Logo of jester cap with thought bubble.

Image source: The Motley Fool.

DATE

Tuesday, May 5, 2026 at 8 a.m. ET

CALL PARTICIPANTS

  • President & Chief Executive Officer — Benjamin Wolff
  • Chief Financial Officer — Trevor Thatcher
  • Investor Relations — Brian Siegel

TAKEAWAYS

  • Revenue -- $3.5 million, up 107%, meeting internal expectations; includes $1.7 million in product revenue (primarily manufacturing) and $1.8 million in engineering services revenue (including GuideTech).
  • Backlog -- Ended at $17 million, up from $13.5 million, after recognizing $3.5 million in revenue and booking $7 million in new contract awards (net of recognized revenue).
  • Full-Year Revenue Guidance -- Management reiterated $24 million to $27 million, implying 357%-415% growth and expectation for sequential quarter-over-quarter increases and higher growth rate in second half as backlog converts.
  • Gross Margin -- Consolidated gross margin was approximately 30%; initial product margins compressed by low manufacturing capacity utilization and first article costs, with improvement expected as utilization and revenue ramp.
  • Operating Cash Usage -- Approximately $10.2 million used during the quarter, above the guided $8 million-$9 million range, driven by inventory build for BRAIN flight computer, accelerated hiring, and production first article costs; $43.7 million liquidity as of March 31.
  • Operating and Net Loss -- GAAP operating loss was $11.9 million; GAAP net loss was $12.6 million ($0.28 per share); non-GAAP net loss was $10.2 million ($0.23 per share), adjusted for warrant liabilities and stock-based compensation.
  • ATM Share Sale -- ~890,000 shares sold through ATM program, raising $6.5 million at an average price of $7.35 per share.
  • Manufacturing Capacity -- Facilities operating at roughly 30% utilization, leaving ample room to increase margin and scale revenue without substantial additional investment.
  • SwarmOS Demonstration -- Achieved coordinated heterogeneous drone swarming with Gremlin-X UAV and Red Cat platforms, demonstrating distributed, autonomous collaboration without preprogramming or central control.
  • BRAIN Flight Computer -- Received $500,000 initial order for BRAIN X2 variant; progressed scaled-down FC1 variant development.
  • Defense and Aerospace Contracts -- Secured a mission-critical propulsion subsystem contract with a major U.S. defense prime for an existing U.S. missile program, expected to contribute nearly $1 million in revenue this year; awaiting government evaluation of first article.
  • Space Domain Entry -- Awarded AFRL HANGTIME project to integrate SwarmOS with space-based satellite sensor grids and contracted with Portal Space Systems for maneuverable spacecraft platform development.
  • IQ 2.0 Commercial Deployment -- Initial robot integration underway with first customer for non-contact surface treatment; established a partnership with a systems integrator, enabling indirect channel access.
  • Patent Activity -- Secured new patent and filed two additional applications for AI software and swarming technologies.
  • Northern Strike 26-2 Participation -- Invited to demonstrate SwarmOS on four UAV platforms in a Department of War joint exercise; validation of cross-platform collaborative swarm autonomy and real-time mission adaptability.
  • Relentless Wolfpack Industry Day -- GuideTech selected as one of 14 invited; showcased SwarmStrike (low-cost cruise missile) networked by SwarmOS, targeting a sub-$150,000 unit cost per swarm strike.
  • Revenue Model for SwarmOS -- Software priced at approximately 10% of UAV cost; typically sold as a one-time upfront license per drone, with expectation of high replacement frequency due to combat attrition.
  • Backlog Conversion and Pipeline Visibility -- High confidence stated in achieving full-year guidance based on committed backlog and ongoing pipeline opportunities across software, manufacturing, and services.
  • Red Cat and Draganfly Partnerships -- Expanded Red Cat partnership with economic terms in place, enabling go-to-market; Draganfly flight test demonstrations planned for current and subsequent quarters.

