Duos Technologies Group, Inc. (DUOT) Earnings Call Transcript & Summary
March 31, 2026
What were the key takeaways from Duos Technologies Group, Inc.'s March 31, 2026 earnings call?
In the fourth quarter and full year of 2025, Duos Technologies Group, Inc. (DUOT:US) reported total consolidated revenue of approximately $27 million, a significant increase of over 270% year-over-year from $7.3 million in 2024. Despite falling short of the previously projected $28 million, management highlighted strong growth driven by services from the asset management agreement with APR Energy. The company has pivoted towards a data center-focused strategy, with guidance for 2026 set at $50 million to $55 million in total revenue, reflecting confidence in new business initiatives and a robust sales pipeline.
What topics did Duos Technologies Group, Inc. cover?
- Revenue Growth Acceleration: Duos reported a total revenue of approximately $27 million for 2025, up from $7.3 million in 2024, representing a 270% increase year-over-year. Management noted, "The company delivered materially stronger gross margin in 2025, generating $7.9 million in gross profit, achieving approximately 29%."
- Divestiture of Rail Division: Management announced the decision to completely divest the Rail division due to "lack of growth and regulatory hurdles". This move is expected to free up resources and reduce SG&A expenses significantly.
- Launch of Duos Technology Solutions: The newly formed Duos Technology Solutions division has already secured $10 million in backlog, which is expected to be recorded as revenue in 2026. CEO Doug Recker stated, "This new line of business has low overhead and is simple to execute while having strong commitments by the end client."
- GPU-as-a-Service Contract: Duos secured a GPU-as-a-Service contract expected to generate approximately $176 million in revenue over a 36-month term, with margins exceeding 80%. This contract marks a significant commercial inflection for Duos.
- 2026 Revenue Guidance: Management provided guidance for 2026, projecting revenue between $50 million to $55 million, with a significant portion expected to be recognized in the second half of the year. Leah Brown, CFO, stated, "Our investment and expanded revenue opportunities give us confidence in our ability to execute and continue building a stronger, more profitable company."
What were Duos Technologies Group, Inc.'s March 31, 2026 results?
- Total Revenue: $27 million (vs $28 million est, +270% YoY)
- Gross Profit: $7.9 million (Gross margin of approximately 29%)
- Net Loss: $9.8 million (Improved from $10.8 million in 2024)
- Revenue Guidance 2026: $50 million to $55 million (Reflects growth from core operations and new initiatives)
- Backlog from Tech Solutions: $10 million (Expected to be recorded as revenue in 2026)
- Adjusted EBITDA: Positive (Achieved for the second consecutive quarter)
Duos Technologies is transitioning successfully towards a data center-focused strategy, with strong revenue growth and promising new business lines. The divestiture of the Rail division and the launch of Duos Technology Solutions are expected to enhance operational efficiency and profitability. Investors should monitor the execution of the GPU-as-a-Service contracts and the realization of revenue from the backlog in 2026 as key catalysts for future performance.
Earnings Call Speaker Segments
Hello, everyone, and thanks for being with us today. We released our earnings press release along with our 10-K for 2025 earlier today. You can find copies in the Investor Relations area of our website. I urge everyone listening to go through the press releases and our 10-K filing so you can better grasp some of the specifics we'll cover on this afternoon's call. Before getting started, I'd like to spend a moment personally thanking Chuck Ferry for the leadership and guidance he has given. Chuck has served the Duos organization and offered me personal mentorship. I'm grateful for Chuck and for the opportunity he extended to me at Duos. Getting guidance from someone who's both a war hero and a corporate Champion is a rare privilege, and I'll always be thankful for it. I'm eager for your continued mentorship and advice as you remain a member of our Board of Directors. Chuck, thank you for all you've done and keep doing for the Duos organization. Speaking now as your newly appointed CEO, it's an honor and a thrill to share where Duos Technologies Group is putting its focus; via our Duos Edge and Tech Solutions division, our full commitment now lies with the data center market, propelled by customer demand that keeps accelerating. I'll get into that in more detail in a moment, but first I'd like to bring you up to speed on our Rail technology and Duos Energy subsidiaries. To start, I'll talk about the legacy operation — the railcar inspection portal. We've said on earlier calls that this part of our business matters less and less to where Duos is headed. We've also mentioned shifting our strategy toward edge computing as a way to diversify. With that in mind, we've made the call to exit the Rail division entirely, and that move should be wrapped up within the next 60 days. We didn't take this step casually. I'm aware that the rail technology means a great deal to Duos shareholders given its long history — in fact, my own connection goes back well before I came to Duos, and I played a hands-on role in designing and constructing the Edge Data Centers that the portal