Tecogen (TGEN) 2026 年第二季財報電話會議：資料中心案源與逾 800 萬美元在手訂單

重點摘要

• 2026 年第二季營收年減 21% 至 580 萬美元，主因是產品營收大幅下滑 64% 至 110 萬美元。
• 毛利減少 11.9% 至 220 萬美元，但在產品部門毛利率提升的支撐下，毛利率自 33.8% 改善至 37.8%。
• 淨虧損自 150 萬美元擴大至 220 萬美元，同時受產品銷售減少及營運成本上升影響，調整後 EBITDA 虧損自 120 萬美元增加至 170 萬美元。
• Tecogen 舉辦了 12 場資料中心產品展示，涉及 12 家機構，其中包括 8 家潛在直接客戶。管理層表示，參與的資料中心營運容量合計超過 8 吉瓦（GW），約佔目前美國資料中心總容量的 15% 至 20%。
• 基礎業務在手訂單超過 800 萬美元。管理層預期未來幾個月將再敲定 200 萬至 300 萬美元的專案，第三季產品營收也將有所增長。
• Tecogen 正建立雙源冰水機與發電模組的庫存以縮短交貨時間，雖然管理層也在這項舉措與現金流及營運資金限制之間進行權衡。

核心財務業績

在 Tecogen 精簡數個服務中心的人力後，營業費用較 2026 年第一季減少約 40 萬美元。

業務與營運表現

產品部門

產品營收年減 64%，自 320 萬美元降至 110 萬美元。去年同期受惠於向尋求《降低通膨法案》稅務抵減之客戶出貨汽電共生系統。

產品毛利率自 29.3% 上升至 48.5%，反映出產品調漲價格以及更佳的產品組合。管理層強調，各季度之間的產品營收可能會大幅波動。

Tecogen 的資料中心策略仍以其雙源冰水機與模組化發電產品為核心。管理層表示，近期的產品展示展現了該系統提供無水蒸發閉迴路冷卻、低噪音運轉，以及接近零氮氧化物與一氧化碳排放的能力。

該公司在追求小型專案的同時，也積極爭取超大規模與知名大品牌資料中心營運商。管理層表示，展示期間討論了具體專案與交貨日期，但並未宣布新的資料中心訂單。

服務部門

在可收費活動增加、設備運轉時數上升以及透過 Aegis 取得之合約帶動下，服務營收年增 10%，自 400 萬美元成長至 440 萬美元。

服務毛利率與去年同期大致持平。然而，該部門產生了約 30 萬美元的一次性成本，與 2026 年出售給 SDCL Kyotherm 之案場的電動冰水機相關。管理層表示，若扣除這些成本，服務毛利率將會高出約 7 個百分點。

Tecogen 還精簡了服務部門的人力，適時調整合約定價，並開始與較大型客戶合作以提升設備運轉時數。

能源生產部門

由於特定案場的正常運作時間改善，能源生產營收成長 35%，自 17 萬美元增至 24 萬美元。毛利率自 25.2% 下降至 9.0%，反映出本季認列了略低於 10 萬美元的保證短缺金額。

管理層展望

受惠於超過 800 萬美元的基礎業務在手訂單，管理層預期第三季產品營收將會成長。公司亦預期未來幾個月將再敲定 200 萬至 300 萬美元的專案，這將有助於進一步收取客戶預訂款並改善現金流。

第二季實施的服務成本降低措施，其完整效果預計自第三季起顯現。管理層相信，服務營收增加、定價調整及營運改善將能支撐更強勁的毛利率，並大幅提升毛利金額。

Tecogen 已開始在取得正式訂單前，提前建立額外的雙源冰水機與發電模組庫存。其目標是縮短交貨前置時間，以便在資料中心商機轉化時能快速回應，同時分階段進行支出以控管現金流。

風險與關注焦點

• 資料中心對產品的興趣尚未轉化為新宣布的訂單。管理層坦承銷售過程比最初預期的要長。
• 較小型的資料中心專案仍面臨承租人決策、施工許可延誤以及其他外部因素的影響。
• 製造產能與交貨前置時間可能會與大型客戶產生摩擦。Tecogen 正透過庫存投資與試點規模部署來減輕這項限制。
• 若預期專案延遲或未能敲定，在收到採購訂單前建立庫存將會增加營運資金與現金流風險。
• 產品營收在各季度之間仍具波動性，同時該公司持續呈報淨虧損與調整後 EBITDA 虧損。

