Tecogen Inc. ($TGEN)

Greetings, and welcome to the Tecogen Fiscal Year 2025 Conference Call. [Operator Instructions] As a reminder, this conference is being recorded. It's now my pleasure to turn the call over to Jack Whiting, General Counsel and Secretary. Please go ahead, Jack.

Good morning. This is Jack Whiting, General Counsel and Secretary of Tecogen. This call is being recorded and will be archived on our website at tecogen.com. The press release regarding our fourth quarter and year-end 2025 earnings and the presentation provided this morning are available in the Investors section of our website. I'd like to direct your attention to our safe harbor statement included in our earnings press release and presentation. Various remarks that we may make about the company's expectations, plans and prospects constitute forward-looking statements for purposes of the safe harbor provisions under the Private Securities Litigation Reform Act of 1995. Actual results may differ materially from those indicated by forward-looking statements as a result of various factors, including those discussed in the company's most recent annual and quarterly reports on Forms 10-K and 10-Q under the caption Risk Factors filed with the Securities and Exchange Commission and available in the Investors section of our website under the heading SEC Filings. While we may elect to update forward-looking statements, we specifically disclaim any obligation to do so, so you should not rely on any forward-looking statements as representing our views as of any future date. During this call, we will refer to certain financial measures not prepared in accordance with generally accepted accounting principles, or GAAP. A reconciliation of non-GAAP financial measures to the most directly comparable GAAP measures is provided in the press release regarding our Q4 and year-end 2025 earnings and on our website. I'll now turn the call over to Abinand Rangesh, Tecogen's CEO, who will provide an overview of fourth quarter and year-end 2025 activity and results; and Roger Deschenes, Tecogen's CFO, who will provide additional information regarding Q4 and year-end 2025 financial results. Abinand?