relies on now. That said, managing this business has been very difficult because of its absence of growth along with regulatory obstacles. By divesting, we'll unlock company resources and reduce SG&A costs substantially. Additional information about a few divestitures will come out shortly. Now, turning to Duos Energy Corporation. For those who recall, Duos signed an asset management agreement last year with new APR Energy, aimed at securing additional contracts for engineering, procurement, construction, and operation of fast power plants. In addition, Duos received a 5% equity stake in APR Energy's parent company. That AMA gave us the interim financial capability to execute and pivot toward our data center strategy. On the Q3 earnings call, we mentioned that the AMA is scheduled to conclude in 2026, and that Duos will retain its 5% equity stake. Now I want to walk through our data center approach and the new business lines we've set up at Duos Technology Solutions. A key piece of how we've built and rolled out data centers so quickly has always centered on reducing costs and keeping our capital expenditures down. Constructing data center infrastructure demands a great deal of capital. Since Duos is a fairly small purchaser when you compare us with the bigger hyperscalers and colocation firms, we required a method to obtain products at lower prices. That's why we launched Duos Technology Solutions. This entirely new division lets us accomplish exactly that, and it also gives us another revenue stream; we began by bringing on an industry veteran who has a demonstrated record of success, someone who grasps our business along with the broader data center marketplace. Kristen Sanderson has come aboard Duos and will act as a Senior Vice President of Duos Technology Solutions. Kristen brings more than 18 years of experience with data center products, extensive knowledge of market distribution, connections to every key supplier partner Duos must engage with, and a deep network of relationships throughout the data center sector; this fresh division enables Duos to source materials for its own construction projects at a far cheaper price than the old approach of buying via traditional distribution. Duos Technology Solutions extends the same strategic sourcing and product distribution offerings to fresh clients, covering large-scale enterprise organizations, hyperscalers, major colocation companies, low-voltage contractors and general contractors throughout the United States. I'm happy to report that in Q1, Duos Tech Solutions has already secured $10 million in new business. That currently sits as backlog, and I expect every dollar of it to be booked as revenue before this year is over. This added line of business comes with very little overhead, is simple to execute, and has strong backing from the end client. The revenue Tech Solutions generates ought to do more than just replace what we lose from the new APR AMA — it should also bring better margins, which gives Duos Technologies Group as a whole another boost toward future profitability and growth. Kristen has assembled a seasoned team possessing the talent and, in a brief 3-month span, an enormous sales pipeline, and we anticipate incredible results from this new venture. Now I'd like to turn our conversation toward the heart of our newly data center focused organization, Duos Edge AI. Demand for edge computing keeps expanding quickly, and I'm happy to report that Duos Edge AI is well positioned to satisfy that demand. The latter half of 2025 turned out to be incredibly busy for Duos Edge. Back in July of last year — 2025 — we completed a $45 million capital raise together with Titan Partners. That money is funding the building and rollout of 15 EDCs, which pushes connectivity and compute capacity out to Tier 3 and Tier 4 markets that don't get served well enough. On top of that, Duos Edge AI picked up a patent for clean room technology that's used when we deploy modular data centers, and that gives us an advantage competitors don't have in this space. When we looked at 2025, the goal we set was to find, construct, and activate 15 edge data centers. It was a really bold aim — honestly, nobody in our industry had really done it before. So we're happy to report today that we got there. For the first six months of 2026, our priority is to keep pushing our sales approach forward so we can bring in new customers across our markets and make full use of the capacity at every EDC. Back in March 2026, we closed a $65 million capital raise in order to roll out roughly 2,300 GPUs-as-a-service, deliver a 4.8-megawatt high-density EDC deployment for a top hyperscaler, and broaden our high-density EDC footprint so we can meet rising power and compute demand spanning AI inference, training, enterprise, and hyperscale AI workloads. Beyond that, 5 new EDCs are now in production, and we intend to add another 20 megawatts of deployed capacity before the year is out. Maintaining stock that our EDCs can roll out for critical situations—that's vital if we're going to keep growing and succeeding in this market. There's a huge amount of buzz and appetite being created by the Duos Edge AI story along with the early victories it has notched. So when fresh requests arrive from the market, holding that inventory means we can move fast. What we're seeing now is that part of this new appetite calls for higher-density power, which is what AI and high-power compute workloads need.