Need a quote from a Motley Fool analyst? Email [email protected]

RISKS

  • Operating cash usage exceeded guidance due to inventory build, hiring, and first article production costs, with management stating, "We expect quarterly cash usage to tend -- to trend toward and remain within the previously guided range as revenue and margins ramp."
  • Product margins in manufacturing were affected by "low capacity utilization and first article costs," with improvement contingent on future volume and approval of first articles.
  • Revenue was negatively impacted by "the federal government shutdown, which temporarily delayed program activity across several of our defense contracts. That work was not." Management identified this as a timing, not demand, issue.
  • Management acknowledged, "not every pursuit is going to result in a win," as the company targets "very large programs" with a "relatively small company."

SUMMARY

Palladyne AI (PDYN 5.83%) reported rapid year-over-year revenue growth, supported by a significant increase in backlog and new contract awards but noted revenue timing delays from a recent federal government shutdown. The firm highlighted differentiated product validation milestones, including live autonomous drone swarming demonstrations, initial defense and space contracts, and participation in high-profile Department of War exercises slated to serve as operational gateways for broader adoption. Management reiterated strong visibility into full-year revenue guidance, reinforced by increased bookings, a strengthened partnership ecosystem, and evidence of high demand for both defense and commercial solutions. The company's liquidity position remains robust, with an expectation to manage cash burn as margin improvement and revenue ramp materialize.

  • Palladyne AI's commercial autonomy offering, IQ 2.0, secured its first industrial deployment in a surface treatment application through a systems integrator partnership, a model the company believes can accelerate future enterprise sales.
  • Management clarified the SwarmOS revenue model as a per-drone license equal to roughly 10% of the UAV’s hardware cost, with no recurring fees, citing anticipated high replacement rates in defense procurement.
  • BRAIN flight computer product line advanced with a $500,000 order and further technological development, reflecting early customer traction and portfolio scaling.
  • GuideTech, an acquired entity, was recognized at a major industry event for integrating SwarmOS and SwarmStrike, targeting cost-efficient, networked weapon solutions for the Department of War.

INDUSTRY GLOSSARY

  • DECA (Decentralized Embodied Collaborative Autonomy): Autonomous AI architecture enabling distributed, predictive, cross-platform machine collaboration in real-time without reliance on centralized cloud intelligence.
  • First Article: An initial sample built to specification for evaluation before commencing volume manufacturing on defense/aerospace contracts.
  • SwarmOS: Palladyne AI's proprietary autonomy software for multi-platform, collaborative, adaptive control of UAV swarms and robotic systems.
  • IntelliSwarm: Product suite combining SwarmOS software and BRAIN hardware for field-deployable, scalable drone and robot autonomy.
  • BRAIN Flight Computer: Modular flight control hardware supporting advanced autonomy and collaborative behavior for UAVs and autonomous vehicles.
  • GuideTech: Acquired engineering services provider, specializing in AI-powered collaborative unmanned systems and included in segment reporting.
  • ATM (At-the-Market) Program: A financing method allowing a company to sell shares directly into the market over time to raise capital at prevailing prices.
  • Wolf Pack Swarming: Autonomy mode enabling multiple drones to independently perceive, predict, and collaboratively complete missions without scripted coordination or central control.
  • SwarmStrike: Low-cost autonomous cruise missile platform intended to operate collaboratively in networked swarms via SwarmOS for enhanced effect and cost efficiency.
  • FC1: Scaled-down commercial variant of the BRAIN flight computer developed for broader deployment and cost optimization.

Full Conference Call Transcript

Operator: Greetings, and welcome to Palladyne AI First Quarter 2026 Earnings Conference Call. [Operator Instructions] As a reminder, this conference is being recorded. It is now my pleasure to introduce your host, Brian Siegel with Hayden IR. Thank you. Please go ahead.