分析師問答集錦

管理層澄清，12 場展示共涉及 12 家獨立機構。其中 8 家為潛在直接客戶，其餘則包括工程師、合作夥伴及一家晶片製造商。規模較大且具策略重要性的展示被安排在流程的後期進行。

Tecogen 認為，試點專案與改造工程可以作為進入大型資料中心客戶的途徑，而無需預先準備全部的製造產能。潛在應用包括在園區擴建期間增加 Tecogen 設備、將其與現有的電動冰水機並排安裝，或將現有冰水機作為備用系統。

這些展示是由 Tecogen 自身的業務與行銷努力所促成，代表了與 Vertiv 合作關係並行的一條途徑。管理層表示雙方關係「整體而言非常良好」，但在未獲事前批准的情況下，拒絕提供主採購協定或具體專案的詳細最新動態。

管理層亦指出，更廣泛的電力限制、高昂的公用事業電費，以及電氣設備漫長的交貨期，皆是支持基礎業務的因素。多戶住宅與其他建築客戶對汽電共生及標準發電設備的需求正開始改善。

法說會完整逐字稿

---

完整財報電話會議逐字稿

管理層陳述

Operator

Greetings, and welcome to the Tecogen Q2 2026 Conference Call. [Operator Instructions] As a reminder, this conference is being recorded. It is now my pleasure to introduce your host, Jack Whiting, General Counsel. Thank you. You may begin.

John Whiting

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 second quarter 2026 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 make about the company's expectations, plans and prospects constitute forward-looking statements for purposes of the Safe Harbor provisions of 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 Q2 2026 earnings and on our website.

I will now turn the call over to Abinand Rangesh Tecogen's CEO, who will provide an overview of the second quarter 2026 activity and results; and Roger Deschenes, Tecogen's CFO, will provide additional information regarding Q2 financial results. Abinand?

Abinand Rangesh

Thank you, Jack. Welcome to Tecogen's Q2 2026 call. Today, I would like to start with the key question that many of you may be wondering about. Given that Tecogen has been working on the data center strategy for a while and the product seems to make perfect sense for the market, what is taking so long for orders to close? Is there something missing in the story?

To answer those questions, I believe we need to start with the objective. In my view, the objective is, what is the fastest way to maximize the value of Tecogen, given we have great technology but have resource constraints as a small company. Originally, our strategy was to target smaller data centers. We could then use that as a reference to get the bigger names interested. As we got bigger names interested, the company's technology starts to become incorporated by big-brand data centers, and now we have technology that everyone wants, making Tecogen very valuable. Now we may have an opportunity to go straight for the end goal, which is the hyperscale and big brand developers.

In March this year, I shared an opportunity pipeline with you. Many of these opportunities were with data centers that were of smaller scale. While many of these projects will likely close, these projects are subject to external factors such as tenants, delays in construction permits, et cetera. The big -- brand data centers don't have the same limitations. In addition to these smaller opportunities, we have been simultaneously working on access to the larger names. It has taken an enormous amount of work on the part of our sales team to make inroads on some of the larger names. But now I believe we have managed to gain some serious traction.

Over the last 2 months, we have hosted 12 product demonstrations, 6 in-person and 6 virtual. The in-person group included 4 data centers that are hyperscale or build-in hyperscale campuses and partners. The virtual visit group included chip manufacturers, engineers and another 4 data centers. These data centers collectively represent greater than 8-gigawatts of data center capacity operating today and multiple gigawatts in construction. Although I cannot tell you the names of the data centers that attended, I'm sure you will know many of them. To put this into perspective, the data centers that attended, either virtually or in person, collectively represent 15% to 20% of present data center capacity in the U.S. The natural question at this point is, does attending this product demonstration show any real commitment or interest?