Even though Duos Edge AI remains dedicated to its initial strategy of rolling out in Tier 3 and Tier 4 markets, the appetite for power we're encountering is unlike anything before—measured in megawatts rather than kilowatts. The data center sector is booming in a way we've never witnessed, and constructing at scale carries a hefty price tag while requiring years to finish. Throughout this rollout—our 15 EDCs—we received a surge of inquiries seeking additional power in the regions where we've established organizations nationwide. Data center capacity and power are so scarce that businesses are approaching Duos Edge AI; as a result, we're going to begin constructing our new EDCs with larger power capacity to satisfy that demand. We've demonstrated in the marketplace that we can deploy more cheaply and reach the market with remarkable speed. Duos Edge AI will now serve customers with high densities, such as neocloud providers and hyperscalers, for their remote edge sites. These EDCs with greater power capacity ought to deliver considerably higher monthly recurring revenue for Duos, which we'll detail in our financial update shortly. Before handing things over to the financials, I want to address our start of the year and our inaugural partnership for deploying high-density power EDCs; this month, Duos closed its first contract spanning two newly launched business lines, GPU as a Service and high-power colocation service for AI infrastructure. Through our GPU-as-a-Service agreement, Duos will place 2,304 NVIDIA GPUs across our Edge Data Center platform, producing reoccurring revenue via a GPU rental model, purpose-built for enterprise and AI workloads; this contract should bring in roughly $176 million in revenue across a 36-month term, with margins above 80% and anticipated annual EBITDA of about $40 million. In a separate deal, Duos won a high-power colocation contract to supply 4.8 megawatts of critical compute power supporting a leading hyperscaler's high-density NVIDIA GPU cluster, housed inside Duos edge data centers; this contract signifies Duos' entry into the high-power colocation market, where demand for AI-grade infrastructure keeps substantially exceeding supply. Taken together, these contracts represent a major commercial inflection for Duos, creating two distinct and complementary revenue streams within our data center platform and validating Edge Data Center infrastructure at the highest level of the AI compute market. Since we announced these contracts, we've seen strong incremental inbound interest from hyperscalers, neocloud providers and other large-scale compute customers looking for high-density EDC solutions, and we see a substantial opportunity to scale the high-power EDC model through 2026 and beyond; now I'd like to hand it over to our CFO, Leah Brown, who will walk through our financials for 2025. Leah?
Thanks, Doug. It's been quite a year for Duos. 2025 brought substantial top-line expansion, deliberate investment, and real strides toward becoming a company that's both sturdier and better able to scale. I'm genuinely eager to take you through the financials for the full year and point out the main operational catalysts behind our results. Consolidated revenue for 2025 came in at roughly $27 million. Earlier, the company had guided to $28 million for 2025. We fell short of that figure, yet we booked just above $1 million of deferred revenue tied to Technology Solutions — it's under contract, the cash has come in, and we'll recognize it as revenue in 2025. That $27 million for 2025 stands as a major step up from $7.3 million in 2024, a rise exceeding 270% year-over-year. Services and consulting revenue under the asset management agreement with new APR Energy largely powered this expansion, reaching $22.4 million in 2025 against $900,000 in 2024. Gross margin came in much stronger in 2025, with $7.9 million of gross profit, or about 29% — a substantial gain versus the prior year. Better cost absorption and sustained operating efficiency drove that. Net loss for 2025 was about $9.8 million, an improvement over the $10.8 million net loss posted in 2024. Higher revenue and a much stronger gross margin were the primary reasons for that year-over-year gain. As we mentioned on the Q3 earnings call, reaching positive adjusted EBITDA mattered a great deal for the company, showing early rewards from revenue scale and better margin. I'm happy to say we extended that momentum in Q4, posting positive adjusted EBITDA for a second quarter in a row; that consistency carries weight and shows the Q3 result wasn't a one-off, but instead a continuation of improving operating performance as the business scales. The back-to-back gains from Q3 into Q4 bolster our confidence in where the business is headed — driving more revenue volume, better gross margin and a more fee-based cost structure. Now to the balance sheet. The company closed 2025 with roughly $63 million in total assets, a meaningful year-over-year increase. Cash rose considerably versus the prior year, helped by a capital raise during the year that bolstered liquidity and improved our capacity to support operations and planned investments. Property and equipment is another bright spot on the balance sheet. Each grew notably year-over-year, reflecting ongoing investment in infrastructure and assets needed to support program execution and long-term growth initiatives. Current contract liabilities above $5 million underpin the company's future revenue recognition; on the equity side, capital raised during the year bolstered our balance sheet and liquidity, while continued investment in the business aligns our strategy to scale operations and drive longer-term value creation. 2025 was a transformative year for Duos Technologies Group. We scaled revenue considerably, strengthened our liquidity position and made strategic investments that position the company for greater operating leverage and margin expansion ahead. Like we mentioned earlier, things have stayed pretty level in the Rail segment. So because of that, we're getting out of the rail business and shifting our resources toward helping our Edge Data Center segment keep growing. Now, on to what we expect for 2026. Across every business line, we're guiding to total revenue somewhere between $50 million and $55 million. That projection takes into account growth from our existing core operations as well as the newer efforts Doug will walk you through, and we think it sets us up well for a solid year. Given how revenue recognition is timed, we anticipate a sizable chunk of that revenue landing in the back half of the year, which lines up with when we expect to reach positive EBITDA. Between our investments and the bigger revenue opportunities in front of us, we feel confident about executing and continuing to build a company that's stronger and more profitable. Doug, I'll hand things back over to you for any further remarks.
