David Spieldenner on PVsyst, 3D Modeling, and What It Really Takes to Scale Solar

0:00with all of theseups and downs on the solar coaster, new opportunities emerge.people say like resiliency, like the industry is very resilient.it's just a group of really smart people that are very good at adapting and seeing theopportunities that come with every transition that happens, whether it's a policy changeor bankruptcy.in a country of a major player, all of these things that look like real big strugglesalways have opportunities attached to them.

0:27this industry, think, at its core is a group of people that are really trying to dotransformational, like revolutionary things.you have a lot of people that are really good at seizing those opportunities.And that's why I don't get myself too wrapped up in any of the like doomsday scenariosthat people talk about when it comes to the renewable energy.

0:44Cause I know that the industry is just gonna keep pushing forward and you gotta stayopen-minded and look for those opportunities and try to seize them when you can.David, good morning, welcome to the show.Thanks, Peter.Finally, we made this happen.So I've been looking forward to this chat for a long, time.We've been talking about what are the optimal ways of calculating the yield as well.

1:09So let's start right there.So you actually said once PV-SYST is not really a future of solar design, but it's aliability.So that's quite a bold statement.Can you elaborate on what's broken with the way of how we actually model solar performancetoday?Yeah, happy to.So just provide a little bit of background.So the product that I represent is Plant Predict.

1:32It's sold by Terabase Energy.I've been with Plant Predict from the very beginning, which goes back to 2010.Actually, there's even some parts of Plant Predict that even started earlier than that.But in 2010,I was with First Solar and First Solar is the company that originally created PlantPredict and they did it because they lobbied with PVSYST for 10 years trying to get PVSYSTto properly model thin film solar module technology.

2:02And after getting fed up with that, they decided that they would throw their money intothe problem and just create their own energy model and just offer that to their customersso that their customers could get good models.But one of the things that I'm really grateful that those original visionaries behindPlanPredict, what they did is they thought about what the energy model of the futureshould look like.

2:25How should it operate?And I remember back in those days, I was on the IT side.We would try to set up PVSYST on a server and find a way to automatically interact withPVSYST, as I'm sure many people around the world have done as well.And it just doesn't work.mean, it's not supported.It doesn't work.It's overly complex.So from the beginning, Plant Predict was built API first.

2:47So there's a series of API endpoints that you can interact with programmatically and doall of your programming or all of your predictions programmatically.And then theThe primary consumer of the API endpoints is the user interface.So then we build a user interface that sits on top of the plant predict APIs and interactswith plant predict through those API endpoints the same way a performance engineer usingPython would be interacting with the API.

3:17So API first is important.Built on the cloud.So the cloud infrastructure that's built out over the last several decades is incredible.So being able to take advantage of all of that infrastructure in order to make theperformance modeling process as efficient as possible is really important.Like, why are we trying to run performance models on your local computer these days?

3:39It's totally not the way you should be doing it, but it's the way that most people do it.So cloud-based.API first.We also wanted to create a user interface that was simple to use.We felt like PVSYST is just way too complicated.There's a long learning curve to really get proficient at it.So as a result, people are hiring experts to simply use a piece of software, which seemskind of wasteful.

4:06So we tried to create a user interface that was such that the learning curve was a littleshorter.so that people that were performance engineers could have the power that they need to doreally sophisticated energy models.But for those who are not as experienced, they still should be able to at least run apreliminary energy prediction in the software.

4:24And so we try to create a user interface that's easy to use and very friendly.And then another thing, because we built the system in the cloud, it natively doessub-hourly modeling.do the modeling at whatever time interval your weather file is coming in.So it just processes through the energy chain on that, starting with the weather data andruns it all the way through your DC, your AC, your transformer, all those calcs on a timestep by time step basis based upon what is in your weather file.

4:53So sub hourly modeling is just native to Pleverdate.It's how it's always been done.So we don't have to worry aboutadders for things like inverter clipping, because we actually do the modeling at the timestep that you need.So the clipping just naturally happens based upon the data that's in the system.So things like sub hourly, cloud-based, API first, intuitive user interface, those aresome of the main differentiators that I think the solar modeling software of the future,this is what all of them are going to have.

5:24These are the main limitations that PVSYST has that the software industry as a whole hasovercome that the industry needs to take advantage of when we think of what the model ofthe future looks like.I hope that kind of sums it up.It's kind of a lengthy answer, but hopefully you get the idea there.It makes perfect sense and it's called performance modeling for a reason, right?

5:46It indicates that there should be performance and if there isn't any performance then it'squite a tricky situation.Before chatting on this show, we were discussing how much solar will still need to beinstalled.So we are talking about terawatts and terawatts of solar.So obviously there we will need a little bit of different levers to move in order to beable to gain and harness those efficiencies that we have previously talked about.

6:13Can you speak to that?Yeah, absolutely.So I personally feel that PV SIST is a bottleneck for the industry.because the energy model is such an important part of every solar power plant, mean, thatis how we know why we are building the power plant is because we want to generateelectricity and we want to sell that electricity.developing the energy model is a really important part of the development process.

6:38And so when you are using a tool that requires an expert to be able to run it, so you gotto make sure you've got an expert, and then that expert's got to have like a superpowerful computer to be able to do that modeling.And it's still going to take that expert weeks.And with individual iterations taking several hours, in order to produce a result so thatyou know what you can do.