Thank you, Jack. Welcome to Tecogen's Fiscal Year 2025 Call. I know many of you would like an update on how the data center cooling strategy is progressing. So today, I'm going to start with an update on the Vertiv partnership. There have been some key positive developments that I'll share, including their opportunities for our chillers. Then I'm going to walk you through Tecogen's data center opportunity pipeline outside the Vertiv partnership. After that, I will provide an update on the other avenues we are working on, including expanding our manufacturing throughput, increasing service revenue and margin and the non-data center pipeline. We have seen significant forward momentum with the Vertiv relationship. First, Vertiv has designed or is in the process of designing between 25 and 50 megawatts of our chillers into various projects. This is equivalent to 50 to 100 of our 150-ton dual power source air-cooled chillers. Second, we have been negotiating our master partnership agreement that expands the marketing agreement that we signed last year. Third, we have discussed bringing Tecogen's hybrid drive technology to Vertiv's chillers. As the sales grow, this may allow Tecogen to scale manufacturing very quickly because we would focus on the dual power source and make this to the refrigeration system that is already built in volume in Vertiv's factories. Last and most exciting of all, we have secured a demonstration project with Vertiv. This is expected to ship sometime toward the end of Q2 for 1 megawatt of cooling or 2x our 150-ton dual power source chillers. Our chiller will go to the Vertiv controlled environment test chamber, where it will operate under simulated AI data center conditions and various outside ambient temperatures. This technology demonstration project gives prospective customers data on how the chiller will operate under real-world data center conditions across a range of ambient temperatures. While we have been furthering the Vertiv partnership, we have also been expanding our own data center pipeline. In this list, we have only included opportunities where the end customer has told us they plan to use Tecogen chillers and have made significant progress on signing data center tenants. This list is sorted based on our current project confidence. Developers that have existing data center experience and financing are higher on the list. Based on past experience, customers typically want chiller equipment delivered 6 to 9 months before the site needs to be operational. For sites that are expected to be operational in early 2027, this would suggest equipment orders no later than Q2 or Q3 this year. Timing is always difficult to predict because there are multiple moving pieces on the customer side, but the time line we have seen to date is completely consistent with our historic sales cycles. Projects can also go through stop-start cycles before closing. For example, in the past 5 months alone, we have had 2 instances where potential customers have told us they were ready to place the purchase order, but then hit unforeseen delays on their end. However, as you can see, we have multiple opportunities of various sizes, thereby increasing the odds in our favor. One project is an expansion of an existing data center. They plan to use our dual power source chiller to handle new tenants. Another is in the final stages of tenant negotiations and expects to use our Dtx chillers to maximize IT capacity. The same developer also has a second project of a similar size and another of a larger scale. Next, there is an opportunity for a demonstration project with an established data center owner for up to 40 chillers. This developer evaluated the cost of power from our chillers against the alternatives and found the value highly compelling. However, they were also looking for some independent validation of our chillers. We believe that the Vertiv demonstration project will be instrumental in unlocking this opportunity. The remaining projects have filed for environmental permits and are in active discussions with tenants. They represent 100 to 200 chillers collectively. We expect more clarity on construction timing as permits are granted and tenant negotiations progress. In our previous call, we had mentioned an opportunity where we have an LOI for 6 Stx chillers. Although this opportunity is progressing, we have moved this further down the list because we believe the others outlined above are moving faster and have more near-term potential. In addition to this list, we also have ongoing discussions with multiple hyperscalers and multiple other data center developers. The current time line on these projects are completely in line with projects and other industries. We also believe that closing the first few opportunities will unlock significant demand. Based on conversations with prospects, even if we have chillers and other critical cooling applications, many data center owners would still like to see our chiller in other data centers or cooling AI loads. We believe this concern will be addressed with the Vertiv demonstration project and some of the near-term opportunities. Aside from data center projects, we are expecting chiller orders from other segments such as cannabis, hospitals and comfort cooling. These represent at least another 6 Dtx tillers. Expected delivery is the fall and winter of this year. We're also seeing a gradual resurgence in cogeneration leads as utility rates rise across the United States. Given the significant amount of interest in our dual power source chiller, we wanted to make sure we could handle a step change in order volume. We have now qualified a vendor for the sheet metal and refrigeration assembly. This vendor already built hundreds of similar refrigeration and sheet metal assemblies for a large chiller company. We have also qualified an electrical assembly vendor for the power electronics and are in the process of qualifying a second vendor. We are also presently building some inventory of both fuel power source chillers and Dtx chillers. Our engineering team has been iteratively improving our design for manufacturability and to reduce build time. Given the cash usage over the last 6 months, I would like to provide some context and then the plan for the next 9 months. Given the size of the pipeline, one of the concerns we had was being able to handle aggressive delivery schedules. As a result, we expanded cash on several fronts simultaneously to get everything we needed to do done. Some of these uses of cash included manufacturing capacity expansion, performing the testing and improvements needed for our dual power source chiller to operate under data center conditions. We also hired a marketing firm that specializes in data centers. In addition to the above, in Q3, we invested significantly in the service group, especially in the greater Manhattan area. We have found over the last 2 years, despite increasing our service contract rates greater than inflation, we found that margin on the cogeneration products was reduced in the greater Manhattan and Toronto service centers. The chiller product continues to maintain solid margins. The cost of labor and increased travel times between sites is one of the biggest contributors to this decline in margin. To counteract this, we invested in new engines in this territory with the latest performance improvements. This allows us to increase service intervals by at least 50%. We expect this to lower labor costs per hour of operation. In Q4, we saw an increase in both 1 hours and margin compared to Q3 in these territories. We will continue to monitor, and if needed, institute aggressive price increases or cost reductions where needed. Our current cash position is $10 million. By Q2, we plan to cut the cash burn down substantially. From 2023 to mid-2025, we managed with $2 million of cash, including a factory move. Roger will discuss the results on the financial plan going forward.

