don't want to wait. But, again, it's, people building a local generation source, generation system, local microgrid, because they can't get the connection from the DNO or TSO.

And, I mean, you just look at look at the order book for NVIDIA right now. You've got basically two to three years of the whole the whole production capability of of chips essentially for AI is bought in advance. And so all of these chips have got to go somewhere, and they they they need an awful lot of power. And if you can't connect to the grid fast enough, then you will build gas generation or you will do something else. So lots to come there. Michael, we're gonna, we've run out.

Thank you for listening to Transmission, a MODO Energy podcast.

Transmission delivers conversations from industry leaders and experts exploring energy markets and the operations and technologies related to grid scale battery energy storage. Check out our other episodes by searching Transmission wherever you get your podcasts.

Check out the Energy Academy, our video crash course on how markets in Great Britain and ERCOT work, or head to motoenergy dot com to see our written research Thanks for listening