Trading Smarter, Not Harder. Managing complexity in global power and renewable portfolios

00:00

Irina Reitgruber Hello and welcome to this webinar, Trading Smarter Not Harder. I am Irina Reitgruber from ComTech Advisory and I will be a moderator for this webinar. ComTech Advisory, for those who don't know us, is the analyst company for the software tools market for energy and commodity businesses. And now I will start with introducing our participants. And let's start with Arne, please.

00:40

Dr. Arne Weber I am Arne Weber. I head up Power, Gas and Emissions for MB Energy in Hamburg. Formerly Mabanaft.

00:47

Irina Reitgruber Thank you very much. Then Sameer, please.

00:52

Sameer Soleja Hi, I am Sameer Soleja. I'm the founder of Molecule, an ETRM Cloud Software Company.

00:58

Irina Reitgruber And finally, Andy.

01:00

Andy McMillan Hey everyone. I'm a product lead here at Publicis Sapient focusing particularly on renewable energy integration into conventional asset fleets.

01:09

Irina Reitgruber Thank you very much and we'll start with discussing different trends and challenges which we see in modeling of modern power market with all the renewables, with energy transition technologies. And I will say a couple of more just to, to set the scenes.

So energy transition from net zero perspective had brought absolutely new technologies and quite revolutionary changes to the power market. And this is from different aspects. So on one hand we have distributed production where we have smaller and distributed across the country's landscape of different renewable sites. We have prosumers, so these are consumers and producers at the same time. We have new incomes to the market, owners of the wind farms, independent power producers. We have smart metering and that's all brings, first of all, a lot of new data. So there is a question of data management where we need to, so energy companies need to manage huge amount of data forecast production, which was not necessary before.

The second impact to that production is intermittent. It means that you need to forecast it, but you also need to substitute missing production capacities when there is not enough wind and solar, and you need to do something about it. Whereas, too much wind and too much solar. So what does it mean?

That's have quite a high impact on the energy system. So in order to stabilize the network, you need to have certain flexibility. So there are energy storage technologies which can help. There are flexibility markets where flexibility of prosumers can be used, aggregated in virtual power plant and used for the network stabilization.

And there are also real-time markets for balance markets, intraday, real-time markets, which also help to do that. So this is a next impact, but it's not enough. There are certificates, tradings, EMOs, which are very complex sometimes and differ from region to region; and if it would be not enough, the technology is the same across North America, Europe, or any other countries. So you have the same wind turbines, you have the same solar panels and so on. But the markets are very different. The contracts, the power purchase agreements are very different. The communication to the market operators are very different. And certification trading scheme is also very different.

And this is the situation we want to discuss in the very beginning. So before I ask our participants to respond to this, I would like to have the first polling questions. So first set of polling questions to show up. And this is about what are the biggest challenge with modeling power in your portfolios.

So as long as participants people on the webinar answering these questions, let's start with our discussion nevertheless. And so the first I would ask our need to comment on European situation. What are the biggest challenges for European market participants from your point of view?

04:36

Dr. Arne Weber Look, I think it's a, it's an interesting, well, there, there's of course many, many challenges and you already outlined quite a few. I mean, there's intermittent production we have to respond to, to changes in generation. I think that's, that's been a known challenge for a while. What I think is changing a lot in Europe at the moment in particular Germany market, we're quite active in is the whole, I would say, say challenge of local versus a countrywide sort of balancing. I mean, we know that in Germany we have a lot of production in the north from, from offshore wind. We have a lot of consumption in the south and actually one of these big challenges that also goes hand in hand with technology challenges. How do you balance that there's too much energy in the north very often and through a clearing price that sells for the whole of Germany or for the whole of region.

We have the same situation in Norway or in Sweden, or you name it, you kind of have to strike a balance and make sure that the, the supply and demand is actually met also physically on the grid. And I think this can be solved by introducing price zones. This can be in solved by excluding certain assets at certain points in time, but end of the day, if you want to have an efficient market, somehow technology has to play a role in between. And we see that there's batteries. Now growing batteries can be quite local, can be quite specific in where they generate, but of course they need a price signal in order to be built in the right places. Otherwise they just feed into the grid as they see fit. So I would say the whole challenge of well re-dispatch or, or local versus global, regional generation is a big challenge and I think technology can actually play a key role in how to, how to help there.

06:13

Irina Reitgruber Mm-hmm. Thank you. Andy, I would like to ask you to add something to European. So you are working cross region. So what is the comparison between challenges in Europe and in North America?

06:26

Andy McMillan Certainly. Good question. I think what Arne just said is think the, we're seeing a trend of kind of more standardization across a lot of Europe. So lot of markets are starting to standardize and promoting, kind of working closely together. That drives a lot of kind of crossover and the way that you can kind of build your ETRMs to work. Where you can contrast that to the US and you are effectively building kind of almost six different ecosystems to account for the market. So each kind of region can account for different real time kind of bidding rules. Location and congestion management is in incredibly kind of complex and, and very kind of very different, I guess, to a lot of the constructs that we see in CE. So trying to have the kind of one ETRM that works and standardizing the way that that works across all parts of the US is incredibly complex and it requires a lot of flexibility in the way that you configure that configure that to work.