Thank you, Abinand, and good morning, everybody. I'll start with the fourth quarter results. Our revenues for the quarter decreased by $800,000 in the fourth quarter to $5.3 million, which compares to $6.1 million in the fourth quarter of 2024. And this is due to the decrease in product shipments and a reduction in Energy Production revenue. Our gross profit also decreased by 28% in the fourth quarter compared to the comparable period in the prior year period. And this is due to the decrease in our Products revenue and an increase in our service costs. The gross margin decreased 8.2% to 36.8% in the fourth quarter from 45% in the comparable period in 2024. As Abinand touched upon earlier, our Services margin was lower compared to the same period last year, but has increased compared to the third quarter of this year. The quarter-over-quarter changes in revenues and gross margin will be discussed further in our segment performance slide. Our operating expenses increased 57% in the fourth quarter of 2025 to $6.1 million from $3.9 million in the fourth quarter of 2024. This is due in part to a $900,000 increase in the asset impairment charge in our Energy Production segment and increased operating costs in our Services segment and increased cost in our Production segment which we incurred for the manufacturing expansion that we're working towards. And we also saw increases in our R&D costs, which were incurred to continue the development and refinement of our dual source chiller, which is focused on our entry into the data center market. Our net loss increased in the fourth quarter to $4 million from $1.1 million in the similar period in 2024. And this is due to the reduction in sales and gross margin, the asset impairment charge and an overall increase in operating expenses. We will discuss expenses in more detail in our full year 2025 numbers. Moving to adjusted EBITDA. The adjusted EBITDA loss for the fourth quarter was $2.4 million, which compares to about $700,000 in the same period last year. And again, this is due to lower sales and gross margin and the increase in operating expenses that we experienced. Moving to performance by segment. Our Products revenue decreased 68% to about $0.5 million in the current period from $1.4 million in the fourth quarter of 2024. This is due to a delay, as Abinand suggested earlier, a couple of projects which we expected to ship in the fourth quarter of 2025, but we now expect to close these orders in the next few months. As we have discussed in the past, our Products revenue has significant variability quarter-to-quarter, and it's borne out in this past quarter. Our Products gross margin decreased to negative 6.9% from 30.9% in the fourth quarter of 2024. This is increased unabsorbed labor. This is labor that we're using to work towards increasing our throughput. An increase in our inventory reserve, a slight increase in warranty costs and all of these costs, which have a disproportionate impact on margin to the revenue decrease. Our Services revenue increased 9% quarter-over-quarter to $4.5 million in the fourth quarter compared to $4.1 million in the comparable period in 2024. And this is due to higher billable activity and higher operating hours of the equipment from our existing service contracts. Our Services margin decreased 7.4% to 43.4% in the fourth quarter of 2025 from 50.8% in the fourth quarter of 2024. This is due to increased labor and material costs in the greater New York City area. Our Energy Production revenue decreased 28% in the fourth quarter of 2025 to just under $4 million compared to $5 million -- I'm sorry, about $555,000 in the fourth quarter of 2024. And this is due to contracts that expired early in 2024 and some of which expired late in 2023 and the temporary site closures during the year. Our Energy Production gross margin decreased to 13.7% in the fourth quarter of 2025 from 39% in the fourth quarter of 2024. And this is due to an increase in cost with our Energy Production business. Moving to the full year 2025 results. Our revenue increased 19.7% or $4.5 million in 2025 to $27.1 million compared to $22.6 million in fiscal 2024. And this is due to a significant increase in our Products revenue and an increase in our Services revenue. Our gross profit decreased about 0.5% in 2025 compared to 2024. And the decrease in the gross margin was 7.3%, which decreased from 43.6% in 2024 to 36.3% in fiscal 2025. And we'll review year-over-year changes in revenues and gross profit further in the segment performance slide. Our operating expenses increased 25% in 2025 to $18.1 million from $14.4 million in 2024 due in part to the $900,000 increase in the asset impairment charge in our Energy Performance segment, increased operating cost in our Services segment, an increase in cost in our Products segment -- again, that's due to the manufacturing expansion -- and increased R&D cost, which we incurred to continue, again, the development and refinement of our dual source chiller, again, that we're focused to utilize in the data center market. Our net loss increased in 2025 to $8.2 million from $4.7 million in 2024. And the loss is due to, again, lower Services and Energy Production gross margin, the asset impairment charge and increasing operating costs. And we'd like to point out that we are working on a program to reduce our OpEx to levels that are consistent with levels from 2024 -- the 2024 spend, I should say, and anticipate to see reductions to commence in the second quarter of this year and further expansion of those reductions in the third quarter and the fourth quarter. Our adjusted EBITDA loss was $5.6 million in 2025, which compares to $3.6 million in the same period last year. And this is due to lower Services and Energy Production gross margin and the increase in operating costs. Reviewing our performance by segment. Our Products revenue increased 105% to $9.1 million in the current period from $4.4 million in 2024. And this decrease -- I'm sorry, this increase is due to an increased chiller and cogeneration revenue that was recognized in the first 3 quarters of 2025. And as we mentioned earlier, the -- the decrease was -- this increase was partially reduced by or offset by the decrease in production revenue we experienced in the fourth quarter due to project delays. The gross margin for Products improved 1% in 2025 to 33.2% from 32.2% in 2024. Our Services revenue increased 3% year-over-year to $16.6 million in 2025 compared to $16.1 million in 2024. And this is due primarily to higher billable activity and an increase -- a slight increase in operating hours of the equipment that's being serviced. Our Services gross margin decreased 8.9% to 38.6% in 2025 from 47.5% in 2024. And this is due to increased labor and material cost incurred as we invested in new engines and new performance upgrades to the sites in New York City. The intention of these investments is expected to reduce labor hours needed per system going forward. The decline in margin is presently only in the cogeneration equipment. Our chillers continue to generate expected and very strong margins. Therefore, we plan to institute both price increases for our cogeneration equipment in the Greater New York City area and make significant cost reductions in a territory -- sorry, significant cost reductions in the territory to restore this region to higher profitability. Our Energy Production revenue decreased 37% in 2025 to $1.3 million from $2.1 million in the fourth -- in 2024. And again, this is due to contract expirations in the latter part of 2023 and early 2024 and temporary site closures for repairs. Our Energy Production gross margins decreased to 28.3% from 38.0% in 2024. That concludes the results review, and I'll turn the call over to Abinand for his closing remarks.