7:01And then because they have to.take shortcuts, they may not necessarily get you the most optimal design because they'regoing to ultimately be just trying to get the system to run.So I feel like PVSYST creates a major constraint for the industry to do the level ofdevelopment that needs to be done.And the other part of it is I think that people are missing the mark.

7:25I mean, we see studies come out all the time abouthow these power plants are underperforming.I think that in some cases, the power plants are underperforming necessarily becausethey're not optimized appropriately.So maybe they could achieve more energy than they were if they had been designed properlyto begin with.But also I think that there's overly optimistic design.

7:46People that are experts in the software know how to make the software do what they need itto do.And so they'llthey'll try to create a more optimistic forecast than what is realistic with the equipmentand the location that they're dealing with.I feel like we could move faster as an industry, but we don't have to necessarily movefaster and compromise our accuracy.

8:08I think the industry needs to move faster.I think the industry needs to come up with more realistic, accurate energy predictions.And so I think the energy modeling software of the future is going to help the industrywith both of those things.And so, yeah, I think that we are at a really pivotal point to where we got to acceleratein order to the targets that we need to hit over the next couple of decades.

8:32And if you look at that, if we are trying to manage expectations of our customers, thesuppliers, of the project developers, and then, it's always a two way battle, a tug of warbetween promising something that will never be delivered versus actually winning theproject.And we need to find that sweet spot of as you as a developer or EPC.to actually get the project but also not to be overly optimistic with what your plant canproduce.

8:57Yeah, that's very true.A lot of people talk about modeling accuracy.But I think that what's more important is to talk about model realism.We want to create models that are realistic, that are going to actually model what thepower plant is as realistically as possible.Accuracy is easy.Accuracy is just, the equations line up?If I put these inputs in, will I get the expected outputs?

9:22Accuracy is pretty easy.But realism is actually pretty hard in the world of modeling.How do you build a model to accurately reflect the real power plant that you're trying tomodel is a much bigger challenge.We are constantly innovating.So we just rolled out our.our 3D modeling capability, which we think is a step above anything that's out there rightnow, not only for how it really accurately models the 3D effects of your site, whether itbe transposition or shade effects or 3D objects near your power plant.

9:58It does like all of these things, But it does it in an incredibly fast way because we'reusing GPU processors in the cloud.We have like the latest technology behind this thing.So it's like incredibly fast, which is really great.This level of innovation only happens if there is not a monopoly.Right now, if there's a monopoly, there's no incentive for anybody to innovate.

10:19But once you break down that monopoly and you have companies like DNV, PV case, Terabase,all trying to build a better model, you end up gettingthese innovations that ultimately are more realistic representations of the power plantthat we're trying to model.And that's really what I think of when I see our 3D model results is people say, itaccurate?I say, well, what it is is it's very realistic.

10:44It's understanding slope and modeling according to that in a very realistic way.So I think that's how we need to be thinking about modelingtools and how realistic are the results to what can be expected.And so one of the things that makes me really comfortable with PlanPredict is PlanPredictis not only used in early stage development, but PlanPredict also is used for operatingpower plants.

11:11So we haveat least 5 gigawatts a day of operating power plants that are being modeled on a dailybasis through Plant Predict.Once again, the API capabilities along with the sub-hourly capabilities work perfect forasset management.Because if you're an asset manager, you can get your Medio data from the prior day inminute level increments, pump it through Plant Predict, and getand expected energy results to help you understand how the power plant should beperforming, not only at the meter, but at the inverters and at the combiner boxes.

11:41And then you can use that to analyze expected energy.I know that PlanPredict is producing realistic results because on a daily basis, it'sbeing benchmarked on at least five gigawatts.That's what I know of.I know that there are probably some customers out there that are doing it that I'm noteven like accounting for.So it's being benchmarked on a daily basis against 5 gigawatts.

12:01And the expected energy results that are coming out of Plant Predict have to be accurate,otherwise they wouldn't use it.The performance engineers, if they were looking at the expected energy results on a PlantPredict, they're like, this is nonsensical.This doesn't make any sense.They wouldn't use it.But they do.They trust it.And they say, this is my expected energy.

12:18And then they look at the actual energy produced.And they see the problems.And then they send out maintenance technicians, and they fix it, and they see that theenergy comes back to the expected energy.So I sleep well at night knowing that I have O&M customers running this model every day.And they're getting realistic results out of the tool.I say, if they trust it, everybody should trust it.

12:40But ultimately, an open market is typically yielding superior products versus a monopoly.from that also the users benefit, the industry benefits and, this whole flywheel of scaleand speed will gain more and more momentum.But if we segway into the biggest misconceptions, perhaps the developers have whenthinking about bankable yield estimates.And again, I want to highlight bankable.

13:04Can you speak to that?ultimately, they're worried that they're going to put forward a yield report and the bankis going to say, we're not going to accept that.We're not going to finance your project because the tool that you used is not bankable.That's the fundamental concern and what people generally mean when they talk aboutbankable.Plant Predict has been used for financing before.

13:27I mentioned First Solar when it was, when First Solar was doing large scale solardevelopment, First Solar was only using PlantPredict for its energy modeling and the bankswere accepting it.So it can be accepted.But we still have situations where I'm told regularly by customers that the banks write inthat you must provide a PV-SYST report.Fundamentally, that is anti-competitive.

13:50Like that should be illegal.but it is like how the industry has operated for several decades.I think the future needs to be where the banks don't dictate the tool, but maybe arewilling to accept a handful of energy models.I mentioned DNV Solar farmer great product.I have a lot of respect for the work that DNV has done over the years on the Solar farmerproduct.

