Innventure, Inc.

Good day, and thank you for standing by. Welcome to the InVenture second quarter 2025 earnings conference call. At this time, all participants are in listen-only mode. After the speaker's presentation, we'll open up for questions. To ask a question during a session, you need to press star 1-1 on your telephone. You then hit an unreaded message advising your hand is raised. To withdraw your question, please press star 101 again. Please advise that today's call is being recorded. I would now like to hand it over to your first speaker today, Lucas Harper, Chief Investment Officer. Please go ahead.

Thanks, Operator, and thank you all for joining us for InVenture's second quarter 2025 earnings call. My name is Lucas Harper, InVenture's Chief Investment Officer, and joining me from the company are Bill Haskell, Chief Executive Officer, and Dave Yablonowski, Chief Financial Officer. Earlier today, we issued a press release announcing our financial results, which is available on our investor relations website, along with a supplemental slide presentation. As referenced on slide five, we will be discussing non-GAAP financial measures during this call. The most directly comparable GAAP financial measures and the reconciliation of the differences between the GAAP and non-GAAP financial measures are available in our earnings release and supplemental slide presentation on our website. In addition, certain statements being made today are forward-looking statements that are based on management's current assumptions, beliefs, and expectations concerning future events impacting the company. These forward-looking statements involve a number of uncertainties and risks, including but not limited to those described in our earnings release Form 10Q for the period ending June 30, 2025, and other filings with the SEC. The actual results of operations and financial condition of the company could differ materially from those expressed or implied in our forward-looking statements. And now I'd like to hand the call over to Bill.

Thanks, Lucas, and thanks to everyone joining us today. Last quarter, we invited Dino Fodoraro, Excelsior's Chief Revenue Officer, to call for a deep dive into Excelsior's technology differentiators, market opportunity, and customer engagement. We heard great feedback from investors and analysts on the value of this spotlight, so today we are pleased to have Josh Kleiman, CEO of Excelsius, on with us to speak about the exciting recent developments at the company and the incredible opportunity that lies ahead. Before passing to Josh, let me just reiterate that market momentum for Excelsius continues to accelerate, and we anticipate an inflection in commercial bookings before the end of 2025. Josh's commentary will provide real-world context for why we remain confident in our ability to deliver on that goal. With that, I'd like to turn the call over to Josh. Josh?

