The Power Problem: Why Legacy ETRMs Struggle With Modern Power Trading
Molecule's Founder and President, Sameer Soleja, and Senior Sales Engineer Barry Lowe examine why battery storage, PPAs, and shaped power products keep breaking the assumptions built into legacy ETRMs. Watch this webinar to see why these gaps trace back to architecture, not configuration.
September 2026 | 22:52
Summary Keywordslegacy ETRM limitations, battery energy storage systems, BESS trading, power purchase agreements, PPA modeling, ETRM architecture, renewable energy trading, power market complexity, energy trading technology, trade capture, settlement data, optionality modeling, energy risk management
Transcript
Sameer SolejaPresident and Founder, Molecule
Barry LoweSenior Sales Engineer, Molecule
Tim Binnington (Moderator)Marketing at Molecule
Tim Binnington Thank you for joining us. Power markets have changed. Higher renewable penetration, 15-minute settlement intervals, battery storage, and renewable attributes have introduced trading realities that many ETRMs were never originally designed to support. In this webinar, Molecule's founder and president, Sameer, and our sales engineer, Barry, will examine why these challenges often trace back to architectural assumptions rather than configuration, and share the patterns that they've seen across power and renewables trading operations.
Tim Binnington So, gentlemen, thanks for joining us. First question really is, what's the difference between an ETRM and an analytics system in terms of how they see the world, and what is each better for? Sameer, what are your thoughts?
Sameer Soleja I mean, basically one is stochastic, one is not. An analytics system speaks in probabilities. An ETRM system is a system of record. It's for predictable daily reporting. It's meant to be right and always right. Analytics is more for poking and prodding at the thing, so stochastic data makes more sense.
Tim Binnington Okay. And how do you see it, Barry?
Barry Lowe Yeah, exactly right. I kind of see the ETRM as the book of records. Like, where do the trades live? How are they priced? What's your daily workflow around those? So that as Sameer said, right, we can create a standardized set of reporting for the team to actively manage their positions.
Sameer Soleja And so the way they bridge, we will usually see people choose, okay, the ETRM system will own the P95 or the P50, and then all of the system of record lives with that information and downstream of that. That's probably what you see Barry too?
Barry Lowe Exactly right. Yep. And data flows in and then we kind of like the hub and spoke out to the downstream systems for reporting.
Tim Binnington And so how does that interact with other systems that people might be running?
Sameer Soleja I mean, I think the analytics system is the odd one out. It's the one that speaks in probability, and it's vitally important. It's how you see the future. It's how you guess at what the future could be. But everything else really speaks system of record as well. The ETRM manages the transactions through their life cycle, produces a predictable daily position in P&L. It feeds the ERP, which produces invoices, and the other systems that may be downstream of that take that information from there.
Tim Binnington Okay. So if we look more broadly, as I said, we've seen a lot of changes within the energy markets. One particular area that we've seen massive growth in is BESS trading and what makes them challenging to manage, particularly with legacy ETRM systems. We're hearing a lot of problems and issues coming about that. What's your take on that, Barry?
Barry Lowe Yeah, I'm primarily seeing it from the presentation side, like given the use cases and modeling that inside the applications. In my mind or the experience I have, right, it's more just everyone's doing it differently. Everybody wants to push something different into the application, into the ETRM, and then get something different out of it into the reporting layer.
Barry Lowe So to me, it's just the variety. There's no standard way everybody's viewing these because everybody's treating it differently, and especially if you're a battery optimizer, right? We're really just taking that data post-decision making and really just picking up the transactions that flow into and out of the battery.
Sameer Soleja And that's sort of a reflection of what's happening in the market as well, right? People co-locate batteries behind the meter with an asset, or they'll put them in front of the meter for some regulatory reason. Or they'll own a battery just by itself. And so all of these reflect in the type of trading that we see happen. It could be that the battery's modeled as a PPA, it could be modeled as a tolling agreement. It could just be post hoc as day-ahead awards flowing into the system. I've heard a lot about people wanting to write TB or TBA derivatives. Haven't seen a lot in action. But that would be another one where it's a derivative between the top four hours of the day and the bottom four hours of the day.
