Building software to increase energy systems flexibility, with Nicolas Höning, Co-Founder of Seita - Episode 13

This is the transcript of our latest podcast episode - listen to it in the link below

My guest today on Luz Climate Tech Podcast is Nicolas Höning from Seita. Seita is a company based in the Netherlands. And Nick is a developer by background. Seita aims to help companies bring more energy flexibility to their energy systems by developing an open source software that takes data feeds from weather patterns, from energy prices, and from energy consuming assets to produce a schedule saying what is best to consume electricity in the following day to minimize your CO2 footprints and your costs.

In this episode, we're going to talk about the data infrastructure in which Seita are bringing their solutions and the challenges that cities are encountering and addressing to roll out their solutions, as well as the ecosystem of partners that they team up with to bring their solution to market. It's a fascinating discussion at a very detailed level on a space that is critical to make our energy systems more flexible and reduce their CO2 consumption and therefore facilitate the energy transition.

I hope you enjoyed my conversation with Nicolas Höning.

Nick, to get us started, I'd love to hear more about your journey and the fields you've been interested in and how that led you to starting Seita.

Yes. Well, I'm already 40, so this is a journey that has been a bit long already. I've been programming for over 20 years now, web applications, the occasional data science, machine learning, things in between. But also I've been busy with the energy transition for more than I think it has been 13 years now. I started my PhD here in Amsterdam, the Netherlands.

I'm German. Originally in 2009 when we weren't even sure if solar and wind power would be as cheap as we see them today. It was fundamental research. It was looking ahead to a system where decisions have to be made decentralized and not in centralized power plants, and where actual loads more and more need to follow the supply. I was working on forecasting market mechanisms and these things in the early days on the topic. The topic never left me.

It's sometimes frustrating to see how much excitement there has always been about these things taking off. But the patience you need to bring to take one step at a time. About the time that I defended my thesis, the Paris Agreement happened and it did make a difference. It was setting people up, aligning minds. But there were still topics like are electric vehicles even a good idea, heat pumps a good idea, and there's more and more of that.

I worked for a smart energy startup for a while here in Amsterdam. They were an energy aggregate urban startup. They have been sold to Vattenfall. That's a big Swedish energy company. After a short while and only a bit later, I did join my previous colleague in research, Felix in founding Citta site as originally a spinoff from that institute we worked at.

So we really tried to take the experience we had gathered into the real world.

And so tell us a bit about Seita and what is the value proposition that you bring to your customers today?

Yeah. So the full name is actually Seita Energy Flexibility. We really focus on that one topic in the energy transition, which is that mostly consumption should be flexible with respect to the time it happens. And I always like to talk about the course of a day. So within a day, you want to decide when energy consumption should happen for the best results, you can save CO2 as has been demonstrated, and you can save a lot of costs, which is now really starting to happen in this new kind of market with bigger spreads, with new kinds of dynamic tariffs.

So our software is getting data about the assets you have, about the weather, maybe forecasting about prices. You see this, this lot of sources. And the outcome is when should you turn things on or off? That's the simple picture.

Who are you targeting as typical customers at this stage? Help us get a sense if this is  a solution for me as a consumer when I turn my washing machine on, or is it for big industrial customers who, you know, decide to turn their machines on or off?

Both are correct. I wouldn't sell my software directly to you though. That seems overly complicated unless you're a tinkerer, because you would use the software for free, we will come to that later. You mentioned two big sectors where flexibility in energy consumption is present. That's the industry and the built environment. Yeah, they are both our target customers. They come with their own problems.

They both have a lot of potential in different ways. But our belief is that the core problem, and you can see the researchers coming out, the core problem to solve is similar. I have to have flexibility in how I plan to use energy throughout my next day. I have to heat my house or I use heat in my industrial process.

You know, these are two super big things. Sure. And how should the software think about that in the presence of incentives like markets and other constraints that happen? And in our view, you can have the same back end software, the same underlying principles, and solve them in a really good way and tackle both of these bigger sectors.