Thanks, Bill. I'm excited to be here today. As Bill mentioned, Excelsys continues to make tangible progress and gain traction in the market. I'm incredibly proud of my entire team for what they have accomplished to date, and we are very excited for the future. You've heard InVenture speak in the past about the size and rapid growth of the data center liquid cooling market, which is only accelerating as the adoption of AI continues to proliferate the market. In a recent analysis of the market by UBS, 100% of the roughly 128 IT organizations surveyed intend to deploy AI workloads, with 14% having deployed AI workloads in some capacity already. The power demands these workloads placed on current data center infrastructure are simply too high for traditional air cooling. And consensus is building in the industry that single-phase water solutions will also rapidly become insufficient for many compute-intensive applications. It is important to note that the average order size for these workloads is still less than $5 million, which demonstrates we are still in the proof-of-concept stage. The bottom line is that the opportunity is still ahead of us, and the direction of the market is very clearly pointed in the need for advanced cooling solutions. With a differentiated market-leading technology, Excelsis has positioned itself at the forefront of what we believe is an inevitable two-phase direct-to-chip liquid cooling adoption cycle. Against this attractive backdrop, Excelsis has been making great strides. Before getting into the details, it is important to remind you all that ExcelSense was founded in 2022 and has only been in the market for less than one year. The third quarter of 2024 was our first quarter of revenue generation. And to be where we are today, less than 10 months later, is truly impressive. So let's dive into several key milestones. As we've stated previously, Excelsis is focused on building relationships with four key groups, hyperscalers and multinational OEMs, global resellers, colocation providers, and AI as a service or neocloud providers. We continue to make strides with each of these constituents. As of the second quarter, we engaged with our first hyperscaler for a proof of concept system. The system is currently being manufactured, and we expect to deliver it by the end of Q3. From there, it will go through the evaluation process, which we expect to take six months. We are also in discussions with two other hyperscalers at the moment and are confident in our ability to advance to proof of concept with them as well. We began taking our first orders from our OEM partner, one of the largest in the data center sector. We expect this order volume to begin to ramp during the third quarter and continue to ramp with demand throughout 2025. We also announced an exciting engagement with Equinix, which is just the latest deployment among several other key co-location players. As with many co-location operators, Equinix is increasingly required to become a thought leader in cooling and power architectures. We have grown our contracted go-to-market partner network to 21, an increase of over 100% since the beginning of 2025. This network includes industry players such as Park Place, Avnet, Computer Center, and more. Finally, we plan to open a demonstration point in the Bay Area with our contract manufacturing partner before the end of the year. Our as yet undisclosed CMODM partner is one of the key suppliers to multiple hyperscalers. and is capable of ramping this demand as proof-of-concepts convert to production opportunities. As you can see, Excelsis has engaged with dozens of key players across the data center ecosystem, and we have quickly built an impressive reputation, which is driven by our differentiated technology and our data-driven approach. Yale industry players are witnessing the technology platform in the field, and results have been undeniable. Bayfield Engineering, a leading data center-focused MVP firm, just completed a study comparing our technology side by side with single-phase water and found an operating cost improvement over water of 28% to 40%, depending on the exact assumption set around infrastructure deployed. Our performance has been impressive. However, the technology is still new, and cell cycles for these relationships can sometimes span several quarters. Because of the compelling nature of the Excelsior's technology, we anticipate the sales cycle will shorten significantly as the technology gains traction in the field at scale. Early signs of this traction are highly encouraging, evidenced by a continued upward trend in leading indicators of demand, namely qualified marketing leads. When we spoke to you during the first quarter, marketing leads had exploded close to four times since the start of the year. While it's still too early to give specific revenue expectations, our pipeline continues to grow, with three times gross coverage of bookings as compared to the plan for the remainder of 2025, and over six times coverage of our bookings plan for 2026. We are also excited to start to see a transition of our pipeline from 100% proof-of-concept engagements to date to 75% production engagements through 2026. And I'll make one final point on the pipeline to frame the rapid pace of momentum we are experiencing. Our Q2 bookings were greater than all bookings going to market before that. And our best estimate for Q3 is that bookings will be a multiple of the Q2 level. You can see that the sales momentum is real. And I'll close with information regarding an exciting addition we made to our team to institutionalize our sales process going forward. We are exceptionally proud of the team we have brought together. The latest illustration of this is our recent hire of Eric Brown as Vice President of Global Cells. Eric is a seasoned leader who joined us after over 23 years at Dell and seven years at AMD. I worked with Eric during my time at Dell, and we couldn't be more excited to have him on board. Eric brings decades of experience managing multi-billion dollar cells engines that will prove invaluable as Excelsys continues its exciting trajectory. Thanks again for having me on the call today, and I'll pass it back to Bill.