Tim Binnington Okay. So looking more broadly at this, we're looking at, say, the volume of data we're seeing with PPAs, because as I understand it, they typically fall outside the core trade model on most platforms. How does that fit in with this arrangement? Sameer, what are your thoughts on this?
Sameer Soleja Sure. I think PPAs are generally more complicated than ordinary trades that an ETRM sees, partly because of the volume, partly because of prices, either forecasted prices primarily—
Tim Binnington Mm-hmm.
Sameer Soleja ...and partly because of settlement intervals that are typically very fine, like could be 15-minute intervals or five-minute intervals or hourly intervals. You combine all of that, fine settlement intervals with predictions of prices out into the future and extreme term, say 20 years on a PPA, you have a trade that could easily take 35 gigabytes of data over its lifetime, and that's one trade. And so also typically pricing and volumes are higher complexity, so it's not just like price times quantity usually.
Sameer Soleja Maybe more often in the US we see that, but in Europe and/or in emerging markets, we definitely see much, much higher complexity in pricing. Weirdly that latter bit is not all that different from, say, a gas plant allocation, which is a complicated topic on its own. So like a simple PPA might have an interface to a SCADA system for actuals and a stochastic model, P50s, say, for long-term high frequency data. That itself is difficult. It doesn't fit in an ETRM's trade model by itself. A complex one may be multiple settlement and resettlement streams, references to other trades, cumulative settlements. I don't know, Barry, whatever you would like to add as well.
Barry Lowe Yeah. I think it's like all the variation to the flavors. You know, when I first started seeing the very, like, either fixed rate or floating rate, and then they just had a shaped profile that was continually getting updated, like an as-gen. More recently, I've started seeing things that got these fixed float components where, you know, they'll take the first X percentage of volume coming off the generation unit, right? But only up to a cap will there be at a fixed price, then they go to a float. Then you start putting all these fees associated with them, right? And it just doesn't really sit naturally into just one transaction, right? You need all these components that take the volumes, the fixed fees, the floating rate fees, and to merge them together. And with SMSs, with the data, like, you know, long-term, short-term, different forecasts, calculating a BAB, using that for your forward valuation, and then building the forward curve associated with that particular generation unit and any attributes associated to it.
Tim Binnington But the complexity of the trade and the complexity of the pricing, is it just the sheer volume there, or is there something more fundamental about the nature of PPA that cause additional challenges?
Sameer Soleja It's both.
Tim Binnington Okay.
Sameer Soleja So extreme data and extreme complexity at very fine intervals. You know, having a conversation with a customer of ours recently about how to get a simple position out of a whole portfolio of these things, and I think from his perspective it was, "Well, can't we just do X and we will get a position?" And the answer is, yes, you can enter this as a series of, you know, as a monthly strip of power, and that's probably gonna get you the 80% answer, which maybe is what you need. But that's not gonna get you far enough for billing. That's not gonna get you the last 20% of the answer. It won't get you the optionality that might be baked into these agreements. It's a seductive temptation. You can enter these things in a monthly strip. Ultimately, you won't be fine.
Tim Binnington Are we gonna be seeing more people moving into this space as PPAs become a more dominant part of the energy trading environment? So it's gonna become rather than a good to have, it's gonna become essential or almost fundamental.
Sameer Soleja I mean, as power is starting to look more and more like retail at the corporate level, right? With data centers and/or tech companies buying shaped power or renewable power or all of the above, you know, we can see what retail agreements look like today. Free nights and weekends, except if you're charging your car. I would think that we can fully expect to see this evolve and become more important.
Tim Binnington Now I do have one little question, and this is just a little aside. You talk about the car, the idea of having the car to grid as a resource selling into the network. I know that's rolling out. There are people who are doing it now. Can we envisage that becoming a significant play within the trading landscape, or is it always gonna be a little bit of an edge case? And Barry, what do you think?
Barry Lowe I think it'll be an edge case, like—
Tim Binnington Okay.
Barry Lowe ...still my, if I had one and it was still my car, right, I still need to get from A to B tomorrow, so am I really gonna play the market? Probably not.