How do you think about the ecosystem in which your software comes in and with which you integrate both on the data feeds side of things which you've mentioned and also on the side of the machines that need to be controlled by your software to be turned on or off. How complex is it to plug into them today and to basically issue instructions to those machines to be able for them to react to what your software is predicting?

So that's a big crook in the whole story. You're correct. I remember when we were a software development company that existed for, let's say, four years, a conversation two years ago that somebody told me that I should bring the hardware as well.

I mean, I was focusing on being an actual machine learning and software engineering startup and really focus on that one thing, solve that one problem very well and the value would come. And somebody said you're not coming to the customer, they need to actually unearth the data first and they need to control stuff. So you will never really have a foothold.

And in the two years since I sense a difference. I see companies coming out with solutions that solve this data step and they are really good at it. So they focus on that. You got sensor companies, this sector is exploding for a few years in the energy sector. You now have a rich set of companies to help you put sensors and get that data together.

But also smart gateway companies, and sometimes these are the same, where a smart gateway has the ability to store some data. For instance, the schedule that we compute to make some very simple decisions to talk to some APIs, but mostly also to connect to devices and tell them what to do. 

A lot of the time that's only one machine protocol that these companies solve. They take OEMs and try to integrate their assets and really solve and say, well, we're compatible with Hyundai, we're compatible with heat pump maker A, we're compatible with this electric vehicle maker or whatever. So this is a lot of work. And I'm happy that other people are doing it and we've built some relationships with these players.

So I think this is all going into a direction where we can collectively really solve this.

So to to come back to the discussion you had two years ago about the need to start building hardware, what you're saying is basically you decided not to do that and you're starting to see the ecosystem shape in a way that you can be positioned as a pure software player, basically.

And tell us a bit about where the open source nature of your software comes into this discussion. Why did you make that choice and what does it enable you and your users to do?

It also has to do with my personal story. I am using open source by myself for more than a decade. Yeah, I'm not one of the Linux users who did it when they were 14, but I've seen some of the journey and I'm just super happy with the general idea of a later part of the story is when I was working for the software application startup, I saw them build nice software.

I was part of one of the scrum teams there. They had built a great team anyway that was really sitting on that code and doing great things. And when they were bought, I saw a lot of people leave after some time and I just know that code without the people around it is probably worse. A lot, lot less, yes.

So it was an experience to to see that this code is now somewhere where it can live, I think, and was wasted resources from my point of view that we spent so much time on it and I've seen a lot of startups actually in my 13 years in the energy sector now come and go because sometimes the timing wasn't right.

I we should, we should actually open up more and stop losing all that value. And now looking forward. So usually people tell me, aren't you afraid that somebody else is going to eat your pie? That might be valid concerns sometimes if you're building a commercial company, right? So this is really in the context of doing well, but I'm looking at one of the biggest opportunities the world has ever seen.

The energy. I actually don't have the correct number right now, but it doesn't even matter. It's it's trillions that is being spent every year. The difference you can make with flexibility can be up to 30% in your energy bills. At least that's true for industry. That's a study that in 2030, that's will be the case. And the larger players have already put down plans that they have 15 years to really actually solve all of it.

That's amazing. We have this huge market and you have the timeline where everybody, almost the whole world needs to be on board. And I'm reminded of the saying that a rising tide lifts all boats. Right? So even somebody who's not your paying customer contributes to your ecosystem. He might lead to other customers. So I'm looking for a win win win with the nation and currency.

Do you already have an open source community that is, you know, taking your code and, forking it or contributing to it? 

It only the beginning. So as optimistic as you have to be as a founder, I ventured into that, which is actually extra work. I remember when we decided to push that button. We had written code already for two years. This originally came out of a project we did with a partner in South Korea and we pushed that button on GitHub.

So this is now open for everybody to see.