Thanks, Josh. As we said in previous quarters, the momentum at Excelsior is real, and the excitement around the opportunity is palpable. We believe the company is equipped with a market-leading technology that positions it at the forefront of a seismic shift to two-phase directed-shift liquid cooling adoption. We also feel strongly that Excelsior's ability to attract top talent, like Eric Brown, is a testament to the work they are doing and how well established the company has become in such a short period of time. Inventor is proud of the progress the company is making with major players in the data center ecosystem, and we believe the value of Excelsius and Inventor's operating model as a whole are markedly underappreciated by the street. Given the breadth of the opportunity and the level of excitement at Excelsius, it's easy for the market to overlook what Inventor really is, which you believe would be a mistake. So before passing the call to Dave, I'd like to reiterate that InVenture is foundationally an engine to create market-changing companies like Excelsius. In the context of our conglomerate model, the success we've had to date is real and the value creation is compelling. While we focus today's call on Excelsius, which we believe represents the most immediate value creation opportunity, InVenture has started three other companies that are doing incredibly exciting things in their own right, and offer proof points of our ability to deliver value for shareholders. First is PureCycle, which Inventure founded from scratch in 2015 and took public in 2021. Today, PureCycle trades at a market cap of greater than $2.3 billion. While Inventure no longer holds an economic interest in the company, it offers a tangible proof point about the value creation opportunity offered by our model. Second is Aeroflex, founded in February 2018 in collaboration with Procter & Gamble, which is gaining significant traction in the $400 billion packaging market. As of the end of the second quarter, the company achieved four consecutive quarters of revenue recognition, outlining its growing customer base and unique ability to serve the sustainable packaging market. At the end of May, the company received critical guidance recognition for recyclability from the Association of Plastic Recyclers, or APR, for its proprietary monomaterial Aeroflex pack. This recognition is one of the most widely accepted benchmarks for evaluating plastic packaging and provides independent validation and credibility that the Aeroflex pack is designed for recyclability. This is a significant milestone and we believe a precursor to unlocking meaningful sales opportunities. We've successfully re-engaged with a handful of notable CPG companies that require APR approval and we are experiencing increased engagement, and a maturely larger revenue opportunity. The company also took steps in the second quarter to further strengthen its IP position and competitive advantage by adding new patents to its growing portfolio. Since the beginning of June, the company was awarded two patents that present a flexible package, plastic package suitable for curbside recycling where all plastic bottles are accepted. The patents align with industry standards used to assess the compatibility of polyethylene-based films and flexible packaging innovations. This achievement demonstrates Aeroflux's unwavering commitment to providing a sustainable and differentiated liquid packaging solution to the marketplace. Third is Raffinity, which we launched in December of 2024 in collaboration with Dow. During our earnings call earlier this year, we outlined three strategic initiatives for Raffinity in 2025. and we are happy to report the company is executing against each of these three milestones. The first was engaging an engineering procurement and construction partner for the first plant design, which was completed during the second quarter. Second was to demonstrate viability of fluid bed conversion of mixed plastic waste to liquid olefins at pilot scale within VTT's plant, which is on track with preliminary results leading to better quality and higher yield than originally modeled. Finally, third, was securing initial site selection and feedstock sourcing for the first plant, which we are on track to accomplish during 2025. As you can see, InVenture's operating model has produced tangible and compelling value creation opportunities over the last decade, and we believe we are just getting started. We have a robust pipeline of high-quality technology opportunities with several MNCs that could represent future InVenture operating companies. We'd remind everyone that we only launch companies that we believe can achieve at least a $1 billion in enterprise value. The exciting progress across our current family of companies is real and observable evidence of InVenture's value creation engine at work, and we believe the future is extremely bright. With that, I'll turn the call over to Dave to cover our financials. Dave?