The fact, yeah, I think that's probably the main, main part. Is it kind of as the trend probably to summarize it, is across the CE, we're we're seeking to standardize and across the US It's more and more flexibility is required to account for kind of different market constructs.

07:31

Irina Reitgruber From the ETRM system perspective, Sameer, how are you resolving those differences and where you see the biggest issue? Is that communication with the system operator or the construction of the market itself, or certificates, or could you please comment here?

07:49

Sameer Soleja Yeah, so from the ETRM perspective, I would say our experience has been completely the opposite to what Andy described in terms of market construction. It may be because we started in the US but we see contracts that are far simpler and far easier to model in the US where in Europe we see things with more complex pricing attached to them. You know, further in emerging markets, we see things with really, really wacky pricing. And so from a systems perspective, the challenge has been modeling term with extremely high resolution.

Adding in, layering in, complex pricing to it, and then, yeah, to your point, connectivity with TSOs or market operators. You know, there's a lot more in Europe and they're a little bit less centralized, I would, I would argue, that has been a little harder on our side, whereas at least in the US you can, you can partner with or buy a solution or even build a solution that connects to the six ISOs. Yeah.

08:53

Irina Reitgruber Mm-hmm. By the way, do already have some responses to our poll, if we could show it. Okay, so managing intermittent renewable production has the most votes followed by real time power markets, including real-time risk and automated and all trading and then regulation compliance changes. Yeah, so this is interesting response. Do we want to comment? Does someone want to comment on this?

09:32

Dr. Arne Weber Maybe if I, if I, you know, just listen to Sameer and, and you know, there's obviously complexity in the US, There's maybe a bit more of the same in, in Europe, but I would say the, the real challenge is actually. I mean, I'm, I'm just an, a, a subscriber, a buyer of the systems. But I would say the market is always ahead of the systems, right?

We kind of do crazy new stuff. There's new regulation, and actually, I would say the, the thing that is for me, you know, the decisive factor if whether a system is good or bad, is whether it can adapt to these changes rather quickly. Because if there's one thing I'm sure of is that, that I don't know what's gonna happen next year.

There's gonna be the next mega trend. There's gonna be the next thing. And then the question is, how quickly can you adapt? So actually for us as a, as an organization, we more and more, stop caring, I would say, at least from a front office perspective, and our, IT guys may have a different view on this, but for me, selfishly, I think the most important thing is flexibility. I want to be able to adapt to the changes, and if you ask me what that requirement really means is the answer is, I don't know. Because the, that's, that's kind of the point. So I, I think the, that, that will be my, my outside-in of perspective on, on tech. Right. I, I don't care as long as it can adapt.

10:41

Sameer Soleja Well, and that's, I think—

10:42

Irina Reitgruber We'll go into this a bit later exactly to discussing what requirements on the ETRM system come out of it. But Sameer, sorry, I interrupted you. You wanted to comment on the responses.

10:54

Sameer Soleja I was just gonna say that that's what makes it fun, right? It's not like our customers will tell us, by the way, we're gonna trade this new thing in six months, please be ready for it. It's more like, I traded this thing yesterday. Why can't I put it in the system? Please figure it out. And I need a P&L in 10 minutes.

11:12

Irina Reitgruber That that's, that's good. Yeah. By the way, we did not touch risk management yet, and with this new environment, risk management is also impacted.

So we have new companies coming into the market, independent power producers and so on. So we have some challenges and counterparty risk perhaps. Then we have all this real time trading. Do we need real time risk management? And finally, the complex deal valuation. If you are saying that European PPAs, so power purchase agreements, are even more complex from project pricing structure than North American one. We are ready for this valuation models for PPAs, for batteries. So who would like to take lead on this? Raise your hand.

11:57

Andy McMillan Yeah, I mean, I'm, I'm happy to answer it. Sameer, maybe I've just been unlucky with some of the PPAs I've seen beforehand because it sounds like we've had different, again, my answer on that arena would, would probably build a bit on what Irina was just saying around, from an IT perspective, we just need to keep pace with the commercializing kind of strategies and kind of approaches that the, that our business teams are taking.

So, you know, as a, as a kind of commercial asset optimization team or as a commercial team, they are increasingly looking for the best ways to commercialize and manage risk around the the products that they're offering out into the market. And the platforms have to keep, keep pace with that. It shouldn't, it should not never be the case. That the platform you've got limits to what you can commercialize and what you can trade. And that's certainly keeping us and IT on our toes around what we're building. So I'm getting new PPAs across my desk kind of every week at the moment that I've got increasingly kind of complex elements to it, whether that's pricing structures, termination clauses, or in the case of kind of best just that, that revenue stacking kind of approach and not just modeling the initial PPA, but all of the other elements that come with it. So the ancillary services and, and any other kind of markets they might participate in.

So yeah, it's, it's kind of the pace of change is probably the, one of the key trends outta that at the moment around your ETRM, you need to be able to kind of plug and play and try something new because yeah, the first you hear about a problem sometimes is a trader's gone to book something in the system and it's broken or it's, it can't do what they want to do.