And that was exciting, but nothing happened. You have to actually nurture, you have to show it around. You have to put more work in it to build this attention, to build in a community. And we are still busy with that. So every month we hold an open meeting where some people show up and we talk about the next steps, and we do that in the context of the Linux Energy Foundation.

So one of our steps was last year to say, well, this is only open source in general, but also business in general is about trust. And we managed that and the news and information accepted our Flex Measures project. And now we are part of a bigger context there where open source project governance is a big topic and we have some more channel partners.

Next week I'm going to the first Dem conference in Brussels. And that's all community building. That's some effort I would love to do. But yes, the energy sector itself, but also open source requires you to be patient and just continue working on it.

What is the profile of people that you see at those community building events that demonstrate an interest for your code? Help us for those not familiar with the community. Are those independent tinkerers or are those people working at some of the companies in the ecosystem that you mentioned, thinking about, well, how could we use the open source energy related projects to make our products better?

What is the sort of, you know, profile and motivation of people looking to use your code?

I see mostly people who work at companies and who for one reason or the other are thinking about energy flexibility, optimization, automation. These companies are usually not the smallest, right? It's usually startups who are just already building whatever they want to build. And it's usually big and large organizations that and that's also who I'm looking for is channel partners who have done their job on finding out how difficult it is to make software that continuously crunches time, series data and forecasts and schedules.

And it talks to several APIs in real time and understands what flexibility means. So what can you do? That's me. The core they've understood that. So they have some i.t people and they understood, okay, let's not spend this effort. Let's not hire six super smart people and give them 12 months and see what they did. And let's look around as the channel partners.

I'd like to talk about that. Sometimes this comes up that they're in a make or buy decision. Oh, sure. With this kind of technology, should they build it in-house? Then they own it. It's theirs. It's a strategic part of operations. Right. You're close to the customer. You interact with them every day. You help them to save maybe 30% of their money.

You help them to be CO2 compliant. You know, it's a great way to connect. It's a high level service, but it's a high risk to build something like that. And buying, well, you just let somebody else do it and it's gone forever, basically, or and you have all this vendor lock in risks and sure, my, my take is buy with us.

We also host it and you can turn this into a make for yourself when you are ready.

Can you make it more specific for me. What type of companies would have that type of decision and potentially decide to buy or adapt an open source solution?

We need to talk about companies who already interact with energy data right now. So they already have a solution where sensoring somehow takes place above the main sensor. Yeah, you need to have a higher resolution both in terms of the timing and also understand what the asset is doing. 

I only care about the assets that really are responsible for flexible demand of a certain size. For example, the dishwasher. I'm not interested. We can just group it under whatever other demand.

And so what would be examples of assets you are interested in?

So in the build environment, I think we should just focus on heating and electric car parked in front.

And can be cooling sometimes as well. HVAC.

Yes. And so an EV charging station provider would be an example of a company that could have an interest in using your software to make their chargers more intelligent?

Yes, you can get some of them ship their own little logic on there. Sure. Maybe following dynamic tariffs. The Netherlands, where we are, for instance, is a great example where there's already, I think, six energy suppliers who just let you follow the day ahead profile in prices. It’s very interesting. And you ,could there's just some OEM suppliers who would bundle some logic in their chargers.

So that your asset becomes smart in that way. What I like to do is to go one level higher. There are some parties who deliver the house or deliver larger buildings, or they deliver utility buildings like offices or hospitals. And they realize now they don't want to stop at one of the flexibility sources like the cars, car chargers in front or on the other hand, the back of the building, which gets electrified now or there is.

So they want to look at the energy bill or energy contract of all of it, but as if it's going you get a lot of startups now that focus on one or the other. Yeah. To build an integrated solution is not out there really. So that's why we're trying to go. We're trying to talk to the companies who have that kind of view.

So the developers and the builders of houses and offices and factories, is that what you mean by the companies who are basically building new things?