Thanks, Bill, and good afternoon, everyone. InVenture's second quarter revenue was $0.5 million. representing 0.3 million of revenue at a Celsius and 0.2 million of fees we collected from our management of the Inventus ESG fund. As with the prior quarter, the second quarter result was in line with our expectations and consistent with our view that we expect most sales orders, bookings, and early stage revenue to be weighted to the second half of the year. As Josh highlighted in his remarks, the size and rapid growth of the data center liquid cooling market is accelerating as the adoption of AI continues to proliferate the market. The company has positioned itself at the forefront of what we believe is an inevitable two-phase direct-to-chip liquid cooling adoption cycle with the Celsius's differentiated market-leading technology. G&A expenses were approximately $18.6 million for the quarter, an improvement from $19.7 million in the prior quarter. This number has four primary components. non-cash equity-based comp of approximately $9.4 million, accounting, audit, and legal professional service fees of approximately $4.4 million, $3.9 million for payroll expense, benefits, and other operating expenses, primarily driven by growth at Ascelsius, and $0.9 million of non-cash amortization related to our intangible assets and equipment depreciation expense. It's important to point out we continue to use outside professional services to keep our long-term fixed costs and headcount low. We do, however, expect to bring in-house a large portion of this activity at an expected lower cost as we build our internal processes and capabilities that meet our public company compliance requirements. Moving down the income statement, we booked $113 million non-cash goodwill adjustment driven by the write down of goodwill on the balance sheet. This was a result of a variety of macro factors that affected business investments in the quarter. But again, this was non-cash. In the non-operating section of the income statement, we booked a favorable non-cash adjustment of approximately 7.2 million related to the change in fair value of our warrant and earn out liabilities. EBITDA for the quarter was a loss of approximately 135 million After adjusting for certain items, adjusted EBITDA was a loss of 16.2 million, an improvement from a loss of 21.8 million in the first quarter. Moving to the balance sheet. As previously disclosed, in the second quarter, pursuant to our securities purchase agreement with Yorkville Advisors, in two separate tranches, we issued convertible debentures for an aggregate amount of $30 million. In connection with these issuances, we received net proceeds of $18 million on April 14th and $9 million on May 15th. This added $27 million of cash to our balance sheet. Additionally, $6.5 million of related party and other term convertible notes was raised into the second quarter. These amounts are reflected in the increase in financing activities on our cash flow statement. These two actions illustrate our active management of cash and both improve our capital position. Other notable items on our balance sheet. We ended the quarter with $12 million of cash, restricted cash, and cash equivalents, up from 11.1 million at the end of last year. Inventory. As we take customer deposits and book new customer orders at a Celsius, inventory is beginning to grow. The second quarter ended with $6.6 million of inventory on hand, up from $5.2 million in the prior quarter. We're acting sensibly with our cash, optimizing the need to keep inventory on hand to meet customer orders while preserving cash on our balance sheet. Investments. The $32.4 million on our balance sheet represents our $22.9 million equity method investment in Aeroflex, a $9 million investment in debt securities in Aeroflex, and our half a million equity investment in the ESG fund. Intangible assets. This represents developed technology and other intangible assets at Ascelsior. These assets amortize at different intervals primarily over the next 8 to 15 years. On to the cash flow statement. In the cash from operating activity section, you can see many of the non-cash items I mentioned earlier. The cash used in investing activities primarily represents the funding given to Aeroflex through our existing facility comprised of debt securities. The cash provided by financing activity section includes the $27 million of cash raised through the issuance of the Yorkville convertible debentures, 6.5 million of related party and other term convertible notes. It's important to say again, the InVenture team remains focused on additional capital raises at the InVenture and operating company level to meet our liquidity needs and fund what we believe are very attractive growth opportunities ahead of us with the goal of creating substantial value for our shareholders. With that, we'll open up the call for questions.

Thank you. And as a reminder, to ask a question, you will need to press star 1-1 on your telephone and wait for a name to be announced. To withdraw your question, please press star 1-1 again. Please stand by while we compile the Q&A roster. One moment for our first question. Our first question will come from the line of Chip Moore from Roth Capital Partners. Your line is open.

Hey, everybody. Taking the question. I wanted to ask, you know, your commentary, I guess, in the press release, you know, on looking for an inflection in revenue growth in the second half here. It seems like your visibility is growing there. I assume that's a function of traction at Excelsius, but maybe a little more color there. I think you talked about you know, first orders with the OEM may be starting to ramp up. But any more color on your thoughts there? Hey, Jeff. It's Bill.

Good afternoon. Nice to touch up with you. So there has been activity, and I'll let Josh amplify this in a moment. There certainly has been activity that's material. There's one thing that I would want to share, which is There can often be a long lag between order and delivery, not because of manufacturing issues on our end, but because these systems are going into very complex environments and oftentimes customers will place orders in some cases two, three quarters in advance of when they actually want delivery. So we're focusing on the bookings because I think that really will be the indicator of the growth in the business. But I'm going to flip it over to Josh now, and if he wants to amplify anything, he's welcome to.

Thanks, Phil. Yeah, I think we're seeing a couple things happen. One is that the bookings are coming in faster than they have, so we're seeing sort of an inflection point in that. I would like to remind everyone we've been in market for about three and a half quarters, so we're new to sort of shipping products and selling product in market. And so I think the adoption cycle's happening about as quickly as we would expect it to. The other thing that I think you'll see a tick up in our bookings through the rest of 25 and into 26 is because the average order size is increasing as well. And this reflects a shift from what were purely really proof of concept orders for the first couple quarters to production orders which are kind of by their nature larger um so i think we're seeing all the signs we would like to see at this point in terms of adoption of two-phase directed chip cooling and all the signs that would evidence um gaining traction on uh for excelsis in the market yeah that makes great sense appreciate the color josh and i guess

To your point, I think you talked about, if I heard you correctly, you know, 75% production engagements next year, right? So it sounds like you've got visibility on some of those growing order sizes. Is that fair?