Leah, I appreciate that. Before we start taking questions, I'd like to repeat how privileged I feel to be stepping in as your new CEO. The fresh strategy centered on data centers is the new Duos Group — Duos Technologies Group — and we're positioned for tremendous achievement; we've received worldwide acknowledgment through the Innovation of the Year award at the biggest data center and telecom conference, held at Pacific Telecom Council, 2026, in January. In addition, we've been put forward for breakout success as North America Digital Infrastructure Leader of the Year from the Tech Capital Global Awards, which takes place in May. These worldwide acknowledgments simply confirm that we're headed in the right direction at Duos, with a thriving future in front of us. We recognize that we have a fresh focus, and this marks a shift away from our legacy business of the past. We're making moves to guarantee that the updated messaging connects with the market and that we'll receive the proper market coverage going forward. We're going to bring on an IR firm for support, and we anticipate that a number of analysts will cover our new focus and business activities soon. With that said, I'll now turn things over for questions. Operator?
Yes, I want to offer my congratulations to you, Doug, and to everybody at Duos. The expansion you've achieved has simply been incredible. Here's what I'd like to ask: you said demand is still extremely, extremely robust. Do you have any concerns about rivals moving into this space? And what is Duos able to do so it holds the upper hand and can stay competitive should new players arrive?
Excellent question — and this is exactly why we run the company the right way. You're going to watch some companies show up in this market, much like you likely just saw that press release from Crusoe; they are getting into it. In terms of constructing modular data centers of 5 to 10 to 20 megawatts. They rank among the biggest players in this industry; they're behind Stargate, they're enormous. So that fact alone confirms we're in the correct market. But here's what we've accomplished, and this part is remarkable — I only recently returned from GTC, and everyone there kept saying how worried they are about deploying with modular because GPUs are incredibly sensitive to particles and dust; and strangely enough, the greatest strength of our business is that we secured a patent in September known as the clean room — we literally hold a patent for something that sits atop our — it attaches to our modular data center and purifies the air before you step inside. So every particle on your body, on your gear gets blown away and filtered out, and then you actually enter the data center. That matters enormously when it comes to deploying, because here's what will happen: the GPU makers such as NVIDIA and anyone producing chips, anyone building servers, they will not stand behind their warranties once the fans become dirty and dust gets inside them. So that represents a massive advantage for us, and it will help set us apart from the rivals arriving in this market. You're going to see them, but we're the only company that has deployed, as a prime example, 15 pods. I dare anyone who enters this business without a 3D rendering to go physically inspect their pod. A customer flew in from China just last week, and they traveled to Corpus Christi and out toward our pods, purely to observe our manufacturing capabilities; so it appears we may gain a new customer on the hyperscale side, potentially. So we possess the experience. We've actually done it; we can genuinely show people our markets where they can physically visit to see our customers, see the year burning, and see how the facility operates. So we welcome the competition, but we're solid in our position.
That sounds promising. For my final question, I'd like to hear about your longer-term strategy. You've concentrated on rural areas that lack adequate service — are there any intentions to expand into larger markets? Additionally, regarding the Chinese customers or partners you brought up, do you expect to pursue opportunities abroad at some point?