For the in-person group, given how busy data center engineers and senior level people are, locking off the whole day for multiple people is a significant commitment. It has taken many months of finding the right internal champions with influence at these larger companies and educating them on the value of our products before we got to this point. Second, beyond the list that attended, we are talking to other large and hyperscale data centers. The ones that are earlier in the process declined attending. This is consistent with our experience in other markets where once a potential customer attends a site visit, our close rate is usually high as they are serious about entering a business relationship. Lastly, I think the market has also evolved. In addition to power, there are some key challenges that the bigger data centers are facing that our products solve.

As you may have read in the press, many data centers are facing opposition for on-site water use, noise and air pollution. Here's where our products really shine. We solve all three problems. The dual-power source chiller is closed loop, so there's no water evaporation. Our products already operate in noise-sensitive environments, such as on rooftops next to penthouse apartments. Recently, during the scorching 4th of July weekend, black smoke could be seen in Northern Virginia from continuous diesel generator usage from data centers. Diesel generators were being used to avoid a blackout as the utility grid was running out of power and forcing data centers to shed load. Therefore, unlike a diesel generator that attracts noise complaints and pollutes, data centers can install our products without worry. This is where seeing a product demo brings these benefits to life. There is a difference between seeing a data sheet with a [indiscernible] reading and hearing our products in person or reading about near-zero NOx and carbon monoxide emissions versus seeing the readout from an emissions analyzer in real time.

There's also a big difference between saying you can add extra megawatts for compute during hot days versus watching in real time a jump in cooling load and seeing the electrical power remain capped or saying you can provide uninterruptible cooling during a power outage versus watching someone shut off the electrical power to the chiller while it's running. These big name data centers have permitted projects, capital and the ability to shape the whole industry. The feedback has been extremely positive across the board and specific projects, delivery dates, et cetera, have been discussed, but we feel confident enough to begin building some inventory of our dual-power source chiller and power gen modules to get a head start.

In addition to the data center push, our base business backlog now stands at greater than $8 million. In addition to what is already in backlog, we expect a further $2 million to $3 million in projects to close over the next few months. This means we expect Product revenue to increase in Q3, and we expect to collect more deposits, improving cash flow. Although our focus has predominantly been on our Data Center strategy, during the last call, we mentioned that we expected to reduce expenses in our Service group to increase margins. During Q2, these reductions were made mid-quarter, so the full impact will be seen beginning in Q3. Onetime costs in Q2 reduced margin by around 7 percentage points. Without these onetime costs, margin is starting to recover in Service.

In addition to cost reductions, we have also started to make contract pricing adjustments where appropriate and are working with customers, especially on larger sites to make improvements that would help them increase run hours and help us increase revenue and margin. Given that our Service revenue was substantially higher year-on-year, this continuous improvement is expected to result in higher margins and therefore, a significant increase in gross profit dollars.

I'll now hand over to Roger to talk about the financials.

Roger Deschenes

Thank you, Abinand, and good morning, everyone. I'll begin with the results for the second quarter. Our total revenues decreased 21% or $1.5 million in the second quarter to $5.8 million compared to $7.3 million in the second quarter of 2025, and this is due mainly to lower Product segment revenue. As Abinand indicated just moments ago, we expect revenue to increase in the third quarter based on the recent increase in our backlog and the anticipation of expected projects that we'll close in the next few months.

Our gross profit decreased 11.9% to $2.2 million in the second quarter of 2026 compared to $2.5 million in the comparable period in 2025. And again, this is due to Products segment revenue. Our gross margin increase -- our gross profit margin increased, excuse me -- by 4% to 37.8% in the second quarter of this year from 33.8% in 2025, and this is due to improved products segment gross margin. Operating expenses increased 11.6% in the second quarter to $4.3 million from $3.9 million in the second quarter of 2025. And this is due to increased operating costs in both our Products and Services segments and the general increase in operating costs incurred for the manufacturing capacity expansion that we're undergoing and the continued development and refinement of our Dual Source Chiller, which, as we all know, is focused on our entry into the data center market.