That's correct. And a big portion of our pipeline for 2026, and we expected and I guess hoped that this would occur are coming off of proof-of-concept orders that we had this year. And I think that will continue for the next two to three quarters, that we'll deploy a proof-of-concept order because people want to see the technology work in their environment. But they're doing that with intent. They're doing that because there is a larger production rollout happening after that. And I think we're starting to see that pattern kind of coalesce. which we find encouraging.

You might talk, Josh, also about the selectivity, I'll just say around the selectivity of the POCs that we have delivered that have, you know, some sort of a multiplier.

Yeah, I mean, I was going to, I'm sorry, you go ahead.

No, go ahead, sorry, Josh.

Yeah, I was going to say just to Bill's point, To protect our bandwidth at Excelsius, really, we've been fairly particular about where we deploy equipment because we want to make sure that when we deploy it, it does have a multiplier effect. So when we deploy a proof of concept, it's really not for someone just to play around with the equipment and become familiar. It's really... because they want to test the equipment, because if it does what we say it's going to do, there's a project in front of it, and that is the sort of overwhelming pattern that we're seeing as we go to market.

Very helpful, and maybe just a follow-up on that point, you know, your first hyperscaler proof of concept, that's obviously a big milestone, and I understand it'll be small initially, and they'll need to trial it for a period, but can you give us a sense of I mean, we all know, you know, hyperscalers are large, but what would they, you know, be looking to source for multiple areas or any indication of, you know, how big something like that could be if you're successful?

I'll just say, as you would expect, it would be big. So, we have been given a phasing plan based on that proof of concept. that if it's successful, and we've been given some timeframes, if it's successful, it'll lead to X amount of CDUs and sockets being cooled. I'd rather not get into the exact detail of that. But I think the way this should be sort of interpreted is that the most sophisticated players in the market, and most of the hyperscalers have very sophisticated thermal architecture teams, in their businesses are starting to get very curious and starting to want to play with two-phase technology, two-phase directed chip, because they know it's going to come. And so we're seeing not just this initial hyperscaler, but we're seeing others follow suit, and we expect a couple other placements in at least one or two other hyperscalers before the end of the year. We're really excited about this, and I think that should be the interpretation. What we're seeing in the market generally is that the most sophisticated players are coming to us in a way because they're the ones that are kind of looking around the corner and saying, well, how long can we use single-phase water or these other cooling architectures with these increasing wattages, these TDPs of chips in the next generation and the following generation? And they're seeing that they're probably going to require two-phase directed chip going forward. So I think these are the early signs of that.

Excellent. And the scale issue at a macro level, you know, the four top hyperscalers have published that they're spending about $350 billion this year on data centers and data center architecture and equipment. and which is, you know, roughly a billion dollars a day, and they've projected 400 billion for next year. So they're spending significant amounts of money in anticipation of the growing demand in the whole, you know, generative AI space.

For sure.

It seems like those numbers go up every quarter. And any more on the manufacturing side, you know, your comfort there, given, you know, what could come down the pipe, you know, on the contract manufacturing side? I think you talked about opening something up in the Bay Area. Just any more color there?

Yeah, sure. Sorry. Yeah, so our manufacturing strategy, which we've really been – talking about and setting up from the very beginning, even before we had a product in the market, has been that we will have a core of manufacturing internally. And we've just signed a lease. In fact, we've just taken hold of an expanded facility here in Austin, about 30,000 square feet, about 27,000 of that is manufacturing. And that will service us. That could actually service the entire planned for next year. But we will subsidize that with contract manufacturing because as Bill mentioned, I think earlier, some of these engagements are sort of binary. If we get traction in some of these engagements, they're going to be massively scaled. And for that, we will rely at least partially on contract manufacturing. But we will always keep for every product a core of manufacturing in-house because we think One, we don't want anyone to have too much leverage over our business. And two, it's just a good quality way to scale manufacturing, to actually do our own analysis of tack times, do our own analysis of refining our build materials for manufacturing, designing for manufacturing. So we just look at best-in-class examples of that, and we think that's a good path forward. So imagine a couple of years from now, we'll probably have 25, 30% of our manufacturing in-house and the remainder with contract manufacturers.