14:12I know PVcase has come out with PVcase yield and they're working hard on that.Solar GIS is doing some work.I feel like there's enough players that the industry should be able to choose, like thelast thing that we want is the industry using like Excel spreadsheets to estimate theirenergy and taking it to the bank and saying, look at how much energy my power plant isgoing to produce.

14:31We don't want that.But there's definitely a handful of tools out there that have reputable customer bases,reputations within the industry, you can look back and you can see the time that they'vespent in the industry, that I think that they should be acceptable tools when it comes todoing energy modeling.Once the industry says, OK, there's a handful of tools that we're willing to accept, allof a sudden then those tools can start competing with each other and trying to raise thebar, trying to come out with capabilities and functionalities.

15:02that more accurately, realistically model the power plant, that's only going to make thebank's lives easier because they're going to be getting more accurate models or modelsthat are going to be more realistic in how the power plant is going to produce becausethese companies are going to be incentivized to try to make their products better.Without that competitive pressure, what we're left with is an industry where there's justa fear of innovation.

15:27Like I feel like what we've seen over the years with PVSys, they don't want to innovatebecause innovation might put their monopoly at risk.And that's not helpful for the industry as a whole.of course, as we can appreciate, of the software providers are working hard at eliminatingthose bottlenecks.And if you still have one major one like this on the whole process, cycle and thedevelopment cycle, it's making everybody's life much, much harder.

15:52Yeah, it is.mean, the one thing that always cracks me up is we're oftentimes asked to make plantpredicts results match PVSYST's results.nobody's done it more than the team that we have within Terabase to do that.And as a result, we've identified bugs within PVSYST where some models are just wrong.They're taking the energy in the wrong direction.And so then we notify PVSYST that there's an issue.

16:19We want to make sure that you're getting more accurate results.People are like, no, we want you to match PVSYST.You're like, but then we'd be wrong.We don't want to do that.And so there's cases where it's like everybody wants the benchmark to be PVSYST.The PVPMC, the people at Sandia National Labs, put on thethe PV Performance Modeling Collaborative.

16:39And this collaborative group does these blind benchmarking studies.So they'll issue like performance specs and weather files for a power plant that's inoperation.And then they'll have performance engineers model that power plant.they'll see how accurate or how aligned their results come out with this actual operatingpower plant.And it's really interesting this last year, it was the 2023 blind modeling competition.

17:03They allowed all the performance modeling companies to provide a submission and.PVSYST came in with their model and they were about 4 % over predicting What wasinteresting is several of the other competitors came in right about the same point, rightbetween 3 and 5 % over predicting.And Plant Predict was within 1%.We were right on top of where we needed to be.

17:29And I just thought it was interesting because so many people are building their energymodels so that they align with PV SYST rather than trying to align with actual operatingpower plants like what we do.so, yeah, we were really proud of the results that came in right on top of the actualpower plant production values that a lot of credit goes to our performance engineer.

17:51who was running the study, but nailed it right on top.But yeah, it was very insightful to see several tools come in right around 3 to 5 %overestimating and PVSYST being right there at like 4 % overestimating.But jumping back to the topic, we've talked about PV assist, we've talked about energymodeling and how the industry is actually still held back a little bit in terms of whatthey could be producing.

18:17And what's your take on what we should be doing?Like how should the playing field be a little bit more level and what are the kind of mainfeatures that would benefit the industry as a whole?Yeah, so.The core algorithms, like the physics, the core physics inside the software definitelyneed to be there.I think the major tools that I've already mentioned a few times on the call are alreadydoing that.

18:42But you should definitely be using a tool that is physically accurate so it understandsthe energy chain of how the power plant is going to operate all the way from the sunthrough to the transformer.So you should be doing that as a baseline.But then we got to be looking at some of the major sources of variability within yourpower plant.So we've spent the last two years investing heavily into our 3D modeling capabilitybecause understanding how the power plant is going to perform in a 3D world is reallyimportant.

19:13Especially as things like terrain following trackers come into play, you've got tounderstand the impact on the energy from a terrain following tracker.So you've got to understand the 3D effects.That's become a new, really important thing to consider.You've got to be doing these simulations in a sub-hourly way.So.Doing hourly predictions, so averaging all of your medial inputs into a single point forthe hour assumes that we live in this linear world where you can just take an averagetemperature, average irradiance, average humidity.

19:47If you think that you can do that and get an accurate result, you obviously don't live inOhio where the temperature can change dramatically.in the course of an hour.The world is just not linear enough to be able to do this hourly averaging.So you should be doing modeling at the sub-hourly level.15 minute, maybe 10, maybe 5.I don't recommend one minute.

20:07That's a little overkill.But definitely something in a sub-hourly way so that you can understand these non-lineareffects on the inputs.And then I think you've got to go after these things like soiling.So we've integrated with PV radar and Fraxon to give you location-specific soiling levels.And now we're going to be rolling out a big integration with PV radar's cleaning andoptimization module so that you can really understand not only what your location-specificsoiling losses are going to be, buthow you can reduce them with a optimized cleaning approach.

20:44I think that cover the basics of physics.Make sure you're taking into consideration the 3D effects, terrain and such.Do things at a sub hourly level, and then start attacking some of these other areas ofuncertainty like soiling.I think is what we need to start doing as an industry to get more realistic energy modelsin the early stages of your project.