Thank you, Roger. So I think the single biggest improvement that we have seen in the last 5 months is really the securing of the first demonstration project. I personally believe that this will be the catalyst for everything else that will come and will also unlock the much broader opportunity that we have been pursuing. In a world where AI tokens are, per unit of power, is the new metric, we provide the simplest and most cost-effective way for a data center to obtain more power, which directly results in more compute and more revenue. We have a robust pipeline of opportunities, the demonstration project, and I think all the pieces are coming together to unlock the larger projects on the multibillion-dollar data center cooling opportunity. Thanks for listening, and I'll open the floor for questions.

[Operator Instructions] Our first question is coming from Chip Moore from ROTH Capital.

Abinand, maybe starting there where you finished on the Vertiv demonstration project. So I think you said expect this to shift later in Q2. Help us think about how quickly that should be up and running? And real-world data, when that should flow? And I assume this plays a role in the 25 to 50 megawatts that you mentioned that they're in the process of designing. Just -- when could we conceivably see some of those move to orders once that demonstration project is up and running?

So I think the two pieces will move concurrently. So the -- this unit, if it ships in Q2, is actually going to go and get tested almost immediately in their thing. Because what typically happens in a lot of these projects that we have is the end customer would like to see -- how is the chiller going to run in, let's say, 110-degree ambient at a certain chilled water output. So what this does, because we -- with our conventional Dtx chiller, our own test cell, we can run a lot of those conditions here. But we can't control for 110-degree ambient or 120-degree ambient. So the Vertiv test chamber allows us to do this, but it also acts as a way for a potential customer to come and see it as well. Right? That's usually a very good way to close projects. So the designing of the projects in the background, that's going to happen, currently, and they're going to continue marketing, and they're getting opportunities. I can't talk very specifically on timing on their projects because at this point, it's either confidential or we don't yet have enough clarity on it. What I think it's going to do, this demonstration project, the hope is to have it actually running no later actually than the end of Q2. What we think it will do is act as both providing the feedback for some of the bigger projects that we have and also for some of the potential customers where they may want to use their chillers in certain environments like, let's say, Texas. They want to know that this chiller would -- what output it will put out at different ambient temperatures. So that's also what you will get out of this, and it's really, I would say, an independent validation as well of our chiller, right? And a massive load of support for Vertiv that they're essentially putting this together.

Understood. That's helpful. I appreciate that. And maybe for your own internal pipeline, that slide, your highest -- I think you stacked them in order, that one where you're in final stages of negotiation, and you could see orders here. I think you said Q2, Q3 for a '27 type of project. Just any more you can expand on that on where that stands, what they want to see, when that could move forward?