Absolutely, that's a wonderful question. At present, we're concentrated on Tier 3 and Tier 4 markets, and I'd like to explain the reasoning behind that. When it comes to deployment requirements in a Tier 3 market, I'm able to install my pods and secure power access within 90 to 120 days. If we enter a Tier 1 market, we're up against bigger data centers and existing infrastructure; our builds will move quickly. So how do you approach that? You target areas where power is easy to get; substations are already there with 5 to 10 meg of capacity available, and you can get permits much faster. That's why we'll keep concentrating on Tier 3 and Tier 4 markets — that segment is enormous, and it will stay enormous over the next 10 years. As for your question about going international, after we begin deploying at scale domestically and keep progressing, we'll be open to it. But right now, our #1 focus is in the U.S. into the Tier 3 and Tier 4 markets.
Nice work on the quarter. Doug, just a few things I want to clear up quickly. I believe you said you expect to have — or actually you already do have — 5 new EDCs being produced that will be rolled out before the year ends, assuming I caught that correctly. Do all 5 of those EDCs relate specifically to the GPU-as-a-service agreement you recently closed?
No, all 5 of those EDCs are dedicated to markets where contracts are already in place; there are markets in Georgia, and we're collaborating with the utility to install on their network too. So those are the standard pods we roll out, the ones we've already rolled out. Just like the 15 we've installed, they're the same. So yes. And let me clarify a bit, since that could help address many questions others might have as well. We're continuing to build using the same model. At our core, you target education, health care, and local government in these markets. But what we're doing at the factory is constructing the pod with additional power. So we're rolling out these units with the same idea, in the same locations, but we're producing more at scale, which lets us serve higher density users. So yes, that's the model.
Understood. Turning back to that initial GPU-as-a-service client you signed not long ago — when should those bigger pods, so to speak, be installed and start bringing in money?
July and August remain on schedule for us. Given permits and that sort of thing, I'd like to tell you August, though things are looking solid. So it's more around August.
That's incredible. And since it sort of connects to what Leah gave us as guidance — Leah, assuming you're on the line, I believe I jotted down $50 million to $55 million in revenue that we should expect for this year. Could you help us understand how that revenue is split up, please?
Yes, thanks for asking that; regarding the revenue figure we're looking forward to this year, on a holistic basis we're certainly expecting to hit that total. When it comes to individual business lines, we as a company don't provide specifics. Still, on the whole, we do expect to reach that guidance.
All right, got it. And since we're on that topic, you brought up PP&E sitting at just over $27 million. Clearly that's a huge jump versus last year, and it's also $12 million higher than what you had in Q3. Could you walk us through what makes up that PP&E?
Sure thing. Most of our PPE consists of our Edge Data Center. We currently operate 15 data centers, and we've begun purchasing ahead for the following batch, which will come online during 2026. So that accounts for the bulk of it. Yes.
Excellent. That's really helpful. And the final item from my side, then I'll come back in. Doug, I believe you said you'd locked in 4.8 megawatts of power for — my assumption is you mean the GPU as a service contract. The original LOI, if I recall, referenced 10 megawatts set aside for that project. Could you walk us through what accounts for the gap between the 4.8 and the 10 megawatts?
Absolutely. The site was constructed to 10 megawatts. So 10 megawatts is there for use. They're drawing 4.8 of that for critical load. That means I'm able to expand at that site fast, all the way to 10 meg. That site has the potential to reach 20 meg, but securing the additional 10 could require roughly another year. So the advantage in this case is that the site possesses a capability already transformed down at 10 meg. So 10 meg is physically on hand right now, should I choose to sell it. I'd simply construct the pods, put in one more section of pods to reach 10 meg; meaning one more 5 meg cluster of pods.
And when it comes to real estate, so to speak, there's ample room to simply place another 2, 3 or 5 pods down should the need arise.
Yes. So 3 acres exist at that location. And what we have seen is that 3 acres is sufficient. Essentially, if you examine our model, we're putting in 5 meg — it's truly comparable to examining 5 school buses.
Understood — that answers my question fully. Regarding that first technology client, the multinational one delivering GPU as a service, is it your expectation that they'll eventually sign for all 10 megawatts?
Correct. In fact — you can count two clients in that category — absolutely, no question. Right now, together with us, they are looking at 5 more locations, and each of those would be 5 megs. Of course, we've already done the analysis, and we pinpointed 5 locations where power capacity exists, but the plan is to bring this one online plus the Iowa one before anything else, and once that's done we'll come back with details on our speed of execution and what the revenue picture looks like. Still, on the demand side — after returning from GTC, we received 21 inquiries tied to sites in the 5-to-10-megawatt range. Interest within this niche is extraordinary. As I mentioned, competitors entering the space doesn't concern me much. What sets us apart is the speed of our deployment and our ability to locate power — that's something we've figured out. Then there's the clean room factor. A clear illustration: within a single pod, you're looking at somewhere between $10 million and $12 million spent on GPUs alone. If someone already has a clean room, I can't see any reason to avoid using it, since it won't add to their expenses.