21:04You mentioned quite a few keywords here that we should definitely double click on.The first one being, terrain following trackers.So that's number one.And when it comes to the slopes and shading and maybe the tracker alignments, how did thatchange from the past to more of a 3d type of model that you guys currently use?So first off, with the emergence of terrain following trackers, we saw it at RE plus quitea few years ago when Next Tracker came out, the XTR and the array technologies came out,the OmniTrack.

21:34We saw the direction the industry was going.We knew that there was going to be this shift.We knew it was going to be a big engineering challenge to address.So we started building that out.It manifested itself within Plant Predict as terrain-pro.So the very first thing was, how do we understand the cut in the field, the benefits of aterrain-following tracker?

21:53That's where it all started.So doing a good cut fill analysis with a rigid tracker and then layer in a terrainfollowing tracker and understand how much of a reduction in earthwork we could get.But then you also have to think about the steel.So taking into consideration like your pile lengths.And if you're doing a rigid tracker, you're going to have pile lengths.

22:11If you have a valley, you're going to have longer piles.If you have a little hill, you're going to have shorter piles.So understanding your pile lengths.And you can reduce your grading by just having a pile reveal window.having a little bit more tolerance for your piles, you can cut down your grading.But then you're going to add more steel.So understanding that trade-off between earthwork and steel is important.

22:31But the part that nobody was really prepared to go look at is the energy impact.So not only should you be thinking about the dirt, the earthwork, you should be thinkingabout the steel, but you should also be thinking about the energy impact.Will these terrain following trackers save me money in the construction phase by reducingthe amount of earth work or the amount of steel I'm doing?

22:51But will I end up losing it by having lower energy production?You should be able to look at that.And so that trifecta of analysis is something that we've reallyfocused on, and the rollout of PlantPredict's 3D modeling capability is the final step ofthat.Because TerrainPro already does the earthwork and the steel.And then you can download a PVC file from TerrainPro and get all of those trackergeometries at the bay level.

23:18So you can see the tilt And then you can upload that into PlantPredict's 3D viewer.and run the simulation and then we'll check.Transposition gain and loss, because you're going to have some transposition gain and lossdepending upon whether or not that bay is tilted in the southerly direction or tilted inthe northerly direction.So you got to make sure you account for that.

23:39And then you're also going to have the shading effects.With terrain following trackers, you could have one tracker on the eastern side that's gota hump that's casting shadows onto the tracker beside of it.You got to understand what the impacts of that are.And one thing that's been kind of interesting is I assumed that terrain following trackerswould always be like a negative from an energy standpoint, because I was only thinkingabout the shadows and the mismatch that would be created by that.

24:06But actually, the transposition gain that you get from some of these, when you take arigid tracker and you allow it to follow the terrain,There's some transposition gain and loss that's going to happen with that.And so I've actually seen cases where terrain following trackers end up generating moreenergy because of some of the sun.If they have a predominantly, it all depends on the overall terrain of the site.

24:28But if you have more of your terrain following bays in the southerly direction, you couldactually get an overall transposition gain on the site.But the key there is you shouldn't be blind.You should.If you're going to consider doing terrain following trackers, it seems like everybody caneasily figure out the cut and the fill these days.They can get a pretty good idea of the steel, but then they just don't know on the energy.

24:49You should at least know.You should have a capability to run that model with those geometries and understand whatthe energy impact is of the terrain following tracker.We make it really easy with Plant Predict because you can go right from Terrain Pro toPlant Predict and run the model.But there's a lot of people that justthey'll just model it as though it's not a train following tracker and just say, well,it's probably going to be 1 or 2 % off.

25:10And it's like, you should know.it's better to be in the known than in the unknown.But if you circle back to the trackers, I wanted to make a couple of interesting points.You mentioned we are looking at how much energy is the plant going to be producing.But what's actually becoming more and more important is when is that plant going to beproducing that energy, right?

25:30And speaking from a European perspective, this year I've heard from more and more peopleas the cost of the mentioned steel and aluminumof the trackers came down also the kind of design and the costs are still decreasing ofthe trackers.There are more and more trackers being installed, for example, in Germany as well, whenyou need to do, we need to be selling a little bit in the morning and a little bit in theevening to avoid the midday peak where you would actually be penalizing for sending itback to the grid.

25:59It's actually a brand new use case that needs to be looked at it also needs to be isolateda little bit ofwhen are we going to be producing valuable energy to the power plant versus being chargedfor the excess energy that we would be destabilizing the grid with.that's where batteries become really, really helpful.And the other thing that I know is a big issue in Germany is interconnection.

26:22So what I see happening is there's a lot of innovation going on in the European marketaround trying to figure out how to build like hybrid systems.So if you've already got a grid interconnection for a wind plant and you have some landaround it, put some solar on there so that you can take advantage of that gridinterconnection.But then you also need to install some batteries and do some energy storage so that youcan have dispatchable energy when the energy is needed and when it's most valuable.

26:52you've got to think about that in your design.design approach.And I know there's some companies that are doing some really cool stuff.I know a few of them that are customers, a few of them that are friends that are doingsome bundling in of hybrid plus storage.think that going back to where we started, our tools need to innovate so that we can keepup with the market needs.

27:19Our tools need to, the people that are building these tools, the teams that are buildingthese tools and the companies that are investing in these tools need to be incentivized totry to keep up.They need to be incentivized to come up with a novel way of doing this modeling to whereyou can maybe do hybrid plants more effectively.You can incorporate storage into it.