Brian Siegel: Good morning, and welcome to Palladyne AI's First Quarter 2026 Earnings Conference Call. Joining me on the call today are Ben Wolff, President and Chief Executive Officer; and Trevor Thatcher, Chief Financial Officer. Earlier this morning, Palladyne issued a press release announcing financial results for the first quarter ended March 31, 2026, along with updated commentary regarding backlog and reiterated its 2026 revenue guidance. A copy of that release, along with the accompanying financial tables, is available on the IR section of Palladyne AI's website. Today's call will include prepared remarks from Ben and Trevor, followed by a question-and-answer session. During the call, management will make forward-looking statements within the meanings of the federal securities laws.

These statements include, but are not limited to, statements regarding Palladyne's 2026 revenue guidance, expected backlog conversion, anticipated quarterly operating cash usage, product development milestones, commercialization timelines, defense program activity, potential customer adoption, market opportunities and future strategic positioning across aerospace, land and maritime domains. Forward-looking statements are based on current expectations, assumptions and beliefs and involve risks and uncertainties that could cause actual results to differ materially from those expressed or implied. These risks and uncertainties include, among others, Palladyne AI's ability to execute on development programs, convert backlog into revenue, scale production, manage operating expenses, integrate acquired businesses, secure additional contracts, maintain liquidity and navigate evolving defense and commercial market conditions.

These and other risk factors are described in detail in Palladyne AI's filings with the Securities and Exchange Commission, including its annual report on Form 10-K and subsequent filings. Palladyne undertakes no obligation to update any forward-looking statements, except as required by law. In addition, during this call, management will reference certain non-GAAP financial measures. In general, management will adjust for acquisition and other transaction-related expenses, stock-based compensation expense, noncash warrant income or expense that are marked to market quarterly based on changes in the company's stock price, expenses related to the change in contingent consideration liabilities associated with closed acquisitions and any tax impact these items may cause.

A reconciliation of these non-GAAP measures to the most directly comparable GAAP measures is included in this morning's press release. With that, I'll turn the call over to Ben.

Benjamin Wolff: Thank you, Brian, and good morning, everyone. I want to cover 3 things this morning. First, I'll walk you through the first quarter's results. Second, I'll discuss what we accomplished operationally across the business in Q1. And third, I'll talk about what's on deck because the opportunity in front of us is larger and more concrete than it has ever been. Our revenue for the first quarter increased 107% year-over-year to $3.5 million, which was in line with our internal expectations. Having said that, revenue for the quarter could have been even better were it not for the federal government shutdown, which temporarily delayed program activity across several of our defense contracts.

That work was not canceled and the contracts remain in place. The work simply shifted in timing, and it remains in backlog. I want to be clear about this because I know it will come up when we get to questions. The shutdown created a revenue timing issue, but not a demand issue. In short, the business performed as we internally expected. With respect to backlog, we entered Q1 with approximately $13.5 million, recognized $3.5 million in revenue during the quarter and exited with approximately $17 million. So we added approximately $7 million in new contract awards during the quarter, net of revenue recognized.

That is a meaningful bookings number, and it gives you a sense of the activity level that doesn't show up in our reported revenues. That backlog provides us with good visibility into the revenue ramp ahead, and we are reiterating our full year 2026 revenue guidance of $24 million to $27 million, which implies approximately 357% to 415% growth compared to 2025. We expect revenue to grow sequentially each quarter with the growth rate accelerating in the second half of the year as backlog converts and new contracts are awarded. Our operating cash usage for the quarter came in modestly above our guided range of $8 million to $9 million on average per quarter.

This was driven largely by 3 things. First, we began building inventory for BRAIN flight computer production for existing customers. That inventory build is not a cost issue, it is a working capital investment tied to near-term revenue, and we expect it to convert as we fulfill those orders. Second, we accelerated some hiring based on the strength of new opportunities we saw in the first quarter. And finally, we incurred costs in our manufacturing business to develop and produce first articles for some of our more recent contracts, but have not yet transitioned to full rate production in large part due to the government shutdown. Again, this is simply a timing issue.