So I don't want to predict timing because I think it's extremely hard to predict the timing. What I will say is the smaller projects -- actually -- firstly, the Dtx is a -- there's -- we've already got the test data, and a lot of these smaller ones, they've seen our equipment in other places. There, the smaller projects also find it easier to get tenants. So the odds of it moving faster is much, much higher. And also, things like financing and getting approvals for environmental permits tend to be much easier for the smaller projects. Because some of the greater than 100-megawatt projects, you've got more hurdles to go through on the back end to make sure the local site approvals, all of that happens without an issue, and then also having the tenants. What we're seeing more broadly in the data center industry is you're getting a lot of the neocloud on the -- as potential tenants. And there's quite a few of the smaller scale neoclouds that are interested in these kind of smaller scale projects, which makes it much more likely that these things will go through. But I don't want to predict timing. All I know is that in many of these cases, the customer expects to be operational by early next year. So to really be able to get equipment in order to get everything delivered in that time and actually constructed on site, they need to move quickly.

Helpful. Very helpful. And maybe a last one for me, just on the manufacturing side. It sounds like you've made some good strides and you've got some potential with Vertiv as well. Just -- what's the highest priority? What are you focused on? What do you need to do here to get prepared?

I think we've gotten most of the key pieces together now because a lot of that was getting the subcontractors qualified and really get first articles from them and then get the first article checked internally to make sure that it meets our quality standards. And that if we can get the different pieces from them and it arrives at our factory, we can just do the final assembly, test it, get it out the door. So getting the subcontractors qualified was really the key. And I think we've now done that. And also, these subcontractors have significant scale-up capability already. So there is -- I believe -- like those pieces are now together, especially for the dual power source chiller. The Dtx, I think our supply chain was reasonably robust to start with. So I think we have the ability -- because a lot of the bigger components from the Dtx are built by some very large companies already. So we can get scale up from them without too much of a problem. It was a dual power source chiller really with the size of the machine as well as making sure that we weren't having a lot of time on the floor in our factory here to get some of the bigger components built outside and brought in. That was really the key. And I think we've now done that.

Got it. And maybe the follow-on. With a lot of that heavy lifting done, it sounds like maybe you don't need to sacrifice non-data center orders if you start to see demand pick up in some of those other areas?

Correct. I think we are going to see quite a bit more in non-data center projects as well. We -- on the sales efforts and the marketing efforts, we essentially split the sales team to have some people handling non-data center and some handling the data center. Part of the lumpiness on the product side is just -- what we've seen overall in the industry is we used to get a consistent amount of small multifamily, like 1 unit, 2 unit orders that were kind of steady flow. With the anti-gas sentiment in some of the bigger cities like New York and Boston, that portion had declined. It's starting to come back. But what we are -- what we have a very good pipeline of is multi-unit, larger projects that are going into bigger buildings. But the problem with those projects is that it -- if 1 project gets even slightly delayed, you've basically moved out 3 or 4 units at a time. So there's a little bit of timing issues there, but I think the pipeline is very, very robust on that.

Our next question today is coming from Alexander Blanton from Clear Harbor Asset Management.

Yes. I'm interested in your outsourcing strategy. Because clearly, to get significant orders from data centers, they're going to have to be sure that you can deliver the quantities that you're talking about. So could you just go into a little more detail about how that's going to work? What things are going to be outsourced? And as I take it, it's going to be -- these components will come to your factory and just be assembled. And so you have obviously changed your manufacturing process significantly in doing that. A little more detail?