27:41benefit from like having a more innovative solution than everybody has to dumb it down inorder to pump it through the industry standard energy modeling tool and then augment withExcel spreadsheets.you mentioned quite innovative solutions, right?And when it comes to the innovation side of things, how does the change of speed actuallyimpact the design workflow in practice?

28:03So you guys been doing some benchmarking on PlanPredict of how much faster it is than theother tools, but the ultimate...kicker at the end of the day is that how is that going to translate into more efficiencyon the next step?Yeah, anybody that's gone through the process, the creative process of trying to optimizea power plant, it's a very scientific process.

28:26You come up with a hypothesis like, I think that if I do make this change, I'm going toget these expected results.And so then you want to run the model and see if you want to validate your hypothesis.the cycle time of that study is too long.You can't run that cycle very many times.Sometimes you just have a deadline that you're trying to hit.So maybe you can only run that creative process like five times before you hit yourdeadline.

28:55How much more valuable could you be if you could run that?Like, up to 250 times speed increases.So what if you were able to run that 25 times or 250 times and keep making minor tweaksand improvements?It would be a lot better.Now, I personally, whenever I'm doing this type of an analysis, I'll get into a bit of aflow.Like, I'll start seeing the next steps laid out in front of me.

29:21I'll be like, I want totest this, and then I want to bring in soiling, then I want to remove some trackers to tryto get rid of the most shaded parts of the site.And I'll see what I need to do, and then I'll start running it.If it takes too long, I'll literally lose that flow, and I'll be like, all right, I forgotwhat I was going to do next.It's easy to have that creative process just get disrupted by processing time.

29:45Or anybody'slike been working on something and then their computer crashes.It's like they lose all the creative flow that they had.the power of getting like quick results I feel like it compounds on itself because yourcreative engineers are going to be able to get into this creative flow and they can startthinking about, all right, how can I get more energy out of this piece of land?

30:08I'm going to start.moving in that direction, I want to see those results, test my hypotheses, and just createa better power plant.I feel like you can only do that if your tools are responsive, if they can get you theresults in a reasonable amount of time so that you don't lose that flow.Yeah, indeed.If you manage to maintain the flow, that leads to better optimization, lower LCOE and alsomaybe faster permitting as well, right?

30:34Absolutely, yeah.if we talk our segway into Terabase and its mission a bit more.So Terabase's mission is to deploy a terawatt scale solar power, right?And when you guys say there is baseload power in the solar context, what do you guys meanby that?Yeah, so that's a whole...purpose behind the name.Terawatt scale solar, baseload solar.we fundamentally believe that solar in the early days, it was always like a bit of like anarbitrage opportunity.

31:04You were trying to replace like the peak or power plants and just try to take off like thepeaks of energy production where electricity is most expensive.And because the equipment was so expensive, it made sense.Like that was the optimum.Or in the case of residential, if you're trying to help a homeowner reduce a little bit oftheir electricity, because retail electricity is expensive, so it all makes sense.

31:29But in order to bring about the type of transformation that is actually going to happen,it's just a matter of time.This transformation away fromburning fossil fuels to generate electricity, to harvesting electricity through renewablesources like photons and wind.This revolution is going to happen.So eventually you're going to be seeing the base load, which historically had beenprovided by coal, and now it's a bit of a transition over into natural gas.

31:56That base load energy needs to be replaced with renewable sources.So the only way that you can make that happen is the...it's got to be cost effective energy and very reliable.Because those are the two main prerequisites for base load.Nobody wants to flip their light switch and not have any electricity to turn that lighton.So it's got to be reliable, but it's also got to be cost effective.

32:22So that's the mission that we are set out upon at Terabase is to try to drive the cost oflarge scale solar down.and deploy not only cost effective but also very reliable solar energy resources.So, we have Plant Predict, of course, which is what I've been talking about a lot.Plant Predict is used in the early stage of solar development, and we've already talked atlength about how it can help build better power plants, drive costs out of the designprocess.

32:50But then,We go into our Construct platform, which our Construct gives you a bird's eye view of yourpower plant as it's under construction and helps you really understand your constructionprogress and also like the, whether you're meeting your timeline, your schedule asexpected, and also seeing all the issues that may be occurring on the project sites thatyou can react and deploy resources to address those.

33:14And then we have our SCADA product.So when I talk about reliable, our SCADA product is designed to be integrated into hybridsystems.We have a full storage battery management system within our SCADA system.So we got to have a SCADA system that's going to allow you to not only be cost effective,but also have a very reliable power plant.And then the big deal on the horizon is automated installation.

33:39which that takes us back into the construction stage, but building power plants usingrobotics and AI so that you not only can drive the cost down to the construction process,which is true if you can employ robotics, there's certain benefits that you get that canactually lower your labor costs and lower your transportation costs and grading But alsoyou can build a more quality, higher quality, a safer power plant that all ties into that.

34:02It's got to becost effective and reliable.And so we believe that through automation and robotics and AI, you can build a betterpower plant using that type of technology.So those are the four facets of those are the four products that we have at Terabase.But those are the four ways that we are trying to really have an impact on the large scalesolar industry.

34:24it's really exciting stuff.I've been at Terabase now for four years and to see thejust the improvements that have happened in the four years is just absolutely mindblowing.a lot of these ideas, TeraBase has a really strong engineering and R &D like core.And we hear about these ideas that seem just like in like, so they're like sciencefiction.And like in six months, our R &D team is like showing it in all hands.