I mean, if I could speak one last one, and maybe for Dave, just on the balance sheet, right, just your comfort, you know, sort of bridge this period here, you know, before some of the commercial traction really accelerates. Thanks. You say bridge the period you've been from the market perspective? Yeah, or just cash needs, right? How you're thinking about access to capital and things like that.

Yeah, I have a high degree of comfort in our access to capital and our ability to put cash on the balance sheet. I'm not concerned in the least. Perfect.

All right. Thanks very much. Thank you. As a reminder, that's star 11 for questions. Star 11. Our next question comes from the line of Ashi Alpesh from Siddoti. Your line is open.

Congratulations on the quarter, guys. I just have one more question relating to a Celsius. And I'm trying to understand the product. Can you explain how accurately does the thermal simulation rack replicate real-world AI and HPC workloads? And if you can just help us with understanding what is the feedback you're getting from the initial proof-of-concept deployments that you've done in the last quarter, and what's the market response been like?

Thanks, Ashi. I'm going to flip this one to Josh, and I think if you can maybe talk about the kind of the product suite that we have and the simulated racks that can simulate things that don't exist yet and the implications of that would be helpful.

Sure. Absolutely. Thanks for the question. So there's an issue with testing thermal loads with GPU servers, and that's primarily because GPU servers are extremely expensive. So a rack of IT used to, maybe three, four years ago in ancient history, used to maybe cost $200,000 worth of servers to fill a rack within a data center. But for GPU servers, it might cost $3 to $5 million. So if you want to test a thermal load, if you want to make sure that the cooling architecture's cooling solution can cool that power density, without spending millions of dollars, you have to rely on some sort of load simulation device. And so what we've done is create what we call TTVs or thermal test vehicles, which look like a processor, essentially same form factor, same sizing as a processor. We put a few of these. In fact, in our latest load simulation sled, we have six of them. each of which can drive about 1,000 watts, so a kilowatt each. So each load sled, which looks like a one-use server, can produce the heat of a six-kilowatt server. So you can fill a rack with those and test the solution, the cooling solution, in what looks like, over what looks like a sort of a GPU reference design for that rack at a fraction of the cost. By a fraction, I mean maybe 1 20th of the cost. So we think this is going to be useful. Now we created, we developed these load simulation sleds and racks for our own product development purposes, for our own R&D. And then we productized them because we found there's a demand in the market for them. So this is going to be, I don't know how large, it'll be a supplemental business for us for commissioning data centers, for experimenting with different thermal loads in the real world. But we are shipping these today and they have been well received. In terms of the proof of concepts that we have in the market, There is a desire to sift through some of the noise out there. Because it's a nascent market sector, there are a lot of companies that claim they can do things. They claim they can handle certain thermal capacities that they can't do. And I think it's the nature of every new market. So we've taken the approach that we want to be very data-driven. We encourage people through our Kickstarter program to buy one of our racks, to fill it sometimes with these load simulation sleds, sometimes with real servers, and test it and make sure that our product does what we say it's going to do. And typically, the proof of concept is simply that. We like what you say your solution can do. We want to see it actually work and perform the way you say it's going to perform in the real world. And as soon as it does that, we'll be ready to buy more. So the initial feedback on these proof of concepts has been extremely positive thus far.

And it also allows them, Ashi, to look forward and project chips that are yet to be delivered in the marketplace, project how our technology will cool those chips, you know, a generation or two forward. So they can, you know, normally test chips that haven't been

around the market today that are being delivered. Great. Thank you so much. Thank you.

Thank you. Once again, that's star 11 for questions. 111 for any questions to populate. And I'm currently not showing any further questions at this time. This will conclude the question and answer session for today. Thank you for your participation in today's conference. This concludes the program. You may now disconnect. Everyone have a great day.