Yes. No, that's a great question. So let me start with the end portion, which is we didn't actually change our manufacturing process necessarily. When we design that -- the dual power source chiller, we always designed it with the option of being able to either do all of it in-house or have large portions of it built in subassemblies that then came internally. The other thing that we always did on that product was to use a lot of components that are built in larger volume to start with so that we could, with volume, also see an increase in margin. So in other words, it was -- but when you think about the dual power source chiller, right, I think of it in sort of certain blocks. You have 1 big block, which is really the refrigeration assembly. So that handles the -- it's similar to an electric chiller. It has your compressors, it has your fans, it has your sheet metal assembly. Then you have the power assembly, which has the engine and the generator. And then we have the power conversion or the electronics, which is really your -- the dual power source technology. The combination of the engine and the inverter and power electronics technology is very similar to our InVerde. The biggest challenge we have in our manufacturing space is just in terms of just physical footprint on floor space. And also -- the -- like we historically have built numerous InVerde units. So we can build those in volume very, very easily. As long as electronics show up preassembled, we can make it with our engine system, our generator and essentially build that power electronics and engine package very quickly. The refrigeration assembly, it's -- you're dealing with a lot of sheet metal. It's not necessarily something that is best done in our factory here because if we can reduce time on the floor by having somebody that built similar assemblies for other electric chiller companies, then we can essentially have that portion prebuilt, pretested, it can come to us, and get made it with the power electronics assembly. The other big advantage of really focusing on the power electronics assembly is it gives you -- our power electronics and engine assembly is it gives you other options as well, including beyond just pure chillers. It gives you the option of being able to power things like fans in a data center or other loads and where you can arbitrage the two power sources. So you can not only get a volume boost, but you can also open up a broader market. But going back to your question of putting these two together, it's -- essentially, a lot of that sheet metal assembly, it's better done by people that's all they specialize in. And they have the volume throughputs. And usually, they're vertically integrated, starting from the sheet metal, all the way to all of the different components, and they're already buying a lot of those subcomponents in volume. So getting that as a single assembly, bringing it into our factory, making it with the power system and then shipping it out is one way to do it. Eventually, we could even ship that power assembly to the sheet metal manufacturer, they do the final assembly. Everything is pretested at each location, and then it's shipped. So there's different ways to significantly improve volume and also hit delivery times using that approach.

Given the constraints, what's your effective capacity, then? If you have your subcontractor do the assembly, it seems to me that it expands quite a bit.

Yes. I mean, I would still say today, I would use what I'd said in our -- I think the last call or the call before, where -- I'd say about 100 units is where we're targeting. I believe it can be increased further from there. But that seems like with a little bit of a ramp-up that's fully achievable. And then from there with some optimization, it's likely that you can increase further from there.

100 units over what time period?

I would say per year. I think it's possible to go higher than that, but I think this is something that we've looked at and looked at the details on what it takes to get there. And it is possible to scale up substantially from there. It's just this, I think, is a good starting point. It will allow us to get many of the opportunities that we have on that pipeline. And then from there, there are different ways to figure out how you scale from that.

And what's the dollar volume of 100 units?

I don't want to comment on exact dollar numbers since we don't put pricing out. But I would say it's at least 3x to 4x of what we've done in our highest year. So I'd say at least $30 million to $40 million of product, likely more.

And that would be just for the data centers?

Correct.

It doesn't include the cogeneration and other products for other markets, right?

That is correct.

Next question today is coming from Barry Haimes from Sage Asset Management.

My question relates to the master agreement negotiation or renegotiation you're doing with Vertiv. Could you talk a little bit about -- what are your goals in doing that? And what are their goals, given that you already had an agreement?

Yes. So if you look at the marketing agreement we have with Vertiv, right, the way it's structured, it says that this is kind of the placeholder while we go through the full master agreement. So if you look at the marketing agreement, it has various terms that are to do with supply and delivery, things like that, that currently are not binding terms within that agreement. And all it does is it takes that marketing agreement and expands it so it has all those different portions. It was always designed from day 1 to go into that broader agreement so that we could actually supply as an approved supplier and have this marketing portion rolled in as part of this broader agreement.

Okay. Great. And what's the timing on when you expect that to get finalized and signed?

At this point, I can't really comment on timing because it's ongoing.

[Operator Instructions] Our next question is coming from Kris Tuttle from Blue Caterpillar.

Great. A couple of questions from me. First of all, I know it's not the sexy part of the business, but I wanted to go back to what you were talking about in terms of some of the investments you've had to make on lowering your service cost. I mean, these are mechanical units, which, I guess, in most cases, have a nonlinear graph of support and maintenance costs over time. And so I'm a little -- I wanted to understand more about how you're situated in terms of if you have older inventory out there. Is this -- how much are you still sort of on the hook for in terms of what would normally be kind of a customer cost? It seems like you guys had to spend some of your own money on that in Q4.