13:25

Irina Reitgruber Yeah, so this great new challenges for the ETRMs and we will discuss that definitely. First, maybe before going into this, let's talk about technology. So we discussed what are the changes in the market, which brought new requirements on ETRM solutions, but technology is changing as well. And when I think about like, five years ago, before COVID time, energy companies were saying, I will never go into the cloud because I don't want to expose my data to any kind of security, I don't want, I mean, this is my very sensitive information and what to keep it in house. But it's absolutely opposite what we see now. So cloud technology becomes, must be people are talking about scalability, about cloud elasticity, about software as a service, as must be requirements in the RFP. So also if we look at the most recent trend is AI. And people are looking to adopt AI-based technology for certain parts of their businesses also within ETRM. And here you have even more requirements on their quality and quantity of the data on, real time on, on cloud and so on.

And in this respect, I want also to mention a survey which was produced this year by a Molecule where more than 400 market participants were asked about modernization drivers for the system and the current condition of the ETRM systems. And the most interesting answer from my point of view, I will read it, that more than 70% of respondents say that modernization is already on the way in the companies, and this is a huge increase to previous times. Nearly 94% have either launched or planned initiatives to modernize their operations, and 91% expect AI to reshape trading operations.

It's quite impressive numbers. Also, what is the driver of modernization from the technology perspective now? Not from functional point of view. It's adaptability, which you already mentioned as the main requirements because of the changing markets, but also scalability. 20% of respondents followed by business efficiency and business growth the main drivers is why people are looking for different solutions in the ETRM space. And also 43% of traders say that scalability is priority number one among their requirements.

So that's well, very large place for technology is a driver of the modernization progress, and I would like Sameer to comment more on this report and just really express how you see the modernization and how you follow it.

16:42

Sameer Soleja Yeah, I mean, to your point on cloud being the default. You know, I'm probably in the wrong place to comment on that simply because, you know, I probably, we probably have selection bias. If somebody comes to us, they come to us because we have cloud technology and nothing else. And so, but I'm glad to hear that, that, you know, I was glad to see that in the report.

I think a lot of what we are seeing is people who've invested in legacy systems or other systems, that, you know, took a couple years to get set up just right and work just fine for say, hydrocarbon portfolios or something like that. And then a new renewables desk comes on and system's not designed for that. And that's, that's a lot of the time when they'll call us. And I think what's new over the last 10 years is people are willing to run multiple systems because they're not trying to report out of multiple systems, they'll pump all that data to a centralized source and then and then report on the combined data.

So in a way that actually buys down the difficulty for the system vendor, all you gotta do is make sure to, to model a portfolio right. However, on the renewable side, there's a lot more data. I know the first time we tried to book a multi-decade long sub hourly settled PPA created like half a billion rows of data and that was re and so either the system has to be very scalable or has to have smarts to enable the scale.

18:13

Irina Reitgruber And how express, how do you see that technological drivers for software modernization? Who wants to, to be first, Arne?

18:23

Dr. Arne Weber Yeah, I think, I think it's, you know coming from the business side of things, it's kind of, I think, I think there's, there's two things that can drive the technology angle. One is really technology driven.

People going into the old school markets, so to speak. I mean, where, you know. I mean, we, we all need know these newer intraday or short term focused companies, very tech driven, very focused on rapid execution. So they really shook up the market big time. Hedge funds, of course, they come from the financial derivative contractors, super liquid. They use very different technology to what I would say the commodity world was used to. So that's one thing that is, is changing it.

We also have pesky tasks to take care of. I mean, then we are looking at, you know, more the, the downstream of the systems. So, so confirmations you know, back in the days, a human being had to look at everything and, and check and, and reconcile, and of course all that stuff is unnecessary headcounts costs. So, I mean, with more and more deals that we do, as a consequence, I would say, of use of more technology in the markets now, it becomes very hard to reconcile the systems or the trade confirmation. I think it helps everywhere along the way.

And then kind of coming back to my original paradigm. Okay, so then I wanna have a system that adapts quickly. I have people coming in every day that have a different background, different training. They bring new tools with them, and then it's the question of how do I get interfaces to my data? How can I quickly access it, analyze it? I mean, I remember working out of systems like Sameer said, you know, one big blob of a system, you had to know the functionality. There was a lot of, I would say, domain knowledge in the system required to do your analytics. I think those days, in my opinion, are over it's, this is, this is. It's not scalable enough. You do the next thing and ah, it's too hard to to implement that and it will cost you millions to adapt your system.

Reality is, and that's why spreadsheets are so successful, they're super flexible and that's why everyone uses spreadsheets. So the first thing I always ask is, can I whack the data into a spreadsheet rather quickly, look at it and then decide on whether it's worth spending more time on it. So long story short, flexibility and then there's a lot of tech driven stuff happening on the side, which kind of forces us to work on the, on the tech angle.

20:33

Irina Reitgruber To support your point, I was participating recently in a webinar where someone said that if you want to attract new professionals to your business, you need to apply more technologies. You need to apply some AI, you need somewhere to have, you know natural language prompts. And so people are expected and the success of the business depend on how good you are in respect to technology. It was interesting idea. I did not ask Andy yet.