During the just concluded quarter, we reduced headcount at a few of our Service centers as we work to reduce our spend there. Overall, operating expenses decreased approximately $400,000 in the second quarter of this year compared to the first quarter. Our net loss for the quarter increased to $2.2 million from $1.5 million in the comparable quarter in 2025, and this is due to lower products segment sales and gross margin and an increase in operating expenses. Moving over to adjusted EBITDA for the quarter.

The adjusted EBITDA loss was $1.7 million in the second quarter of this year, which was higher when compared to last year's loss of $1.2 million, and this is due to lower Products segment sales and gross margin and higher operating costs.

Moving to performance by segment. Products revenue decreased 64% to $1.1 million in the second quarter of 2026 from $3.2 million in the second quarter of 2025. The revenue -- Product segment revenues last year benefited from the shipment of Cogeneration Systems during this period to customers seeking tax credits from the Inflation Reduction Act of 2022. As we have discussed in the past, Product revenue has significant variability quarter-to-quarter. Our Products' margin -- gross margin increased 19.2% to 48.5% in the second quarter of this year from 29.3% in the similar quarter of 2025, and this is due to price increases and change in product mix.

Our Services revenue increased 10% quarter-over-quarter [indiscernible] 2026 to $4.4 million, which compares to $4 million in the second quarter of 2025. And this is due to higher billable activity and higher operating hours of our equipment from both our existing Service contracts and our Aegis acquired contracts. Our Service gross margin was essentially flat compared to last year.

During the second quarter of this year, as I indicated, we spent approximately $300,000 of onetime costs at a handful of sites that we service, which are electric chillers as part of the energy sites that we sold to SDCL Kyotherm in 2029 (sic) [ 2026 ]. In fact, one of these electric chillers had a catastrophic failure at the start of the cooling season and needed significant repairs and the additional cost of rental cooling during the time the systems were being repaired. Excluding this onetime cost, as Abinand indicated earlier, our gross margin from our Service operations would have been 7% higher in the 2026 period.

Our Energy Production revenue increased 35% in the second quarter of this year to $0.24 million from $0.17 million in the 2025 period, and this is due to increased uptime at certain Energy Production sites. The Energy Production gross margin -- gross profit margins decreased to 9% in the second quarter of 2026 from 25.2% and this is due to the guaranteed shortfall of just under $100,000 that we recognized in the quarter.

I'll now hand the call back over to Abinand for closing remarks.

Abinand Rangesh

Thank you, Roger. At this point, there are two conclusions you could reach. The first is that it has taken a long time today, so [ today ] the company's prospects in the data center market are slim or the second, that it has taken a while to navigate the data center landscape and get access to the right champions of the largest data centers. Now that we have the potential addressable market is massive. After all, why would some of the largest data centers bother to take a day out of their busy schedules to attend a product demonstration. Given that we now have a range of opportunities from small data centers to the largest players, what is the likelihood that nothing closes, especially if we have some inventory on hand and delivery risk is eliminated.

If we land even a pilot project with a big brand data center, what does it do to our prospects industry-wide? Management and the Board have consistently bought stock because we believe the company has technology that solves some fundamental problems facing not just data centers, but a whole host of markets. You decide which conclusion makes most sense. Thank you, and I'll open [ for ] questions.

Operator

[Operator Instructions] Our first question comes from the line of Bobby Brooks with Northland Capital Markets.

分析師問答

Robert Brooks

So very exciting to hear about the 12 demonstrations. I wanted to unpack that a little bit more. Were all 12 demos with 12 different potential customers? And you disclosed having those 6 demonstrations at the beginning of July, so had another 6 over the last month. Was there any acceleration in that? Just curious to hear more there.

Abinand Rangesh

Great question, Bobby. So out of the 12, 8 were potential direct end customers. The rest were engineers or other people, either designing data centers or partners that could help us accelerate some of that. So really that have key influence on the end result. And then one of them was one of the chip manufacturers. So generally, it was all high-profile visitors.

Definitely, what we tried to do with regards to the scheduling was to schedule the most important ones towards the end, just so we got better practice with the demonstrations as well. And it just so happened that's how the schedule works. So we gradually escalated in terms of size towards the -- so which is why our press release earlier in July probably showed a lower installed capacity, and we had more towards the end.

Robert Brooks

Got it. That's helpful. So 12 -- so it was like 12 different customers across the 12 visits?