Yes. So that's a great question. So it is -- so the way the service contracts work. On the cogeneration units, we charge per run hour. So every unit that machine runs, we charge per run hour on those. And the service contract includes components, everything inside the cogeneration cost. So -- and we -- most contracts are either 3-year, 5-year, but they auto -- in many cases, the customers renew it. The reason the contracts were structured that way was so that the customer has some essentially predictable expenses, and it would allow the business to have a recurring stream of cash flow. But as the costs in places like New York have gone up, like the labor costs have gone up substantially and the time to get from sites has gone up, the labor efficiency has gone way down. We've also seen material cost increases, but we're -- in the other territories, the material cost increase has been absorbed by any increases in service contract rates. So we had two choices. We could either turn around and say, "You know what, these service contracts are no longer profitable. We either get out of that service contract or figure out a way to make those service contracts profitable." The concern we had with essentially walking away from some of these contracts was that over time, like just as you're starting to get your data center side of things ramp up, the risk of reputational hit, right, if you walk away from a number of service contracts is high. So we felt when we looked at the numbers that with putting in new engines, we could substantially increase the service intervals. I mean in our test cases, we got almost a 2x increase in service intervals. I commented about a 50% increase on average. If you can do that and you don't have to go to the machine as often, the numbers suggested that we should get back up to our previous margins, which were somewhere around the 50% gross profit margin. That's kind of why, yes, it was very expensive in the short term. Right? And it took us -- I mean, it pulled our loss down. But at some stage, it -- if the units -- if we can increase that service interval, the numbers should play out. If for some reason that they're not happening, then we'll go back and just raise our prices. Or in some cases, we will get rid of service contracts that are not profitable anymore. That's -- but in the short term, we felt that this was a better way to go, especially because you've got ongoing cash flow that comes from this. So it's much better to figure out a way to make them profitable then walk away from them.

Yes, that makes a lot more sense now. Were those the territories where you feel like you had the problem to address?

Yes. It's really been the urban environments. Like just -- actually, just predominantly Greater New York, and to a certain extent, up in Toronto. Those are the 2 territories that really pulled it down. The chiller services in those territories continue to make good money. And part of that is just because the chiller product is built a little differently. It tends to be a flat rate contract. And it also tends to have already very long -- like one of the reasons actually we did this was because the chiller service intervals are much longer to start with. And that's why we took some of those improvements from the chiller product, applied it to the power generation thing and said if the chiller can make money, we should be able to make the same things with the same kind of structures and same improvements to the cogeneration and get the cogeneration units making the same margins.

Okay. All right, two other quick ones from me. Just -- one of them, can you remind us in terms of when you -- like your pipeline of business with the data centers and all that, and we understand they've been delayed. They've been delayed for lots of other companies. It's been very topical. What are the terms? In terms of the revenue recognition and payment terms, are there any upfront payments involved like deposits, do you recognize revenue on delivery? Is there a customer acceptance period? And then what are the typical payment terms where you'd be getting that cash?

This is Roger. Typically, we require a down payment from customers. It can go from 25% to as much as 40%. And then revenue is recognized when title transfers. In most cases, it's ex-factory. Sometimes, it's a destination. But for the most part, we recognize revenue when the product ships. Obviously, there are some holdback on the revenue rec for start-ups and minor things like that. But for the most part, when we ship a unit, we will recognize the revenue at that point.

Okay. And payment terms, 30, 60, 90 on the balance?

Payment terms are generally 30 days upon customer acceptance.

Okay.

So usually, that would add another 30 days to [indiscernible]. Generally between 30 and 60 days, I'd say.

Okay. And then last question, really helpful update on the pipeline. It sounds like things that got delayed from Q4, like they got delayed a bit into -- like not just looking into March, in the March quarter. I mean, it's -- we're now almost done with the March quarter. It sounds like expectations should be we're going to see more deliveries really starting more in Q2. Am I sort of -- did I not hear that right? Or I know it's a little bit awkward in terms of timing, but it seems like more of these things are going to be flowing starting in 2Q, Q3?

So the -- there's two portions of it. So some of the projects there, that's non-data center-related projects, right? Those -- there are some cannabis, some non-cannabis like hospitals and comfort cooling and things. Those are the ones that pushed out a little bit. The data center pipeline, it has been, I think, within the range. Like your typical sales cycle is longer than what we've seen already on the data center stuff. So I would say the two are likely to come together around the same time. With data center projects, it's very much -- something could suddenly start moving [ equipment, like ] close the projects as soon as they get a tenant or a lot of pieces start to move very quickly after that. With the non-data center projects, usually the timing is contingent on if they're doing, let's say, air conditioning load, then they would usually do that off season. So they would plan to take equipment deliveries in the fall and winter so that they could do the construction of the chiller plant in the off-season. So that typically moves that timing. And then with cannabis, a lot of it is just tied to their financing timing. If they can get financing, the project moves.