Tim Binnington But if you're plugged in at work or you're plugged in at home, and you've permissioned your supplier to do this, can we see this becoming an interesting area, or as, you know, or am I just going, yeah, there'll always be people who are willing to sell their fuel?
Barry Lowe I think end of day, right, you know, someone's gonna go ahead and do it. My brother would, I know that for certain. So I think it just comes down to how sophisticated is the homeowner, right, or car owner. Do they really wanna play that game, or it's just not in their wheelhouse, right?
Tim Binnington Okay.
Barry Lowe And then at the end of the day, how much volume's really out there of people who are willing to push back? And so maybe peak load—I mean—weather extremes, I could see it. But day to day, who knows?
Tim Binnington All right. Okay. Well, what we're seeing here is people are running into limitations. You know, we now need to think about what are the causes. You know, is it the system configuration, or is it something deeper around that volume and complexity of the data? Beyond the PPAs, across the board, what are we seeing?
Sameer Soleja It was interesting. So when we started Molecule 14 years ago we were in the early part of the 2010s, and a constraint that had been a problem for systems built before Molecule, which was data storage, wasn't a problem, right? It just didn't matter. Data storage was cheap.
Sameer Soleja Well, now if you have a single trade eating 35 gigs of data and you're doing hundreds of these, the systems have to evolve and, you know, I know for us, we've had to develop optimizations around it in order to show high resolution data in a deterministic way at the time it's needed, but also be fast. I think that's probably a challenge for anybody.
Tim Binnington Yeah, but is it something that we're going to see—power producers, traders having to make fundamental decisions about, you know, actually we can't afford to do it this way, or we've got to make an investment in our systems to get us to a position where we're able to act at the level that the market's demanding. I mean, Barry?
Barry Lowe I mean, end of the day, right, we have to act in the capacity that our clients, prospects need us to behave, right? So we have to be able to handle these large data sets and also this changing market and, you know, as the PPA structures change, be able to support those and quickly implement them as necessary.
Sameer Soleja I mean, I think the customer or user's choice is either take a rough approximation of your position and P&L or get it right. The choice of the latter means you must have a system. You just can't do it at scale without that, or you can throw hundreds of people at it.
Tim Binnington But I mean, the limitations of some existing systems, you know, is it about the data model, or is it something more to do with handling and enhancements? Or what's the challenge out there?
Sameer Soleja That's probably a better Barry answer. Like, I've been here for 14 years and, you know, all I know is secondhand through our customers. What I can say is that I know the things that we have done to optimize this, to make it fast and right all the time are unusual. And so I would suspect others struggle with it—
Tim Binnington Mm-hmm.
Sameer Soleja ...potentially more.
Barry Lowe Yeah, I think there's kind of a couple of things just from my experience, right? Number one is, is the solution built to be able to be easily extended and developed over time by the core application, by, you know, the technology group building out new functionality, adding flexibility to the products and instruments that they build. That's the first piece, right? And then the other thing that I think comes into play is also the consultants understanding how best to use the application. What was the intention of the software provider when they started going down this path? Like, where's their goal and endpoint? And making sure that those firms take that into consideration when they build those workflows, internal processes. And I think that's kind of two key pieces, right? The flexibility of the application to evolve over time, and then also having those relationships with those vendors, service providers, or even just their clients, and making sure that they really understand how the system was designed and built to be forward compatible with enhancements as they come along.
Tim Binnington Okay, which kind of leads us into the question of, so what does an ETRM actually need to get right, you know, architecturally, to be able to handle power today and the prediction of where the complexity of the market's taking us tomorrow? I mean, Sameer, what are your thoughts on that?
Sameer Soleja Extreme volume and tenor, and it needs to be able to optimize that for storage, throughput, determinism, and arbitrary query.
Tim Binnington Okay.
Barry Lowe Yeah. Yeah, all that, right. And then to me it's, like, making sure that it can evolve as the market changes. I think from my mind, that's a key piece, right? Sameer's more struggling with how to maintain all that data, and I'm more thinking about, okay, how do we handle this variation of something that's coming down the path, or we're starting to see rumors of this kind of maybe coming through. Is it really gonna take effect, and do we need to get ahead of it?