Abinand Rangesh

Yes. So exactly 12 different entities.

Robert Brooks

Got it. And then when we spoke after the last print, you mentioned how the larger data center operators were sort of split in two, both interested, but group one, let's say, would learn of the capacity limits you currently have and sort of lose interest and say, call me when you can do more. But then interestingly, you had mentioned that, that second group were some who were seemingly willing to scale up with you. So I wanted to follow up there and hear, is today's commentary on these larger potential customers a sort of evolution of that second cohort getting more comfortable on the technology? Or just any additional color you could share on that dynamic?

Abinand Rangesh

Yes. Yes. No, that is a great question. I believe once -- again, when people come for demonstrations, there's generally a certain level of interest, right? The way at least I think of the sales process, you've got two portions to it. The first part is getting the interest up to the point that somebody says, this makes sense for me. And then the second part of it is, of course, reducing the -- anything that might add friction to it or might act as a negative to them not closing the deal, right? And out of that capacity is one of them. That's part of why we're starting to build some inventory, so we can cut the lead time and we can turn these things around without having to tell somebody, you need to give us an order right away, and this is our lead time, right? We can start getting a head start on some of that.

Having said that, I believe based on all of the discussions that we had during these meetings. Again, the advantage of having, especially the ones in person is you get a decent amount of time to really probe their requirements and what they're looking for and where they're thinking. We're starting to find, I think, applications that we could essentially start small and grow with them. That is also one of the advantages of these bigger data centers that have existing locations because you could either add almost as a retrofit as a pilot or you can do it as part of one phase of a bigger project. So there's ways to do this without having to have all that capacity upfront. And we believe that this is -- the capacity issue, although is important, I think there's a way around it in this case, but we will see as further discussions go forward.

Robert Brooks

That's really helpful color. And then maybe just on building that, the building capacity ahead of orders as the confidence has grown. I was just curious to get a sense on the size of that and how much that might cost?

Abinand Rangesh

I'm not sure I can comment on that exact number, but I'll give you kind of how we think about it. What we're thinking about is what would people want, let's say, in the very end of this year towards the early part of next year, right? What's sort of a rough number that -- and then if we can -- if we, let's say, have three or four different potential customers saying, I need roughly this amount of capacity, then we'll say, okay, our odds -- if we build x amount, then we'll sell it to at least 1 out of the 3 or 1 out of the 4, right? We -- our odds are good in terms of getting that sold.

So that's how we're thinking about it in terms of risk, cash and in terms of also timing because I think part of the issue I'm sure everybody is thinking about is what does this mean in terms of cash flow and inventory. And if we time it right, in theory, we should be able to bridge -- we won't necessarily need to put up all the capital upfront, right? It's -- if you look at the timing of when these things could potentially be shipped or delivered, that's what we're hoping for. But even if it ends up slipping a little bit, that -- it allows us to manage the risk versus the cash flow.

Operator

Our next question comes from the line of Eric Stine with Craig-Hallum.

Eric Stine

So maybe first, talking about those 12 demos. Just curious, is there a way to think about that, which opportunities you may pursue under the Vertiv relationship or not? Or is that even the right way to think about this? I mean, is that kind of a separate path?

Abinand Rangesh

I'd say the two are parallel paths. Pretty much all of this -- all the demos that came were arranged through us and our sales team and our marketing efforts. We, of course, if it makes sense in these projects, there are places where it may make sense to collaborate. But right now, those are all -- came from our marketing efforts.

Eric Stine

Got it. And then maybe sticking with Vertiv. I know you got the initial order. Just curious if you can speak to both the progress towards finalizing the MPA? And then also, I know you've talked about in the past that Vertiv was specking you into, I think, 25 to 50 megawatts of projects. Maybe if you could just give an update on where those specific projects stand?

Abinand Rangesh

Yes. So one thing on this particular call, we -- as we mentioned anything, of course, on either party, right, we have to get prior approval on both sides. In this particular case, I talked to my counter-parties, and we decided that let something close so that we save the approval process with Vertiv for something that's more substantial. So at this point, I can't comment too much on that without -- so I think stay tuned, let things happen, and we'll announce deals as they happen.