We expect quarterly cash usage to tend -- to trend toward and remain within the previously guided range as revenue and margins ramp through the remainder of the year. Liquidity as of March 31 was $43.7 million, and we believe we remain well positioned to execute our 2026 plan. Before I get into the operational highlights, I want to spend a few minutes on a topic that I think provides important context for everything else I'm going to say. There remains a lot of confusion in our industry about what our technology actually does and how it is different from what other companies are offering.

I published 2 white papers with my co-Founder, Denis Garagic, during the quarter, specifically to address that confusion, and I want to walk you through the core ideas. The first paper is about what we call Decentralized Embodied Collaborative Autonomy, or DECA for short. The central point of that paper is pretty straightforward. Most software platforms that people associate with modern artificial intelligence lives in massive centralized data centers. They are optimized for thinking, analyzing data, recognizing patterns, generating language, supporting human decision-making, and they are very good at that. The challenge is that machines operating dynamic real-world physical environments cannot rely on that kind of centralized cloud-based intelligence.

A drone can't wait for a round-trip comms link to a data center. A robot on a factory floor needs to react in fractions of a second. A missile system operating in a communications denied environment has to be self-sufficient. Nature actually solved this problem a long time ago. Think about how human intelligence works. At any moment, the human body is generating an enormous amount of sensory data from sight, sound, touch and our other sensors. Almost none of that ever reaches conscious thought. The vast majority of it is filtered, processed and acted upon locally, automatically by fast systems that operate far below the level of conscious reasoning.

You don't think about how to keep your balance when you walk. You do not reason through catching a falling object. Those things happen automatically, locally in real time. And that architecture works because it has to work that way. There is no other way. The physics does not allow for anything else in a world where reaction time and energy efficiency are constrained. Our technology is modeled on that same biological principle. Intelligence lives on the machine. Perception is filtered locally, not centrally. Decisions are made predictively rather than reactively. Machines collaborate through decentralized interactions rather than waiting for instructions from a centralized controller.

That is the essence of DECA, and that is what we have built into our products. The second paper applied to the SAE automotive autonomy framework developed by the auto industry for self-driving cars and applying that to drone autonomy and swarming. For reasons we did this -- the reasons we did this is that there is enormous confusion in the market about what autonomy and swarming actually mean. We recognize that all software is not the same. It has different purposes, uses, capabilities and compute requirements. Similarly, not all autonomy is the same and not all swarming is the same, but the same basic words are used to describe a myriad of different capabilities. We are changing that narrative.

The paper walks through a clear taxonomy from basic remote control all the way up to what we call Oracle-Class Wolf Pack Swarming, which is decentralized, predictive, collaborative autonomy where the swarm is not just reacting to what it observes in the moment, but participating -- but anticipating what is likely to happen, positioning assets and allocating sensing resources in advance of events rather than in response to them. As far as I know, we are unique in having developed this capability, and we are actively working to bring it into the commercialized version of SwarmOS.

Our current SwarmOS product already operates at the Wolf Pack Swarming level, which is genuinely different and more capable than what anyone else in this space is offering, regardless of how they describe their systems. Oracle-Class is the next step, and we are further along toward it than any competitor we are aware of. I encourage investors to read both papers. They are on our website. They're not long, and I think they will give you a much clearer framework for understanding what we are building and why we believe it is different and highly valuable. Now let me walk you through what actually happened in the business during the quarter.

The most significant operational milestone of the quarter was a demonstration of true heterogeneous autonomous swarming. We flew Gremlin-X, our reusable mini bomber UAV platform, that was previously known as Project Banshee, running our IntelliSwarm product in a coordinated test swarm alongside multiple Red Cat platforms also running our SwarmOS autonomy software. I want to explain why this is different from what you typically hear described as drone swarming because the distinction matters a great deal. A lot of what's called swarming in our industry is really preprogrammed flight coordination, where the only function that happens automatically is collision avoidance. This is akin to lane-changing sensors on a modern car. Otherwise, the drone follows a script.