21:02

Andy McMillan Yeah, I'm, I, I echo everything Arne just said around. I think it's surprising a lot of the clients that we work with are kind of sitting on legacy ETRM applications. And the more you dig into it, the more you see huge aspects of this business are just run off of Microsoft Excel. Where workarounds are kind of like found or different approaches are found to, to get the exact answers just because it's quicker and it might not be perfect, but it gives a better answer than working through a kind of like legacy ETRM, kind of like IT landscape where we've seen some of that kind of, where they're kind of non-cloud native. On that, that kind of the trade example, you're right, you've got multi-year PPA, they've got five or six different financial legs on the PPA that have got a different data series you need for each leg of like, you're suddenly into like a lot of data that if you are running in through your kind of a risk reporting on a, on a really frequent basis and trying to move to anywhere, like near real time risk reporting, then you're gonna need significant kind of firepower to, to run that.

And we see challenges on that when we're looking in at the legacy ETRM place. And then on the AI front around technology, I think, the, you know, the whole industry has gone AI crazy and AI's gone really wide and, and gone for a huge number of use cases where we've act, where we've actually seen it applied and created value is maybe more on the, the smaller instances at first, at the moment.

So, PPA clause extraction standardizing some risk reporting. Tied some confirmations type, like when it's more like focused and, and where, as opposed to trying to do like huge, like industry-wide application. Then we've seen kind of, we've, we've struggled to see some of those returns on AI just yet, but it's a huge disruptor, but it's still finding its feet in terms of the best applications and that will differ across different industries and different companies.

22:48

Dr. Arne Weber I think if you go further down the value chain and use AI for that, everyone would love to do that because that's kind of the more exciting stuff. The problem is then you know, the logic of corporates is kind of making it really hard. And actually, personal risk, right? I mean, I'm not gonna bet money and bet my book on something that I don't really understand, and I was like, well, some random black box told me that's what I need to do now in the market, I would never do that. I mean, that's—

23:12

Irina Reitgruber That's, that's a not the areas where AI, first of all, is usually applied. So we created a report in the beginning of the year analyzing different use cases. And I mean, besides obvious application and forecasting where you use machine learning, you have the automation of back office processes, operational processes. So this robotic process automation is really gaining bit attention. So large attention. And we have even introduced new software particularly, there are a few companies which are doing back office and operational automation across different systems, not particularly within the ETRM. That's very interesting topic.

But this is not what we are going to discuss now. So let's go back to the ETRMs and see how different ETRM vendors should address the challenges which we have already mentioned. And first of all, I would like to have another set of polling questions show up because that's exactly about ETRM challenges.

As long as the polling questions I hear on the screen, I can provide a couple of more points from the report we have. And so nearly one of the three survey respondents say that their current ETRM does not support everything they trade. And 25% said that operational risk is the biggest blocker to modernization. So companies have concern that the usual workflows, they are processes, established processes can be jeopardized by introducing new solutions.

And what else? There are 38% of their respondents say that the main driver of decision on ETRM/CTRM is the need to replace outdated system, which is different from previous times where a lot was just replacing some spreadsheets and then similar homemade solutions. So this is, there some information from the same report as we discussed before. And generally, so the challenge is what we usually hear on the market is interoperability due to the focus on ecosystems of tools. So what we see now that the modern newcomers into ETRM world, so called open ETRM/CTRM, this is the system which is based on what the technology, but still miss certain functional deepness, which is usual for legacy ETRMs. And with open APIs, with good connectivity and cross-industry collaboration, this functional circumstances can be improved. And due to the best of breed approach, companies gain more and more different solutions within the ecosystem, and so interoperability becomes a problem.

Another one is the user experience where visualization becomes extremely important and visualization also in terms of certain AI support. Then, as we already mentioned, agility and flexibility to cover different kind of requirements and to cover all this changing market condition, cloud nativeness, and finally, ability to support certain trading, EMOs, certificates, risk management, and so on.

So this are the issues which ETRMs need to address. And I think, when we get the results of the poll. Do we already have them?

Yeah, we see how people value that. So legacy software, outdated architecture, it's one of the biggest challenges as expected, followed by interoperability. If you have modern architecture solutions and integration of ETRM in the local market. We discussed those regional challenges. And then on the place, number four, missing functionality to model certificates, PPAs, and so on. So then let's discuss this. Maybe we'll start with comment from the user perspective. So how do you see this issue?

28:09

Dr. Arne Weber I guess user perspective. I think that's, it's kind of what I expect to be honest. Well, of course we have a lot of, I would say, legacy software, legacy systems out there. There's been a lot of evolution in technology, so of course a need to upgrade and if you can do that in such a fashion that hopefully next time you do an iteration or there's a, well, if you can do the independent from hardware, so to speak, that's of course something that is very attractive and I think that pays into solutions your cloud base and so on and so forth.