Yes. I think in both of the main sectors, it differs a bit. So our company Seita has four people now and mostly we need engineering. Apart from me, we're all just building the product and making it better. But the next step is that I will need managers that understand how it works in the build environment and in industry.

So because in the build environment I think you have to find the parties who deliver these integrated solutions, you know, building and installation companies, the ones that actually control all of it. So through this it differs a bit. And in the industry world, I think you have a different kind of company, you have more ESCO roles. So ESCO stands for energy service company that really help industrials to manage the energy data and make decisions.

I'm in contact with a very interesting company in Germany who just completed the whole transitions transition journey for industrials. These sites have to be transitioned over years to go to CO2 free. And we come in then to help them do the last 30%.

Well, it's different relationships we have to build, but basically giving these same kind of technology service in the end. And there's a lot I think that makes our business a bit hard to scale is that there's an onboarding or consideration effort always in the beginning because we'll almost never get the same configuration a couple of times if you're thinking like we integrate.

So the certain pieces that come together to form the overall energy picture of a, of a building in certain built in industry or in, in SMEs, in business parks where said, listen, you sit microgrids all over the world. You know, that's what I tried to also think about. You never seldom get the same thing, you know, 100 times in a row.

Usually we have to think about this at the beginning and to help people make decisions. That's why one of our enterprise support services is simulations. Before you put these pieces together, before you have that solution, you can also run simulations with us to understand which solutions make sense. Under which market conditions should you add a battery, should it be small or large.

Yeah, and, and we tried to make that. So that's one income stream that's basically enterprise support. And to get people into our software as a service revenue stream is something that it's not a one click event that's later part of the customer journey for us.

In terms of customers and revenue, where are you on the journey? Do you have life paying customers for either of those two services or are you in the pre pre-revenue development phase?

No, there's some revenue. I mean, me and my co-founder have managed to carry ourselves for two years or so. And now the team has expanded to four. But that's also due to our recent funding. Actually, there's a few paying customers. We use this simulation service, also a larger one in the built environment that I'm really happy to work with, but it wouldn't carry four people just yet.

So it's a mix for now and we have to try to get the right mix within the next two years, let's say, to always move towards a mix that we want. So I want the software as a service part to slowly grow percentage wise in our revenue from.

From a product standpoint, is it the same underlying product or do you need to build more to be able to deploy the live software as a service solution that basically is on 24/7, enabling customers to optimize their energy consumption. Do you need to build this out more or is that basically ready?

Oh, the underlying principles already we have also done some projects, some with paying customers in some in demonstration settings to show that we can handle certain different situations. So in essence, that's an industry project where big centrifuges have to be scheduled during the day to run at the best times. And that customer doesn't have an flexible energy contract.

So they only care about CO2. So we have our own CO2 model and help them schedule that. I think they saved 27% of CO2 and that's their deliverable.

Only on the days that there are some flexibility. Right. So so this is helpful to understand we are not saving energy. We are iterating what will giving you the best times. So if that's a day where these centrifuges have to run 24 hours of that day, then we cannot help that day. Right. It's just going to happen. But on the day so we can help.

This happened in another project. We built our own vehicle to grid's lap with another small party bus. We simply wanted to show that we can handle real time and also buying and selling because that's what makes vehicle to grid so nice. In that circumstance, the occupants of the house actually got a dynamic energy contract, so they were really making money.

Some of these days. So we're covering these bases and what for us, technology wise is really the next step is to to do sector coupling. I want to have an optimization that takes care of different kinds of flexible assets in and one optimization. So I make I make a schedule for a site and in that site you have these most important flexibility assets.

Let's say you have heating or even several heat sources, a heat pump and maybe there's even a gas powered solution. You've got your solar panels. But also another flexibility like in mobility is part of the the site. Now, maybe there's some kind of charging pulls out there. Yeah. And to couple these two different, you know, very different modes of electricity use and also how the flexibility actually works for them into one solution.

This is our next technology. Most of them got it.