If you have seen drone light shows, that is a form of swarming, but every drone knows exactly where it is supposed to be at every moment because someone programmed it that way in advance. If something unexpected happens, the system does not know what to do. What we demonstrated in Q1 is fundamentally different. Each drone running SwarmOS was perceiving its environment independently, reasoning independently about what to do, acting on its own judgment within the mission parameters and collaborating with other platforms in real time. There was no script, there was no centralized controller calling plays.

It is the distributed adaptive intelligence that the Department of War says we need, but many thought was 5 to 10 years away at best. It is resilient in ways that preprogrammed systems simply are not, particularly in contested and communications degraded environments. What makes that kind of distributed autonomous operation deployable at scale is the hardware underneath it. During the quarter, we progressed the development of our BRAIN flight computer variants, including a scaled-down version of the commercialized X2 variant called FC1. BRAIN is the hardware that when combined with SwarmOS forms IntelliSwarm, a product deployable at scale across autonomous platforms. We recently received a $500,000 first order from a defense tech company for the BRAIN X2.

Next, we expanded the Draganfly partnership during the quarter by conducting a lab simulation of SwarmOS running on Draganfly's commercial defense platform. The next step is to integrate SwarmOS into their drones and run flight tests. Q1 was also the quarter we established a real presence in the space domain through 2 separate engagements. Through our HANGTIME award with the U.S. Air Force Research Lab, or AFRL, we will integrate SwarmOS with a space-based satellite sensor grid, enabling UAVs to develop even better situational awareness. This is the first planned integration of our collaborative autonomy platform with space-based assets, demonstrating that our AI can leverage data from all domains, air, land, sea and space, to improve mission effectiveness.

Separately, we secured a contract with Portal Space Systems to support development of next-generation maneuverable spacecraft platforms, providing navigation, guidance, spacecraft modeling, embedded software and avionics support. Portal is building spacecraft designed to reposition across orbits on compressed timelines with minimal ground intervention, a class of problem well suited to our edge native architecture. Looking ahead, we see opportunities to expand the partnership to include Palladyne's autonomy capabilities. Through Palladyne Aerospace and Defense, we secured a contract with a major U.S. defense prime contractor to deliver a mission-critical propulsion subsystem for an existing U.S. missile system program, and we expect that contract to contribute nearly $1 million in revenue this year.

This is a validation of our precision manufacturing capabilities and continues to expand our footprint in long life cycle defense programs. This contract is an example of the government shutdown impacting our first quarter revenue as we are still waiting for the evaluation of our first article. On the industrial autonomy side of our business, we are in active deployment with our first IQ 2.0 customer with the initial robot system integration currently underway. This is a non-contact surface treatment application, and it is a use case where IQ's combination of teleoperation and simplified path planning addresses a real industrial problem that no robot manufacturer or AI company currently solves with an off-the-shelf product.

The customers' operations offer what I would describe as a potential land-and-expand opportunity. The initial deployment is one robot. As the customer builds confidence in the system and experiences all that it can do, the natural progression is to add more robots and expand use cases. We think that is going to be the typical adoption pattern for IQ, and is -- it is consistent with how enterprise automation technology tends to scale in industrial environments. Matt Muta transitioned from the Board to an operating role during this past quarter, joining us as President of Commercial and Industrial. Matt has real experience building and scaling enterprise technology businesses and his focus will be on converting the IT pipeline into customers.