Okay. So therefore, you can have some reasonable volume in Q1?

Reasonable. But I think a lot of this right now, right, I think some of the bigger projects will close. I think we've got enough over there. That timing is very hard to predict.

Yes. Understood. Understood. All right. I've got some other technical questions, but those are best left for another time.

Our next question today is coming from Maj Soueidan from Geoinvesting.com.

A quick question. I guess I just recall, I think from previous conversations with you, at least maybe during earnings calls, that you really see hyperscalers as an opportunity. But in the last few calls, you've kind of mentioned them. So I'm trying to understand what's changed there. Is that from the cooling side or power side? And maybe you could maybe touch on that a little bit for a few seconds?

Yes. So that's a great question, Maj. So originally, we felt that the hyperscalers, the validation process and a lot of this thing might just be out of what we would be able to do. But as we have started to go after many of these projects on the colocation side of things, we found that we've gotten -- we've either met hyperscalers at trade shows or direct outreach to hyperscalers has resulted in actually very positive engagement. And the -- I think with a lot of this, I can't really comment on specifics on any of them, but it does seem like there is significant interest on -- from the hyperscale side of things. So it -- we're kind of letting the hyperscale conversations continue. And we'll see whether that leads into projects or pilot projects or what that leads to, we don't yet know. It's just they appear to be happening concurrently. So we're -- we're just -- we're going to pursue it. We have presented to a number of them, and there has been clear interest on the technology for the chiller side. So -- and I think the power side, at this point, we're not leading with it, but there may be interest on some of the ancillary loads. But that is something that -- the chiller seems to have significant interest.

Great. I have two more additional questions, real short. I'll start with the service contracts, that business and the things you've done to decrease maintenance needs on-site, increase in service intervals. Does it make sense to do that in other jurisdictions other than where you're at now to increase margins there, too?

Yes. So we are in other jurisdictions. We're doing those -- because one of the biggest costs on the service side of things is your oil change intervals and your engine component intervals. It's better to do that when you're replacing the whole engine rather than do it on an engine with higher time. So we typically -- in other service territories, we're doing those changes, but we're doing them as we get rather than do it proactively on units. Because at some point, right, there is an expense associated with it. So it's better to do it where you're going to have the biggest return on that expense. So in other territories, we are doing it, but on a much more gradual basis.

So at some point, you could see the overall gross margin on that business go up as places like New York catch up to being where they used to be in other areas, maybe getting an improving gross margin profile. Am I understanding that correctly?

Correct. I mean, the target is across the whole service territory. We would like to have at least a 50% gross profit margin.

Okay. And finally, I just have a question on the modular data center space. I don't know. You kind of mentioned it in the past, like in fleeting comments about seeing that market kind of heating up a little bit. I was wondering if you could give us a little bit of color on what you're seeing there and if you are, if you can? And do you see opportunity for Tecogen to play in that growth potential there?

Yes. So we have seen quite a few leads in that space. As yet, nothing has got far enough that they made it into that opportunity list that I presented. We believe, just looking at the broader picture, that there's going to be a lot of modular data centers being built, but also, there's going to be a lot more smaller scale data centers being built. Because we're seeing some of the really large data center campuses run into other hurdles such as local opposition from people that live in the area or permit problems on the really large data centers. So I think there's going to be a push for the modular as well as the smaller scale data centers being built in urban environments. And in that sense, our product is like a perfect fit for that market. At this point, we're not seeing...

And the cooling side, right? Sorry.

Correct.

Thank you. We've reached the end of our question-and-answer session. I'd like to turn the floor back over for any further or closing comments.

Thank you very much for listening. And if anybody wants a further conversation on any of this, management is available to have more in-depth discussions. Thank you.

That does conclude today's teleconference and webcast. You may disconnect your line at this time, and have a wonderful day. We thank you for your participation today.