Eric Stine

Okay. But beyond the deals, in terms of just progress towards the MPA, I mean, is that just still in process?

Abinand Rangesh

Yes. So I think things are in a very good place overall. But I just cannot comment at this point because, again, I didn't this time around, get specific [ permission ] for it.

Eric Stine

Yes. Understood. Okay. Last one. So building inventory, I know you've built some inventory over the last 3 quarters. When you're highlighting specifically building inventory in your prepared remarks and now, is that saying that there is more inventory build to come? And then how do you kind of manage that versus some of your capital constraints that you've talked about?

Abinand Rangesh

Yes. So we had a little bit of inventory earlier. But as our backlog has increased for the non-data center projects, a lot of that inventory is going to get absorbed as part of that. So we have to add additional. It's also, as I mentioned, it's -- we're trying to manage this in such a way that it's really -- it's less a matter of total inventory, it's a matter of figuring out how do you compress the lead time because typically, with some of these projects, right, there's a lot of things moving in the background and you may get various progress and verbal approvals. But as you may not get a purchase order for a while even though we might know that things are progressing, things are happening.

So we don't want to have a delay at that point. We'd rather know that, okay, we're going to get -- we feel very confident we're going to get these projects, in which case, let's get -- so as soon as we get a purchase order and get a deposit, we can start shipping. And that's how we're thinking about it. So we're trying to balance the cash flow versus that, but we also feel that it's important to be able to move quickly, especially because at this point, we have the interest of these bigger names. We don't -- we want to make sure that we're able to respond quickly. And then hopefully, once we get some units starting to ship, then we can start ramping up very quickly from there.

Operator

Our next question comes from the line of Chip Moore with ROTH MKM.

Alfred Moore

I wanted to follow up maybe around the demos and some of your comments, which makes sense around balancing working capital and compressing those lead times. Just the discussions you've had, I think you alluded to even talking about dates. Maybe just help frame that out like a range of whether it's sort of those pilot early phases versus larger potential opportunities? Just any sense of where those type of dates are panning out.

Abinand Rangesh

So part of that is it's -- like I can't comment too much on that. What we are trying to do is to have really work with these end customers to figure out how we can scale with them. So we think at this point in the company's like where we are, it's probably more important to get the right names that as -- in terms of first projects, if we can do that just because if you get the right names, you end up shaping the whole -- all the future development, right? Once you have the right brand name, everybody else is going to follow, hopefully, based on that -- the right name using it because it's, I think, reputationally probably more important to get that done right than necessarily -- I mean, whichever order we get, of course, if we get a smaller data center first, we're going to take that. But our priority right now is to try to get the right brand names closed, if we can.

Alfred Moore

Understood. That's helpful, and we look forward to hearing updates. And maybe just for my follow-up, the base business, right, with that backlog increasing, good to see, and it looks like you expect some more here. Just the trends you're seeing there and any changes in sort of base markets?

Abinand Rangesh

So we're starting to see power constraints across the board all over the place, which is -- in the past, a lot of our sale, whether it was chillers or cogeneration was made predominantly based on economics. Now we're starting to see lack of power, electrical equipment being -- having long lead times, those kind of things affecting the ability for -- non-data center type customers have access to either cooling or power generation or any -- all of that equipment. So we're starting to see that driving some of this.

We're also starting to see a little bit of the cogeneration and standard power generation side of things start to come back for a while. That had significantly reduced, but we're starting to see some of the larger multifamily buildings and other types of buildings that are seeing high utility rates that -- and in some cases, not enough power that are looking to use our equipment. So I think the base business is growing just as a result of some of what the data center space is also facing.

And part of what we've seen in that base business, right, things like switchgear, things like that, that are having longer lead times is also what is making us feel like having some of that inventory on hand will also likely pick up potential customers that are trying to get other electrical equipment, they're not able to install electrical-based equipment just because the switchgear and things like that might be longer that we could essentially pick up projects just that way as well.

Operator

Our next question comes from the line of Alex Blanton with Clear Harbor Asset Management.