Got it. Also, do you think you'll be in a position to reveal the identity of that initial technology client anytime soon?
I can't say for certain. At this point, the NDA is extremely, extremely, extremely tight. So perhaps if we — after we've demonstrated our value to them, that could become possible. To put it another way, they're Tier 1, so we're in good shape.
Thanks for that. I'd like to fit in one more before I go. You brought up a $10 million backlog within Tech Solutions that you anticipate converting into revenue during this year. Is that normal for this line of business, where a signed contract might need multiple quarters before it actually flows through to the revenue line?
That's right, precisely. Let me walk you through an example. We move a great deal of product — and our pipeline is enormous. A large portion consists of items such as cabinets, PDUs, and fiber connectors; those turn around in 60 days, 90 days at the very most, correct? Okay, so we book the order, and then it goes out the door fast. But for UPSs and other switchgear, lead times run six to eight months — a few stretch to nine. So here's what happened: we landed a sizable booking late in the year, yet billing it took us three months, right? Larger products just move slower. That said, anything we book that sits in the pipeline — the stuff you watch us report this quarter and the next — all of it gets built this year, since most of it is, I wouldn't call it off the shelf exactly, but more like UPS, PDUs, cabinets, cold aisle containment, those kinds of things. And there's plenty of it.
That's amazing. Well done, everyone.
I appreciate that. This is exactly what I'm here for. What a great question. Thank you, sir.
Perhaps I can add to what Dan just asked. So it's not just that $10 million from the distribution business is actually being recognized as revenue. Is $10 million, say, a quarter a normal run rate for that operation? Is that a really big quarter? Is that on the low side? Clearly, I'm not asking for hard guidance, just hoping to understand what you anticipate or what that capability < > under something like a normalized scenario.
Yes; and this business is new for us, but what we're observing is when we're able to recognize it and how steady it is; so let's say the funnel exceeds $150 million depending on which product it is.
Just to make sure I've got this right — when you mentioned the annual capacity for the funnel, is that the top-end figure you'd be able to achieve in...
That $10 million figure covered a 2-month span, and it was back when they were just getting going. So clearly, what we're anticipating now is much bigger than that.
yeah, stepping back in and then out again is going to stick around somewhat more effectively. I believed your comment was that the pipeline sits at $150 million. Is that something like a yearly TAM, or the capacity you folks could handle? Did that figure actually reach my ears?
correct, that figure comes from a pair of sales reps we brought aboard, and it represents what's sitting in their pipeline for the present year — plus that covers merely 3 months of conducting business. That team is brand new for us. I mean, check it out, on the data center side they purchase $1.6 billion of product, correct? So in this sector, whether you believe it or not, that's just typical.
So you could reasonably say that if people viewed losing that $20 million 2-year deal negatively, the upside is much bigger. You brought up substituting that revenue, but from what it sounds like, in a normal environment this might be several times larger.
Precisely, and that's the reason we took it on. Nico, one quick point on that division. Keep in mind, the biggest factor behind us taking on that division is that in today's market, it's common knowledge that constructing a megawatt runs somewhere between $10 million and $13 million, correct, to construct a megawatt. The reason they're coming to us is that my cost to construct one megawatt is $6.5 million. How is that possible? Our infrastructure team now sources straight from the factory. That means we skip distributors like Wesco or Graybar, which takes 20% to 30% off the line item since we purchase directly.
What I'm hearing is that you're providing a solution that can be deployed much faster than, say, a standard data center the size of a football field, and for less money.
Correct.
Has there been any consideration given to getting rid of that cheaper option and simply targeting an improved margin profile?
Correct. Even so, we're able to get things up and running faster. Keep in mind that this means capital expenditures aren't quite so heavy. You end up putting in 5 megs for $25 million — that's a substantial gap.
So most of what I've got here is really just asking you to clear some things up. There are clearly a lot of moving pieces, no question. I'm simply attempting to figure out what the company was — you've got the AMA, the equity software — and then it shifts toward this modular data center — school, hospital, anchor tenant — and the metrics surrounding that came across as extremely black and white, such as cost, what the revenue opportunity looks like. And then it feels like we're kind of pivoting once more. So I just want to get my head around all these moving pieces. And maybe the — it would help if you could clear up the deck that's posted on your website from February, I believe that's the month. Is that, like, solid information? There are just a few discrepancies in the metrics between what appears on the slide and, say, what got reported. And I've just got a few clarification questions. I'm simply wondering, like, how firmly fixed those numbers were based on that particular presentation.