So that's one. And it kind of also emphasize the point, at least that's how I interpret the result here, we do new business, the markets are evolving. We're doing new deals, new structures. So ultimately we need something that captures new business and that, that is, I would say the second big requirement.

And, and honestly, from a use perspective, you don't want to care about the tech stuff as little as possible and be fully focused on the new stuff. And like Sameer said, then you shoot trades into the system and hopefully it somehow is able to digest them. And that is a question, how flexible is it? How much do you need to bend it, or do you end up, like back in the day, it's just booking cash flows and hoping for the best and not actually capture the risk of the position accurately, but accurately, but just somehow while doing your duty as a trader. And I think the truth is always somewhere in between, right? I mean, the system won't ever be there to fully do what you need from the business side, but how flexible, how quickly can you adapt? I think that's, to me, that's always the question.

29:32

Irina Reitgruber Then I will ask Andy as next working on the consulting side, you face different customers with legacy system, with modern systems. So how do you see this issue?

29:44

Andy McMillan Firstly, Arne, please don't book cash flows, 'cause that's, I spent a lot of my life unpicking those and finding ways to rebook them.

29:50

Dr. Arne Weber That's Sameer's business model, to remove systems that only do that.

29:54

Andy McMillan Yes, it's ours as well, right? We go into a lot of a lot of clients and, and it's unpicking those cash flows because the ETRM isn't able to do what they ultimately want it to do. So there's a real reason why traders do that. It's not that they just don't follow the rules, although sometimes it could be that.

Well certainly from a legacy ETRM kind of perspective, we hear from clients all the time around the UX. It's not maybe where you'd expect it to be in a kind of modern application. That is probably being stretched even further based on the volume of data that's being fed into it. The different visualizations you need to do, the different risk reports that you want to run. That's just further stretching it even more.

Around PPA modeling, like I mentioned, the kind of a prime use case of AI. We are seeing that across nearly all of our clients in terms of the PPAs are getting more and more complex. The legacy ETRMs are struggling to then handle those. There's different ways that you can model those, people booking them in different ways. They're understanding like, you know, the way that we are all global organizations. You might be agreeing a PPA and it's written in German, and the person modeling your PPA could be a English native speaker or Romanian or something very different. And there's a lot of very industry specific terms.

Things get missed and, and mis-booked all the time. And the big, and probably the biggest one overall, I'd say is around the integration layer between your ETRM and kind of any of your downstream systems. So if you go into any of our kind of legacy clients and look at that architecture map, the map is mind-blowingly huge with all of these different systems that have just been brought in to do one hyper-focused kind of task at one point in time.

But we need to get that into the ETRM as quickly as possible now, and we need something that can calculate a data series that excludes negative prices, that can add two series together and create a new hybrid series and send that on a, like a real time basis too. So we, it's, it's being more and more exposed, I guess, by the, the challenges of the, the trading landscape now. So yeah, ETRMs have a huge kind of extra push there to meet the requirements then of the trading landscape.

31:49

Sameer Soleja Yeah, and so we thought we had a fancy solution for this and we had a really nifty formula price builder with a language we had developed ourselves. And you know, upon onboarding a couple of new PPA focused customers recently, just threw our hands up in the air.

We're like, okay, well this language is never going to keep up with all this weird shit. And so we actually decided to build a Python interpreter into the system and say, okay, fine. You know what? You have a weird PPA or an odd pricing model, throw it into an LLM. Have it generate a pricer for you. Attach it to the trade, be done with it. We'll run it going forward. It's been an interesting solution an interesting journey to that solution. But it seems like a really flexible one.

32:32

Irina Reitgruber Sameer, I have some more questions for you as a software vendor. So in many recent conversation there turn out to be kind of open ETRM concept. What is open ETRM? It's modern architecture. It's quite basic functionality, not very deep in functionality. It's priced based on different metrics related to portfolio size rather than number of units. It's multi-tenant, which has huge cost advantages. And it based on ease integration of some custom tools on different solutions so that you can easily build ecosystem of the solution. So that's basically how we defined this open ETRM. So is that your vision as well? Is that what Molecule is heading to, and what is generally your notion about future ETRM system? How do you see it?

Also, with respect to all the flexibilities all the complexities of current market and what we mentioned before. So the changes on the market are very rapid, so you need to be very agile and adaptable for different market requirements. So what is your view for the future?

33:48

Sameer Soleja Yeah, sure. So, I mean, to your point about being agile and flexible. Molecule supports 50 commodities, something like half a million different instruments, and you know, we've had to design it for flexibility, which is resulted in trade-offs. Like some things get more simple in order to be able to scale. But one of the choices we made many years ago was exactly what you described which was having an open architecture. So basically, we have a proper normal restful API. Customizations are built outside the API. We guarantee that the API always produces the same thing across versions so that people can be comfortable allowing us to upgrade every two weeks or every four weeks. I think that's really important and where we have been pushed from there is to make time series data easier to use, to pump data to a customer's data lake, to publish our models, et cetera, like that. That's all stuff we do, and that's and I would hope to see everybody else do it too.