Tim Binnington Okay. Now, one thing I'd like to perhaps touch on is, more specifically at the moment, we've obviously seen over the last few years some significant geopolitical shocks to the energy trading environment. Things to do with Russia, and now things in the Middle East again. On both occasions, various significant parts of the energy infrastructure, be it oil, be it gas, be it imported LNG, be it the knock-on to the domestic energy markets. Are we seeing anything from our customers about needing new technology there or new challenges, or is this all fitting within the models that we can deliver today? Sameer?
Sameer Soleja I personally haven't seen hydrocarbons being traded in new ways other than perhaps LNG. It's usually power that has all the weird variation in it. Barry, I don't know if you have anything to add to that.
Barry Lowe No, I haven't seen anything unique coming through in those hydrocarbon space other than—
Tim Binnington So they're just handling the price hike in the same way, and there isn't a sophisticated alteration. But then again, the knock-on effect this is gonna have, particularly into this winter in the European market and potentially the North American market, the pricing impact and maybe new participants in the market coming on stream, either renewable stuff, projects coming on stream, things that people have been working on or people suddenly getting in because frankly, there is... it's an active market. Do we think there's anything lurking around the corner that's gonna give anybody any surprises? I mean, I know we're doing futurology here, but—
Barry Lowe Yeah. That, I think that's a good question. I'd never really thought about it before until then, is maybe, you know, given everything going on around the globe, maybe there is some new, like European nat gas denominated products coming through in terms of maybe different pricing indices or some sort of like complexity to it. I think that might possibly be coming down in the future, but I haven't seen anything of that. But that probably the only one I can think of and maybe something, some new benchmark for crude, but—
Tim Binnington Okay. Sameer—
Sameer Soleja Yeah, I mean, I was gonna say maybe we'll see the frequencies converge.
Barry Lowe Yeah.
Sameer Soleja Because then you can track, you know, a molecule of oil all the way through to the charge it produces on the grid. You know, there's... I could see that converging. I would think that pricing could become a little bit more standardized. Typically in emerging markets you'll see it be less so, and as the market matures you'll see more standardization. But I would bet, if anything, on the standardization being at high frequency, possibly extreme tenor.
Tim Binnington Okay. Earlier you mentioned data centers... and obviously this is a political thing, but it's a reality. There are huge demands on supply market and interestingly, new variations on generation, 'cause we're seeing people not only buying power, but trying to find ways of supplying energy on their own sites. So we've got people strapping jet engines onto stuff, you know, because they're in various places. We've got demands for power from the main grid. Do we see anything interesting coming there that's gonna provide a shift or an impact on trading volumes or technology do you think?
Sameer Soleja You mean before or after the bubble pops?
Tim Binnington Well, whilst the bubble is inflating... people are having to adapt to that new reality. So yeah, either way. Give us now and after.
Sameer Soleja Barry sounded like he had—
Barry Lowe So I think yeah, like I'm not gonna go too deep into that. I think the big manufacturers, you know, retail providers possibly might be getting more into the energy space. You know, as you said, like maybe they become something where they start putting generation units at their facilities to be carbon credits or, I don't know, some aspect of it. So maybe we'll start seeing kinda like people downstream needing ETRMs because they start to trade more structured products than just getting a fixed price contract from the power supplier. So then they'll start doing more hedging, start building, you know, purchasing RECs or GoOs to go along with their energy. So that to me might be interesting to see how that plays out over time. But in terms of, you know, the variety of how to produce power is kind of like outside my wheelhouse.
Sameer Soleja Yeah. I mean, I think the promise of data centers in years past has not been for base load consumption. It was for interruptible consumption. And so perhaps as we see models get more efficient, we'll see data centers needing to respond in more real-time to signals to, you know, drop load or to shed load or to ramp back up. That might drive sophistication as well. I think a lot of people had intended that that was going to be the case over the last few years, but it hasn't been. It's just the data center's running all out.
Tim Binnington Okay. Well, gentlemen, thank you very much for your time today. Hope everyone who's listened has found this interesting, and obviously if you've got any questions, just you can drop them to us. Thank you so much.