We also received a new patent during the quarter supporting advanced swarming and decentralized autonomy architectures, and we filed 2 new patent applications related to our AI software products and technologies. Our intellectual property portfolio is growing alongside our product portfolio, which is important for the long-term defensibility of what we are building. Next, I want to spend a few minutes on the broader context because the environment we are operating in has changed significantly, and I think it is worth being explicit about what that means for us. The Department of War is committing an unprecedented amount of resources to autonomous systems, collaborative swarming, counter-UAS, long-range precision fires, hypersonics and missile defense.

These are not abstract priorities, they are specific trackable programs and budget lines that we are actively engaged with. The Defense Innovation Unit has seen its budget grow substantially and its funding programs, specifically around multi-domain collaborative autonomy. PAE Fires, the Army's portfolio acquisition executive responsible for artillery, missile defense and sensor systems, oversees a set of programs spanning long-range precision weapons, hypersonic weapons, integrated air and missile defense and counter UAS, each of which represents a potential opportunity for our product and service lines. And Golden Dome, the administration's flagship missile defense initiative, is one of the largest single defense investment priorities in a generation.

We are pursuing opportunities across these and other programs and budget lines for SwarmOS, BRAIN, IntelliSwarm, Gremlin-X, SwarmStrike and our engineering services and research and development groups, including the Mark XL program. I want to be honest about this, we are a relatively small company pursuing very large programs, and not every pursuit is going to result in a win, but the alignment between what the Department of War is prioritizing and what we have actually built has never been stronger, and this is the environment in which we are operating. One of the most meaningful near-term proof points for what I just described is our invitation to participate in Northern Strike 26-2.

Northern Strike is a premier Department of War joint exercise hosted August 2 through August 14 at the National All-Domain Warfighting Center at Camp Grayling, Michigan, which was designed as the drone dominance -- which was designated as the drone dominance range in the recently enacted National Defense Authorization Act. It is a joint national training capability accredited exercise involving more than 9,000 participants operating across contested multi-domain environments. It serves as one of the most demanding operational validation environments available to emerging defense technology companies as well as a recognized gateway to operational programs of record.

We will be demonstrating SwarmOS on 4 distinct UAV platforms from 4 different OEMs, including our own Gremlin-X, with each drone collaborating autonomously and managed by a single operator, by a single ATAC interface. The exercise will validate cross-platform swarm collaboration across multiple UAV classes and manufacturers, decentralized decision-making that is resilient to denied or degraded communications, real-time mission adaptation across dynamic conditions and significantly reduced operator burden relative to conventional approaches. For us, Northern Strike is also a direct engagement with military end users and acquisition stakeholders who influence programs of record, which is exactly where we need to demonstrate this technology.

I also want to highlight something that happened just after quarter end, but that directly reflects the work we did throughout Q1 and before. GuideTech was selected as one of only 14 companies invited to participate in the AFRL Relentless Wolfpack Industry Day hosted by the Air Force in collaboration with the Doolittle Institute on April 28 and 29. We were the only company in that group that most people would describe as a small cap. Our inclusion reflects the maturity of what we have actually built. GuideTech's submission combines SwarmStrike, our internally developed low-cost cruise missile, with SwarmOS to deliver a networked collaborative autonomous weapon solution.

We are targeting a cost of less than $150,000 per swarm strike, which means you can put 10 of them in the air for the price of a single conventional cruise missile and then network them through SwarmOS to combine the effects on targets simultaneously. That cost per effect argument is precisely what the Department of War is focused on right now. SwarmStrike has completed its initial flight test, and we are actively advancing the program through multiple government channels. Separately, and I want to be clear, this is a distinct development, a different defense prime participating in the same relentless Wolfpack cohort independently chose to incorporate SwarmOS into their own submission.

They evaluated the platform on its merits and built it into their own hardware solution. We didn't arrange that. That is the beginning of the platform adoption story we have been working toward where SwarmOS becomes the autonomy layer that other companies build on, not just a product we sell directly. Taken together, Northern Strike and Relentless Wolfpack are not isolated events. They are evidence of something broader. The strategy is working. The products are being validated in real operational and acquisition context, and the market is beginning to recognize what we have built. That is what I want investors to take away from everything I have described today.