34:54

Dr. Arne Weber Now, a cheeky question from the use side, if you, if you have all these instruments and all these different products. And you adapt something to one client, how do you ensure that this doesn't, somewhere down the line, you know, the domino falls somewhere else and all of a sudden, you know, client number 101 has a problem with his markings or whatever, just because he changed a tiny setting. I mean, that requires huge oversight. So how do you manage that? Because the complexity grows, right? As with all these products.

35:22

Sameer Soleja It does. And, and so two things really. One, obviously, automated regression testing, but two, we try not to change things in the calculations. We try to add new calculations instead so that some customer has some particular way of doing something. Either we'll tell them actually we think you're wrong because everybody else does it this way or, sure, we'll add a field, we'll add a field to the output and we'll always produce that as well and nobody else changes.

35:51

Irina Reitgruber Anything else before we go to the questions from audience? Do we want to give a kind of final statement to, to this discussion? Andy, maybe this time I will start with you.

36:06

Andy McMillan I think we've covered it briefly so far and it was interesting, Sameer, to hear your kind of history around the kind of like data series, preparation tool you've got. I think that's something that I'm particularly interested in seeing kind of the market try and resolve a little bit more around, 'cause that's only grow as CFD schemes evolve as different, different subsidy schemes evolve, like the German EEG model. As PPAs become more complex. The use cases we've got on this is just becoming more and more complicated.

And so having a good tool to understand how that works and to give the flexibility to create new ones would be really important. I think maybe I'm saying this because it's very front of mind, 'cause I spent the last week trying to unpick a bunch of trades that have been booked use using the wrong data series. Where, and that's kinda, you know, a business user is kind of trying to book trades to the best of their knowledge. They're using a series which they think is kind of right, but it's subtly different. And there's documentation that they've got to enable that is, is quite challenging. So having a kind of fully functional kind of tool to support, or as part of your ETRM would, it is something I'm particularly interested in seeing. So yeah, Sameer, that would be great if we could have something else like that.

37:19

Dr. Arne Weber I think some of my personal bias was confirmed by the discussion. So front office can breathe easy. We're not gonna get replaced by AI just yet. We have a few more years left, I guess. Let's keep it that way. And there's a lot to be done right with, with technology.

I mean, all joking aside, there's growing challenges. There's a whole new set of systems out there. And I would say the rate of growth and the growth of complexity is ever increasing. So that means there's still lots to be done.

37:46

Irina Reitgruber Thank you.

Shall we then go to the questions we have already some questions from the public. The first one is for American market specialist. Will the new FERC rules for fast interconnection have risk effects? Probably it is aimed for Sameer.

38:08

Sameer Soleja I am a tech guy man.

38:13

Irina Reitgruber Okay. But if the question can be also answered in writing, if somebody has that knowledge. The next one. What features should a modern ETRM system prioritize to stay relevant in fast moving renewable markets. I think we were talking a lot about that. Maybe someone wants to give very short, summarize, answer here, otherwise the whole webinar was actually mentioning this I guess.

38:42

Andy McMillan Yeah. That's, that's essentially what we've covered so far. Right. I guess some of the key things for me is around focusing on the modern UX that enables people to navigate the system, data series kind of preparation, consume, and then being consumed by the ETRM. Support for traders in the instruments and tooling that they've got available to be used, and battery energy storage building more complex models and how they're gonna be modeled in the system. I think they're probably the ones that are front of most front of mind for me, and the ones that we kind of see most often from a lot of our clients.

39:15

Sameer Soleja Irina asked a question around future modern systems. I wonder, and then this is something that I've been pondering for a while, if a future open modern system doesn't even bother with a UI and exposes its data in a way that an LLM can consume it and, you know, done.

39:30

Dr. Arne Weber I would echo that. I mean, it is, it is radical, but I like that because it's end of the day, this is, I mean the second you've got your interfaces set up, the second it looks like home again, I don't open any. Actually, that's the target setup for me. I don't want to even know what my password is.

39:46

Irina Reitgruber That's a hope of everybody. But there are a lot of risk related to this because you must be sure that the machine understood your sentences, your prompts, exactly. You must be sure that what you get back is really trustable. So this is a set of questions which people are dealing with and industry will resolve this, I'm sure, but at the moment, concerns from this still prevail.

However, I also believe that this is the future because everybody is using ChatGPT similar LLMs in our real life. So why not in the business? Yeah. That would be the expectation.

There is interesting question about forecasting models and I think that's not only forecasting, it's also about optimization, for example, for batteries. So how do you see forecasting models evolve for intermittent renewables, especially the batteries and prosumer aggregations are becoming more common. And maybe I will add from my side, so how do you see this models for ,say, better optimization? Is it something bespoke for the businesses or there will be some standard ways to model that?

41:07

Dr. Arne Weber Well, I would say if you think about forecasting of batteries or weather or whatever it is, I think the trend of where this is really going, in my opinion is that there's gonna be specialist systems for all of these activities. And almost to our discussion earlier on, I don't know where these specialists are. There's gonna be, well, we see companies that specialize in whether there's companies that now specialize in battery electric assets and optimization there. Experts on certain grid phenomena that you can trade if permitted by the local regulation. So there's more and more of this sort of super specialized firms. And same for, you know, decentral power and decentral dispatch of flexibility, that pool flexibility from end customers, household retail. There's loads and loads happening. And then there's small specialist firms, or actually big, some of them are now big specialist firms or somewhere in between, right?