Alexander Blanton

I've got a question about the direction you see the market going for you, is it going to be original data centers built from scratch? Or is it going to be retrofit of existing centers? And in those instances, would it be participating in expansions of existing data centers rather than just retrofitting what's already been built?

Abinand Rangesh

That's a great question. Actually, we see both. Its -- earlier, a lot of our opportunity pipeline was with newer type of projects, but we're starting to see more with existing data centers in a few different types of applications. One is the AI side of it gets a lot of the press, right? But there's still a lot of existing cloud data centers that are supporting either the AI infrastructure, but also just regular web workloads. And many of those are running out of power. And their cooling load may not be quite as big as an AI data center, but it's still there. And if they can free up some of that, then there's benefits there. So we're starting to see some potential in those kind of applications.

We're seeing some potential customers say, we can incorporate this as part of a bigger campus where they do it in one phase or as they build it into a certain phase of an expansion. And then there's also like the smaller data centers that might use it as a primary cooling source. So there's different applications that we're starting to see. And then there's also a push, I think, more broadly in the industry to start moving towards maybe smaller data centers that are closer to urban environments just to -- where you can pick up pockets of power. And in those areas, again, our cooling products could be a very, very good fit. But we're seeing a mix of different types of applications right now.

Alexander Blanton

But you see an opportunity to retrofit where someone has an electric chiller and they say, well, let's take that out and install Tecogen instead. Is there any -- what is the opportunity there?

Abinand Rangesh

That actually is -- we are seeing some potential opportunities like that. But what is more likely to happen in those kind of applications is because with an existing data center, the actual -- doing the retrofit tends to be because it's a live site. So they might add it to either as [indiscernible] almost like an expansion into -- because many of these data centers typically have some additional space they set up for future expansion, where they may not necessarily remove the electric chiller, they might leave it in place and add us in and put the electric chiller now as a backup. So that's a more likely application than a pure remove electric chillers. But we are seeing some people that are looking to remove electric and move put ours in.

Alexander Blanton

I'm interested in what you said earlier about solving the -- helping to solve the problems of pollution and noise. And there was a third one.

Abinand Rangesh

The water usage.

Alexander Blanton

Water pollution and noise. How do you do that [indiscernible] in your case?

Abinand Rangesh

So if you look at -- yes, so there are two different ways you can do it, right? One is with the chiller because typically, when a utility wants you to shut down your -- or reduce power from the grid, it's usually on the hottest days just because that's when everybody else has their air conditioning system on. And when the utility starts to get very constrained, they start asking large load users to drop load.

In which case, your alternative and what happened in Virginia, for example, was the diesel generators were turned on. What we would do in our case, right, is just have that load either dynamically move to natural gas or have those those chillers just be turned on in those applications so that load moves over to natural gas. Because our equipment is designed for continuous usage and originally, like our Ultera Emission System was designed for Southern California. So we have very, very clean emissions. The -- our Carbon Monoxide NOx very, very low, right? It's [indiscernible] so in that sense, substantially cleaner than diesel generator. The other thing is because our machines are being designed for continuous usage, whether it's the InVerde product or the Chiller, designed for pretty low noise urban environment. So you could be standing next to it and you hear it, but it's not going to be bothersome, no more than ambient noise. So in that sense, it's substantially quieter than those diesel generators. So that's the -- the first is the chiller application.

The second application is also potentially the InVerde. It's not necessarily going to run all your loads, but there are certain standby loads or are other loads, again, that you may not want to have your massive full data center load coming on smaller power systems like InVerde. But because it's modular, you essentially could add the InVerde in conjunction with the Chillers to shed some of that load during that peak time and not turn on the generator and keep the diesel generator really just for emergency use.

Operator

And we have reached the end of the question-and-answer session. I would like to turn the floor back to Abinand Rangesh for closing remarks.

Abinand Rangesh

Thank you very much, everyone, for attending our Q2 2026 conference call. I will keep everyone updated as things move forward. I believe we've gotten the interest. We're going to do everything in our power now to turn this interest into projects, but I'll keep people posted. And hopefully, we'll update -- we'll have some interesting news over the next few months. Thank you.

Operator

Thank you. And this concludes today's conference, and you may disconnect your lines at this time. We thank you for your participation.