34:51call like a video of this concept that six months ago just seemed like science fiction.And so it's really incredible to see the progress that's being made at Terabase.And you mentioned AI, is a huge buzzword of today and which actually begs the question ofdata centers as well.how do you guys feel like data centers will actually change the landscape in terms of bothdeveloping projects and also consuming that electricity?

35:13Will that impact the customers, the average on the street as well in terms of energyprices and different behaviors or what is the tendency that you guys see?it's become a big factor.I live in rural Ohio, and we have a pretty solid electrical infrastructure in this area.And we have relatively low cost land in this area as well.So a big.Part of the Northwest Ohio economy is data centers now.

35:42So there's a lot of data centers that are coming into play.So I see it firsthand, in my community.I see them when I'm driving down the road.I see them coming into place.we got to find a way to power these things.And so what's great aboutSolar is really fast to dispatch.And what I mean by dispatch, I mean like you can conceive of a power plant, design it,permit it, construct it, commission it in a relatively short period of time.

36:12And when I say relatively, I'm talking like 12 to 24 months.And so as these data centers come into being, they're going to need energy and they'regoing to need it fast.So solar the perfect solution for that.In some sort of combination with grid power, can definitely get the power that you need,but not have to overburn the grid that you have.then ultimately, one big concern,is everybody's worried about their personal electricity bills going up with the emergenceof data centers.

36:42It's like, OK, so now there's going to be a huge source of demand on electricity in myarea.Am I going to end up having to pay for that?So we got to continue to find cost effective sources of new energy in order to make surethat it doesn't become a big burden on everybody.where these data centers are being built.Perhaps the first step of or of the solution is right regulation, of course.

37:05So, so people are not going to be outbid by the data centers and ultimately I will consumeall the power that's being produced in a local area and then you won't be able to turnyour kettle on.And the second thing is what you personally can also do is install solar, right?A little bit of solar, a little bit of storage on your own house and you will be able tocover your basic needs for a couple of days.

37:26So you won't have to suffer.that's the thing that's amazing about solar.I was just reading an article that's talking about how innovations in solar, but moreimportantly, cost-free money transactions and innovative financing players are totallytransformingthe energy infrastructure in Africa and bringing power to millions of people.So that's the beautiful thing about solar is that it works well in a large scale utilityscale configuration because you can get the economies of scale and you can deploy largeswaths of solar very quickly and cost effectively, but it also works in a distributed way.

38:05So like you said, if the energy pricesin my local grid start going up, then that further incentivizes me to do someself-generation.so, and the sun shines, even though it doesn't shine as much here in Ohio, but it doesshine everywhere in the world during the day.So everywhere is, it's possible to put solar panels on your roof and harvest some of thosephotons for yourself.

38:27there's no question in my mind that100 years from now, solar is going to be like, people will look back on the 20th centuryand say, I can't believe people burn fossil fuels to generate electricity when we have allthese free photons fallling to the earth and it's so easy to harvest them and use them inwhatever application I need.And I think that the work that we're doing right now is totally laying the foundation forthat in the future.

38:50And if you play the devil's advocate, where do you see the biggest technological gaps thatare still holding solar back?Yeah, probably the biggest issue, at least in the midterm, is definitely our gridinfrastructure.Interconnection seems to be a big challenge.Much more, I think, than some of the geopolitical issues, which seem to always create alot of like,narrative around what's gonna happen.

39:17And it definitely does have an impact, whether it be tariffs or subsidies or whatnot.the world that we live in here, at least like in the United States and in Europe is somuch based upon the grid infrastructure that has been put into place.That's where like what's going on in Africa is really interesting because all of thiselectrification that we're seeing in Africa is happening without a grid infrastructure.

39:37So in...Places like Europe, the United States, I think what's really a big challenge is definitelythe fact that the system is designed around this grid.And we've got to figure out a way to either.de-bottleneck our grid, which means you got to build more grid capacity or build more gridconnection capabilities.what you're also seeing, at least in the world of data centers, is in some cases, they'rejust going off the grid and building microgrids in order to not have to be tied into thegrid, which is kind like what you're seeing in some of developing parts of the world.

40:15But you may even see that a little bit more in the Europe and the US area.And coming back to the African example, they actually might have a blessing and a curse atthe same time as well, right?That they are going to be energy first.So if they won't have an old grid, which we do have in Europe and in the US from the 50s,60s, 70s, might be able to actually build a better grid that's a bit more decentralized,much more stable.

40:39And with the use of solar and batteries as well, it's going to be hopefully much moreresilient than the things that we currently have here.Yeah, it's pretty amazing.I agree.so much of the grid system puts so much power in just a few players' hands.And so in a distributed world, there's much less disruption come from like the powers.that's being held in just a few people's hands.

41:03there's a little bit more freedom in those environments.And so, yeah, it's gonna be interesting.I've been in the solar industry going back to 2006.Prior to that, I was in the automotive industry and the company that I worked for wasbuilding the drive trains for large trucks.And during that time, we saw $3 gas here in...in the US for the first time, not in California, because California is kind of their ownspecial gas market.

41:29But like in Ohio, Michigan, Indiana, like $3 gas just like shocked everybody because itgot so expensive.And so the transition into solar was like so amazing because I felt like I could actuallyhave an impact on what I was seeing is like the ever increasing cost of fossil fuels andenergy in general.And then I started to think about the impact it could have in these developing parts ofthe world, Africa, India, Southeast Asia, or even like in the United States, we have someareas that are energy poor.