And then the question is, well, they all need to somehow tie back together at some point in time and, and unfortunately, ETRMs or risk management systems then have to give you the full view as an integrated energy company like MB Energy, we have to deal with that. I mean, there's a highly complex small scale sort of stuff cooking. At the end of the day, we still need to make it a company report and have a company view on risk. So somehow that all needs to channeled and funneled. And then actually, I think part of the magic is that you need to cut the data also. The final risk report doesn't need to know every small position of every tiny household. You don't need to see that 24/7. Once a day is plenty, right? Because that's not gonna change much. A PPA, you cannot get in and out of without negotiations of a contract. So don't care whether I have to mark to market on it every five seconds because it's totally irrelevant. Can't even trade it, right?

So, my point being, there's a lot of ecosystem left and right of us, and somewhere we gotta make cuts to the data somewhere and we gotta make cuts to our business activities and bring them all back home again to something meaningful on a company level. I think that's typically where the ETRMs come into play because they have to have this company view.

43:03

Irina Reitgruber And from the ETRM perspective, Sameer, do you see this kind of things like, you know, forecasting of production, better optimization, do you see it as a part of ETRM or that should be some third party valuation models, so third party risk solution so you need just to provide integration to those?

43:23

Sameer Soleja I mean, I think today we provide integrations to those. But what we see, I guess in the economics and market dynamics of the ETRM software market is that companies are trying to bring that in-house and under one umbrella. I don't think anybody's done it. Well, only a few have done it successfully yet. And, and certainly very few modern folks have done it successfully, but I think that's ultimately where it will be and where the sort of investor capital will lead us.

43:58

Andy McMillan We see similar with our clients. I think a lot of the clients I've worked with are building battery energy storage at a huge rate, but the typical way of marketing that, really that we see is, is via kind of specialist right now as they're still working out the commercialization strategies as how they want to operate them internally. That probably has a good linkage with ETRMs or the other kind of supporting software that's available to support it. But it's definitely in an exploration and kind of a find out stage right now. Where there's real opportunity to kind of do this better than other people because everyone's still working it out in terms of the right way to kind of stack the revenue that's available through these assets.

So I know that's a really sort of cop out answer, but I guess that's the kind of, that's what we see in the market at the moment.

44:40

Dr. Arne Weber And I think if, you know, if you look at the market players, how they evolve. I mean, typically as they start with some sort of new idea, I mean, it's either pooling flexibility and then, okay, they sell the service, the platform, and that's one business model.

But then the question is, of course, do you wanna sell it as a service or at some point do you think well actually, we can attract enough capital to take risk on these, on these activities. And then boom, you have a trading house, right? So it's a bit of a question of where do you want to go with it. I mean, you can stay in the weather forecasting for corner to stay with this example, or you can say, I have a superior weather model and I can trade this weather model intraday, and then I'm just gonna not sell the model anymore, but I'm just going to express my view in the market and make a buck there.

And that's what we see a lot these days, right? I mean, we see that these players push into "our" markets, I mean the power markets and they're not ours anymore. There's now investment banks in these markets and Okay, so they kind of play in the same game. That's the interesting part, right? These people push into, or these new players push into the market, they do things differently, and then they force the old kids on the block to do things differently too, which is, I guess, why these markets evolve.

45:46

Irina Reitgruber Okay. Yeah. Thank you. There is, I will jump to some other question, which is really, sounds very interesting for me.

So if you compare European and those American markets, European market for renewable certificate trading is very advanced, continuous intraday day market, ancillary service market, the US market is very fragmented. Does it mean that European ETRM vendor had it easier to enter North American market than vice versa? That probably something for Sameer as you are working on both markets.

46:21

Sameer Soleja Yeah, I would say at least from our perspective, the inverse has been true. It is hard for a system to enter another market, but the advantage that we have had starting in the American market is that we had to deal with six different ISOs right off the bat and 20 different power blocks right off the bat, and all sorts of really weird complexity that can then be distilled into easier time blocks and a few more integrations, et cetera. So for us, it's been easier to go that way. But you know, definitely curious to hear what others think.

46:55

Irina Reitgruber Yeah. Andy, what is, from your perspective, you see different system, different vendors?

47:01

Andy McMillan Yeah, I mean, I'd probably agree with Sameer on that one. I think the project I've been on for the last few years, we've been doing a full kind of Excel driven kind of certificates business establishing in a legacy ETRM. And the rollout we took for that started in the American market for exactly the reasons that Sameer said.

So it drives that complexity. It drove out all of the kind of different ways of working. It made you have to integrate with all the different registries, understand how you're pushing down your revenue across multiple different assets. We've now taken all those learnings and we're rolling it out across central Europe and UK and we see that then, okay, the UK do it completely differently because you've got the different kind of like factors that you need to apply and so it fully stretches your rollout.