How do you intend to do that?

Well, I hiring more smart engineers and to making its algorithm smarter. But so we are not a startup that simply will scratch that simply. I know that deep learning isn't simple, but yeah, we're not doing a this kind of blackbox optimization where we are just saying let's let's define that there's a lot of inputs and in train for the outputs and we'll get something we are trying to describe the problem well and so for me it's half making the algorithm better.

But the more important half is for me to actually model the problem. So to model, well, what do we mean with the flexibility of heat if we describe all the coefficient coefficients, how the heat is lost and what you can do with the heat, how long you can wait should be very well. If the better you describe the problem, then you don't have to be so good on the other and especially describing different composite situations in a language that we as engineers can can work on rather fast, you know, for for new let's say there's a new channel partner and they have 50 sites that are quite similar.

I want to be able to have them onboarded fast and that shouldn't be a four month project. Right. The the the time that this stuff can be operational has to be brought down. And that's, I think, the the main engineering challenge.

You mentioned R&D funding earlier. So tell us a bit about that. What did you secure as funding and what are your plans for how you're going to use it.

That it's actually very exciting. So we got a grant from the Dutch government, this case, this is one that secures something like 250 years for two years. So that's how we grow. And I really like the consortium, so it's awesome to other startups and the proposition is to really help us in the built environment with serial renovation.

So as you know, there's real estate companies looking at the portfolio of, I'd say some 300,000 apartments and they already have the demand. So let's say from BlackRock investor, you got to have CO2 fill friendly portfolio here until 2036, whatever it is. So they already know they have to really go through the whole portfolio. And some of these are how to renovate.

You can't get the occupants out to do the whole renovation to to get heat pump and very well done insulation and ground heating in. You will always pick the best ones first because the ones you can do, but there will be parts left over. And the thing that we're working on with the three startups together is that is going to be high temperature heat pumps that from one startup and hot water tank that is very forward looking.

It can capture the heat a long time so you can maybe even last through a so-called dunkle flower to go this this time when there's no wind and no sun for a couple of days. Yes. The great thing is you can get your gas heating out and the high temperature, the heat pump in and basically be done within a day.

And the the apartment will use more energy than the one that is super insulated and has a, you know, a high efficiency heat pump that works in lower temperatures. But you've also saved a lot of money and a lot of effort and also even material for these well, this group of apartments. And you didn't have to send the the occupants out.

So we think we're working on a proposition here for real estate companies that that can scale. It solves the giant problem they have. You know, let's say these these 20% of their portfolio that they have no idea right now how and when they should tackle. That's our job is basically the energy management system to manage one apartment at a time, basically to manage the combination of the heat pump and the hot water storage to make sure that's always in hot water that you might need and to buy it when the prices are low by the electricity to make cold water with the prices are low.

Sure. But you also have the, of course, a problem, especially with high temperature, heat pumps, that they use a lot of electricity. And you have to look at your neighbourhoods and the transformation, the substation there, and you don't have to you can't overblow it. So we will actually work on a neighbourhood level. Well, our software is supposed to work on a neighbourhood level.

In this research project, we will work, I think, on at least five apartments at once. And to make sure that they shuffle around the times when they actually use electricity. So to really help the network operators, otherwise network operators wouldn't even accept this. So it has to be fit in this project. I'm actually hoping that finally the energy transition will pick up the speed.

I've watched it for 13 years. We're not really doing it, but there has to be some exponential pickup quite soon. And hopefully this is one of the projects where we can try to do some commercial deals before these two years are even up.

That is an exciting project for sure. Nicolas, thank you so much for your time. We we are we've got to the end of the time we we gave each other is been fascinating to dove into you know, the detail level of the of the data and the ecosystem in which you fit in to solve this this critical problem of bringing flexibility to our energy system, which is very hard from, you know, from this conversation and others, there's a lot that needs to change for this to be where it should be when it comes to reducing our CO2 consumption.