42:01Specifically like in the Southwest, whether it be like Indian reservation areas orwhatever, where it's like energy poor.And the only hope they have is that some bureaucrat is going topass a bill to bring energy to their area.So they feel like so helpless.So like the impact that I think that the work that I've seen happen over the time, I'malmost at 20 years into a solar career at this point, on millions of people who are justgonna have better lives and better opportunities because of the energy access that they'regonna have is one of those things that I'm really proud of.

42:34because I love to see the progress that's being made.You've certainly joined the right movement.And when talk about motivation and how do you keep the guys on your team motivated?is there any sort of mantra that you guys stick to or what is the main force of movementin the team?Yeah, so,So being part of TeraBase, we're part of a startup.One thing that I think happens, you know, I've worked for Fortune 500 companies.

43:00My early career was all with Fortune 500 companies.I dabbled in some startups in the early 2000s, during the dot-com boom.And then, you know, I ended up, I went to First Solar when First Solar had onemanufacturing line.And I was like, so happy to be part of this small startup again.And then First Solar just grew into this big multi-billion dollar company.

43:20so coming back to Terabase I joined Terabase when there was less than 50 people.And Terabase is a small Bay Area startup.So.I think that there's something special that happens in startups that creates a real strongsense of motivation.It could be equity-based.like pay structures where there's stock options or whatever.So you feel like you've got a piece of the overall company that you're building.

43:45I think that that has a real powerful effect.It's also like Matt Campbell, our CEO does monthly all hands calls where he likeintroduces all the new hires.you feel like you're really a part of the team when you're part of a startup, which it'ssometimes you feel disconnected when you're part of a large.thousand people organization.But when you like know personally, like everybody and the CEO is introducing people byname and knows their dogs and stuff like that, those factors are really motivating.

44:16So I think that there's a special motivation that exists in startups that is reallypowerful and creates incredible innovation.And I think that's whyYou know, we have a real startup culture in both Europe and the United States and probablyaround the world.I mean, you see it as well in China But I think that that's a really important thing is tomake sure that everybody feels like they're recognized, they're known, like the workthey're doing is meaningful.

44:48And if there's some sort of equity basedcompensation structure in place, they feel like at some point, if all the stars align, I'mgonna get a nice payout at the end as well.I think that that really is kind of like all part of the motivation story.Indeed, motivation is one of the key drivers of innovation and especially in industry likesolar and renewables.

45:10it's ever more important, especially these days, 2025 has been a crazy year across theglobe.Sometimes it feels like it's a telenovela that somebody is writing and you actually wentthrough all the twists and all the jumps in the book, but we actually went through it andthen hopefully 2026 is goingbe a much easier year on the ride of the solar coaster,Well, that's, so I'm fortunate in that I've had 20 years in the solar industry.

45:38And so, uh, the solar coaster doesn't freak me out as much because I've seen the hypecycles that come along with the solar coaster that make you really feel like the sky isgoing to fall on you.And then an election happens, the dust settles, and it turns out that like, it's not asbad as you thought it was going to be.I remember the ITC cliff, I think it was in 2016, 2017, like that was going to be like theend of solar.

46:09And December of that year, ITC got preserved and the industry was able to keep going.I was there in 2012 when solar panel prices like fell off a cliff.A lot of companies did go out of business during that time.But like with all of theseups and downs on the solar coaster, new opportunities emerge.that's what's really, people say like resiliency, like the industry is very resilient.

46:34Well, what it is, is it's just a group of really smart people that are very good atadapting and seeing the opportunities that come with every transition that happens,whether it's a policy change or bankruptcy.in a country of a major player, all of these things that look like real big strugglesalways have opportunities attached to them.And this industry, think, at its core is a group of people that are really trying to dotransformational, like revolutionary things.

47:01so...you have a lot of people that are really good at seizing those opportunities.And that's why I don't get myself too wrapped up in any of the like doomsday scenariosthat people talk about when it comes to the renewable energy.Cause I know that the industry is just gonna keep pushing forward and you gotta stayopen-minded and look for those opportunities and try to seize them when you can.

47:19I think that that's gonna be the case.This this transition from fossil fuel generation torenewable generation, I think is going to be one of the main defining factors in thiscentury.Definitely what we're seeing from an AI perspective is going to be a big story as well.Digitization, the internet, celestial-based internet.All these things are going to be like major, major factor, but underwriting all of that,think one of the main differences when we get to the end of the century versus thebeginning of the century is people will say, I can't believe we burn fossil fuels forenergy.

47:51how crazy was that?That's, that's nuts.But, yeah, that's, that was the world we, we've grown up in.We will make it happen for sure.speaking of opportunities, what advice would you give to a new engineer or new founder whoactually ended up in the solar world?what are the tools or perhaps the opportunities that could be built that would really movethe industry forward?

48:14I would definitely tell them to sign up for a Plant Predict subscription.That's where I would tell them to start.Don't waste your time on PVSYST.I would say that.I mean, I feel like Python programming is so fundamental at this point.So you've got to have some understanding of how to do programming.Python is definitely a place to be.You got to understand how to interact with APIs.

48:39So Plant Predict, with its API-first infrastructure, the people that know how to interactwith the software at an API level, I think, have a real advantage.So I think that those are some of the things that I would say.I can see a world where some of the established desktop applications that people think arealways going to be there are not going to be there.