Yes. So actually, following — clearly, following the call, we're going to refresh it, because at this point we've recognized and disclosed certain information. That's going to get updated. But keep in mind — and I don't want to create confusion. I'm attempting to — that's the reason I'm trying to alter the model here somewhat. Our business has 2 components. One would be the Edge Data Center business, and the other one is the infrastructure. Within the Edge Data Center business, the GPU business sits inside the Edge Data Center business. Keep in mind, it's the same pod. It's the same concept. It's only that I'm building them larger. Just view the GPU as another category of customer. So I'm simply bringing in other categories of customers. So it's the same model, and the revenue is much higher, clearly, since they're consuming power. We generate money from power space and cross connects, correct? So the more power we move, the more money we bring in; so — but clearly, the CapEx rises in the pod cost. The model — and I'm quite certain we shared this — the model on the GPU is a major difference. Prime example, keep in mind, our pod model at 15 cabinets runs $350,000 to $400,000 a year. That's the objective, correct? From that, if you stack it against the GPU model, 1 meg, you're looking at $1 million a year; so at 4.8 megawatts, you're now at nearly $1 million a month. So why not build the pod larger and bring in the customers who require that power. When for us it's nothing more than at the factory, we simply install bigger panels in.
So, when you refer to that identical model, you've mentioned only the initial, the regular edition of this being deployed in, sort of, Tier 2 and Tier 3 locations or countryside places where you find, like, a 500-mile data center, what is the…
Nico, your audio is breaking up; I can't make out what you're saying.
I believe the service I'm getting right now is pretty poor. What I'm trying to ask is whether the metrics really vary unit by unit, because with the earlier versions or the model it seemed far more standardized. And when you refer to that same model, are they being placed in particular spots where, rather than a colocation arrangement that still includes the hospital and the school, it sits in a rural community and simply has a smaller portion available for leasing, basically, to other companies in that town at this point...
Yes, that's — you've got it exactly, Nico, that is precisely right. Our anchor customers, then, are what we consider our core customers: education, health care, and enterprise within that market, correct? Then the carriers arrive to occupy space for peering and cross connecting with one another. That's the best — somebody's talking over someone, I apologize for that; but yes, Nicco, if you're able to hear me. That's the model as it originally stood. That's the reason we remain committed to it; we're merely adding additional capacity so we can attract those customers requiring greater density. So that market is something we serve at all times. And that is what assists us in reaching those Tier 3, Tier 4 markets, particularly regarding permitting and all of that, since we stay low on the radar. We aren't 10, 20, 30, 40 megawatts that they'd need to build out that stream within the community; what we pursue is power already in place. Power that is in excess, which the utility is eager to profit from. Consequently, it benefits the surrounding community too. in tax dollars. So they actually give us a warm welcome.
Yeah. I was curious whether you could sort of speak about scale as we move into the back half. From a gross margin standpoint, what does the model look like? And then, given all the selling of the Rail business plus winding down that management contract, what sort of OpEx ought we to anticipate going forward?
We'll address this quickly. Let me handle the rail piece. Regarding the Rail business, our hope is to offload it or shut it down within the coming 60 days; that's what we're aiming for there. So currently there's no cash burn from that operation on our side. So with luck, we'll get out of it. It releases a substantial amount of SG&A; clearly we won't be carrying that employee burden and every other cost. So that's a positive, and it ought to take place within the next 60 days. But I'll pass things to Leah for your figures there.
Absolutely, Carl — good afternoon to you. Yes, so an improvement in gross margin is something we'd anticipate during the back half of the year. As a quick reminder, because of how revenue recognition works for certain business lines of ours, that revenue is going to show up in the latter six months of the year. So we're anticipating gross margin of approximately 76%.
Question from an unidentified analyst: On gross margin, shouldn't it be — well, the data centers on their own carry gross margins of the 70 to 80 type.
Right, precisely; thus we ought to be looking at roughly — regarding gross margin, you're sitting at about $7 million, $6 million as the year draws to a close. Right, precisely; so when we file our report here in May, you'll get a look at our Q1, yet it will be possible for you to observe that revenue climbing in Q3 and Q4.