I think you could make a case for focusing either one to be more complex depending on where you start and maybe what markets you know better. But certainly it's the key challenge, not just for certificates, but across kind of general like all power trading is, all markets have their own differences.

You can't just have one cookie cutter kind of approach across all. And there's regional differences that we see. But yeah, North America is, has been challenging, but has driven a lot of learnings around kind of building a certificates business.

48:07

Irina Reitgruber I had the same impression from my former jobs and different software vendors. Arne, do you want to add something here or should I go to the last question?

48:17

Dr. Arne Weber I think, I think maybe I can, alway always throw in the user angle and that's, we run a global company. We are active in the US, in APAC, in Europe. So for us, we need a system that works across regional, across borders. Sounds easier said than done.

48:33

Irina Reitgruber Easy to say, huh?

48:34

Dr. Arne Weber Yeah, exactly. It's a very easy requirement to write down. It's very hard to bring it to life, but on the business side, we are quite capable of, of course, we know how to do this. We have our entity set up and so on and so forth. Our client base is there in their respective countries that really, the system then mustn't stop us. That's the only requirement.

48:51

Irina Reitgruber You use the same ETRM system across all the regions where you are active in, right?

48:57

Dr. Arne Weber Yeah.

48:57

Irina Reitgruber That's good. Okay, so we have still some questions unanswered. I can maybe we take a short time to answer the last one for today. For portfolios with both centralized and distributed generation, where do you see the biggest data quality gap today? Who volunteer to answer?

49:21

Andy McMillan Yeah. I can talk about what we've seen. I think it's a huge barrier to, to building a lot of the processes that you want to. So for example we've worked with a lot of quite old wind parks. Some of those wind parks have very different kind of technology around them that are built that are kind of different stages of the life cycle. So obviously some wind parks we work with are brand new and some of them we work with are reaching the kind of decommissioning period and some of those standards for connecting to those wind parks are very, very different. There's lots of point to point kind of integration on some elements of it. There's lots of, there's real kind of workarounds to enable it.

So a significant part of our kind of work was focusing on that data quality and data availability upfront that required huge amounts of data, kind of ingestion, data storage, data manipulation. So you have kind of huge amounts of work just to get you to your kind of starting point of then being able to build that into the system. So I guess, those kind of connection standards, I'd just say don't underestimate it and don't make it too lightly into your kind of project plan because it takes significant amount of time and effort to establish those connections to all those different disparate sources.

50:26

Irina Reitgruber Thank you. Someone else want to add something?

If not I guess, I mean we have two questions left, but with the last few minutes, I know if you want to have very, very short answer, how much pressure are you feeling to consolidate power certificates to a single view of exposure? And I would expect it's important, but I ask you to confirm, you know. It's a company exposure, right?

50:59

Andy McMillan I can go first again, more and more especially as more and more PPAs are kind of bundled, so your position is converging more and more. Typically, a lot of the clients still have very separate desks and manage that view differently, but it's coming closer and closer together.

51:15

Sameer Soleja Yeah. We've definitely seen a lot of folks, sorry.

51:20

Irina Reitgruber I thought that, I expect that Molecule also support that.

51:24

Sameer Soleja Yeah.

51:24

Irina Reitgruber Right.

51:25

Sameer Soleja I think, when we interact with a customer for the first time, I would say three out of four times, managing certificates is almost like a scheduling job and rather than an optimization job. But I think every so often, we'll run across a really forward thinking organization who finds that there's alpha or maybe beta to be found and be in the certificate optimization as well.

51:49

Irina Reitgruber And Arne, are you in your solutions have the same system supporting certificates trading, power trading? Can you have consolidating view, I guess?

52:00

Dr. Arne Weber Of course. I mean, that's, that's always the target, right? I mean, have everything in one system, one consolidated view. Otherwise, you're forced to consolidate outside your leading system. So, I mean, I would say textbook answer is all in one. Reality not always like that, but okay. Until Sameer has fixed everything.

52:23

Sameer Soleja Amazing.

52:27

Irina Reitgruber Okay. You see we have even a question about quantum computing on the roadmap, but I'm not sure if there are some other points on the roadmap which are more relevant at the moment.

52:40

Sameer Soleja I mean certainly for optimization that, that would be the right place to, to use it. You know, the new tech advancements that we're really focused on are things like an MCP server to connect to LLMs.

52:53

Irina Reitgruber Mm-hmm. Okay. So that's interesting.

52:57

Andy McMillan Yeah, with our client base, we see a huge runway on AI machine learning. Expanding LLM usage, as Sameer just said there, there's, there's a huge runway of opportunity there. Quantum maybe has a little bit way to go until it's kind of in that kind of commercially commercial landscape just yet.

53:16

Irina Reitgruber With AI development, it will come sometime, hopefully.

Okay. So thank you very much. We're the last minute of our conversation. Want to thank all participants. It was very interesting and yeah, have a nice day and a lot of success in your businesses.

53:34

Sameer Soleja Thank you. Thank you. Thank you everybody.

53:35

Andy McMillan Thanks everyone.

53:36

Sameer Soleja Take care. Bye

53:37

Irina Reitgruber Bye.