49:02So you've got to keep an open mind and keep an eye on what's going on.I think that a lot more of the design process is going to be browser-based in the futurethan it has been in the past.So you know, I.Keep your eyes open on some of the design tools that are coming out that arebrowser-based, because I think that there's some really interesting work going on in thatspace.

49:23Yeah.That's the right answer.as we are coming to the end of our time, are there any exciting announcements that youwould like to makeI definitely want to highlight this integrate.So we're actually entering into a software reseller agreement with PV Radar to allow allof our PlantFordict Pro subscribers the opportunity to purchase a PV Radar subscription ata bundled price.

49:49So PV Radar is allowing us to sell the product under list price toPlantPredict Pro subscribers.So if you're a PlantPredict Pro subscriber, you get this benefit of PV Radar So PV Radaris just amazing software.It demystifies soiling and cleaning optimization.I find myself doing it.cleaning optimization run on almost every project I do anymore because it's so easy to dothrough PV radar.

50:15So PV radar are something that we're really excited about.We are raising our plan predict prices in 2026.So this is something that's really big.the reasoning behind it, like the rationale is we are makingour 3D modeling capabilities available all the way down at our core level subscription.So our lowest level subscription, you can buy a PlantPredict Core subscription today for$125 a month.

50:38So for $125, you can get access to PlantPredict Core and have the full 3D modelingcapability completely enabled.We've spent two years investing in that technology.We have a whole other server stack running the 3D model.So there's a lot of cost associated with that.So we have to raise the prices.We're going to take our pricing from$125 a month to $150.

50:58You can get an annual core subscription for $1,500.That's the new price.It's been $1,250.So $1,500 US dollars as an annual core subscription.We offer small volume discounting at the, we call it the business level.That's three core subscriptions for three core licenses.Yeah, so through three users was $3,000.We're going to raise that up to $3,600.So 20 % price increase.

51:25But that's still quite a bit cheaper than buying three individual licenses.Then our Pro package, the price increase is going to be a little more muted on that one.So we're going to go from $7,500 up to $7,800.So $300 per year price increase on the pro.three user package and then the pro 10 user which we call our enterprise package we'reactually keeping the pricing the same so enterprise customers won't have any increase soit's mostly in our in our Core and and pro level that we're going to be raising ourpricing now what's important for everybody to know is the pricing this price increase doesnot go into effect until january 1st 2026 so we are going to be signing up new customersand we're going to be signing renewal agreementsall the way up until December 31st.

52:13do this, I've done this several times.It ruins my Christmas every time because I'll be writing proposals and signing deals rightup until January 31st.But anybody that wants to take advantage of current pricing can do so.All you have to do is reach out.Support@platpredict.com at is the easy way to get ahold of us.Locking your pricing, we also are offeringtwo year subscription.

52:37if somebody wants to lock in 2025 pricing for the next two years, you can do that.If your subscription doesn't renew until the summer of next year, you can actually renewit in December and we'll just extend 12 months to your subscription that renews out inJune or July.So we've got a lot of flexibility here.We want to make sure that anybody that's a current subscriber has the opportunity to renewnext year subscription at 2025 prices.

53:06Because on January 1st, we've got to raise the price.another thing, I mentioned PV Radar.Our friends at PV Radar have offered us the opportunity to sell, not only are you going toget a bundled version ofof PV Radar for $15,000, which is $5,000 less than the list price.But we're also offering an introductory price.So we're going to offer up to a $3,000 discount, introductory discount, to anybody thatsigns before December 31st 2025.

53:34So you can get a full year of PV Radar for $12,000 as a bundled package with thisintroductory discount,So a real opportunity to do some saving here in the US.We talk about Black Friday coming up next Friday.This is our Black Friday special, but it's going to be like all the way through the end ofDecember.So I plan to be writing a lot of proposals and booking a lot of deals over the course ofthe next month.

53:59whatever money Santa brings you, you can turn it into some real savings.So we will link to those in the show notes below.And on a final note, David, who should reach out to you or Terabase?What type of people would you like to speak to?So I love solar developers.Solar developers are my favorite.They're doing the hard work of identifying where we're going to put the solar and gettingthe permits and all that stuff.

54:22EPCs, our energy modeling capability is really helpful in the world of EPCs.So we definitely love EPCs.I love independent engineers.IEs are a special group within the industry thatWe have lot of IE customers and friends in the industry that PlanPredict gives them acompetitive advantage in the work that they do.They can build capabilities through our API that can add value to their practice.

54:49And so IEs and definitely consultants, consultants that help consult and maybe their ownerengineering or just solar consultants.Definitely the tool set works really well for independent consultants because it's not avery expensive piece of software.And then I have a large network of universities that are using Plant Predict.I love students and professors.

55:13Plant Predict is a very easy tool to learn, so it works really well in the classroomenvironment.And it's also very transparent in how we do some of our calculations.It seems to catch on really well within the ranks of students.those are the people that I like to talk to on a daily basis about Plant Predict.Awesome, so let's make sure that those guys reach out to you so we'll link to all of yourcontacts in the show notes below and David, thank you very much for your time and forsharing your knowledge.

55:42I really appreciate you taking your time out of your busy day and let's keep each otherposted.Thanks, Peter.I appreciate the work you're doing to get these messages out.Your podcast is awesome.You do such a good job and you have such great people.I'm just grateful to be part of it.So keep up the hard work you're doing to make sure that you keep everybody as informed aspossible.

56:01We are trying to do our best.thank you, David.See you soon.Bye.