On the operating expenses side, those really ought to be declining concurrently.
That's right, the OpEx.
And so, Doug, given what you mentioned today, what's your sense of how long this demand situation is going to persist?
The elevated demand we're experiencing right now — for instance, when I show up at GTC and 21 individuals are all vying to speak with me in order to get contracts signed — I believe that stays robust over the coming 3 to 4 years. After that, what matters most for our business is the existing major data centers, and I believe we may have discussed this previously, those existing major data centers are going to view us as a hub and spoke, since they'll be eager to seize the markets where we operate, such as the Dumas, such as the Corpus Christi, Lubbock — these Tier 3 markets we're entering — they require compute available there. The mobile operators do as well. As we move to 6G — we're at 5G, we're moving to 6G now — they require the compute positioned out at what we refer to as the eyeballs. So all of that data is going to require a great deal of fiber to return, a great deal of network, correct? So they wish to be in a position to possess that network and they wish to possess that customer. The optimal method to accomplish that is clearly to purchase these numerous data centers in every location, and bring them back to the core. Since, frankly, they're all returning to a core regardless. So it makes total sense. So I believe the growth is going to be quite strong and exceptionally strong in the 3 to 10 meg range, because at present — and I just performed this exercise for a different prospective client, who required 2 meg worth of power, 2 meg, which doesn't seem like much these days, but it is a lot — I was unable to locate it anywhere in the country within a single data center. I'm referring to a legacy data center. So the market is overlooking the need for the 10 to 15 meg data centers. And a prime example, like Johnson & Johnson, they keep their equipment at a local data center. They go to something like a QTS. They go to Flexential. That's where the house. They don't go to a hyperscaler. They don't go to these large ones they're constructing. We're losing sight of the fact that the demand exists and they're still expanding. So I believe you're going to see the market over the next 5 to 10 years concentrating on that 10 to 15-megawatt range. So we have a long haul, but we do need to build quickly.
You gave us quite a bit of detail on GPU-as-a-service, the economics, and the revenue from it. I'm trying to figure out the revenue exit run rate for this year. So could you add a bit more detail on the high-density EDC — how many total megawatts, and what is the revenue value per megawatt for that high-density colocation customer compared with leasing and GPUs that you bought?
Sure; regarding the GPU model, let me step back. The objective for this year is to deploy 25-megawatt. That can happen through 300 kW pods that we deploy. At present, we have 15 of them on the ground at 300 kW; but the total megawatt — because that's the basis on which we're being judged right now, everyone is judged by megawatts, not by kilowatts or cabinets — so the plan is 25 megawatts; and when we examine the GPU model, for every megawatt, we're looking at $2 million a year in revenue. That's exactly on target. That's what they're billing, that's what the industry shows, and that's what we're building to. So obviously it's a very strong model for housing GPU for customers.
Okay, I appreciate all of you being here right now, and we can’t wait to talk with you during the Q1 earnings call. Many thanks for giving us your time.
As we wrap up this call, I want to share Duos' safe harbor statement, which covers key warnings about forward-looking statements that were given today. Within the context of the Securities Litigation Reform Act of 1995, this earnings call includes forward-looking statements. Terms like believes, expects, may, will, should, anticipates, plans, along with their negatives or comparable wording, are meant to flag forward-looking statements. Be advised that such statements don't promise future results or occurrences and are exposed to various uncertainties. Risks and additional factors, a lot of them outside our ability to control, could affect how accurate these statements and the underlying projections turn out to be, and could lead Duos Technologies Group's actual results to vary materially from what those forward-looking statements anticipate. Among these risks and uncertainties are those outlined in Item 1A and Duos on annual report on Form 10-K, which is expressed incorporated herein by reference, plus other factors that might be described from time to time in Duos' filings with the SEC. We appreciate your participation today in Duos Technologies Group Fourth Quarter and Full Year 2025 Earnings Call. You may now disconnect.
Read the full transcript via the API
You're viewing the first half of this call. Get the complete Duos Technologies Group, Inc. transcript - plus 131,000+ transcripts from 7,000+ companies, speaker segments and full-text search - through the EarningsAPI REST API or hosted MCP server.
Get an API key View API docs →For developers and AI pipelines
Programmatic access to Duos Technologies Group, Inc. earnings transcripts and 131,000+ others is available through the
EarningsAPI REST API and the hosted MCP server.
Quarterly plans from $162 - full transcripts, speaker segments, full-text search,
and the /api/v1/transcripts/recent polling endpoint for ETL pipelines.