Let me close by putting all of this in context of where we are in the progression. On our Q4 call, I described our strategy as crawl, walk and run progression, not as separate strategies, but as stages of maturation. I want to come back to that framework because I think it is the right lens for understanding Q1 and what comes after it. 2026 is the crawl year, as I said before. Crawl is about proving that the integrated model actually works at scale, converting backlog into revenue, executing live demos and trials and advancing development stage assets towards defined milestones.

Our wins in the first quarter achieved all of these objectives, $7 million in new contract awards, a successful swarm demonstration across multiple platforms from different manufacturers, 2 new space engagements, active deployment of our first Commercial IQ customer, 2 white papers that established our intellectual framework on the public record, a Northern Strike invitation, a key patent issuance and 2 new patent applications. That is a lot of activity that progresses us towards our objectives. In 2027, we will walk. Walk is when proof becomes repeatable. We expect broader SwarmOS and IntelliSwarm integrations, more IQ wins, more BRAIN wins and expanded defense programs with multiple product-based revenue streams running concurrently.

That is when growth starts to become more systemic and less dependent on individual contract timing. And then we run. Run is where the full vision becomes operational across aerospace and eventually land and sea, where IntelliSwarm enables larger and more complex distributed systems, where the autonomy and propulsion architectures we are developing today start to converge and where the revenue is systemic rather than episodic. In conclusion, we know what we are building. We know why it is different, and we believe the work we are doing in 2026 is laying the foundation for everything that follows. With that, I will turn it over to Trevor.

Trevor Thatcher: Thanks, Ben. I'll focus on our first quarter results, liquidity position and capital outlook. Revenue from the first quarter of 2026 increased 107% to $3.5 million. compared to $1.7 million last year. The increase was due to the inclusion of post-acquisition revenues from the acquired companies. Within that $3.5 million, product revenue, which today is mainly derived from our manufacturing business, was $1.7 million. Engineering services revenue, which includes GuideTech, was $1.8 million. We did not recognize meaningful product development contract revenue this quarter, but we expect this will pick up beginning in the second quarter as awarded business turns into signed contracts as we execute on recently signed contracts and as our existing contracts get extended through contract options.

This quarter represents the first full quarter of revenue flowing from the businesses acquired in November of 2025. Cost of revenue for the quarter was $2.5 million compared to $0.4 million in the prior year period. Consolidated gross margin for the quarter was approximately 30%, which reflects the current revenue mix. Product margins in our manufacturing business were compressed by low capacity utilization and first article costs. As utilization improves and revenue ramps, we expect manufacturing product margins to improve accordingly. Our software products, when they begin generating meaningful revenue, are expected to carry the highest margins in the portfolio, in line with typical software margin costs.

The 30% consolidated figure is not representative of where we expect to be as revenue ramps and mix evolves. Research and development expense was $3.9 million compared to $2.9 million last year, reflecting continued investment in autonomy software, avionics and product development programs from both Palladyne and the acquired companies. As we've discussed in prior quarters, we are investing in Gremlin-X and SwarmStrike development, the former of which was a major focus during the quarter. We expect continued investment over the next couple of quarters to bring that platform closer to commercialization. General and administrative expense was $6.9 million compared to $4.2 million in the prior year period.

This increase reflects the incremental scope of G&A and overhead functions from the acquired businesses as well as select hiring to drive and support growth. Sales and marketing expense was $1.9 million compared to $1.2 million last year, reflecting expanded marketing programs and business development efforts. Operating loss for the quarter was $11.9 million compared to $6.9 million in the prior year period. GAAP net loss for the first quarter was $12.6 million or $0.28 per share. On a non-GAAP basis, net loss for the first quarter was $10.2 million or $0.23 per share.

The primary differences between GAAP and non-GAAP results were a $1 million noncash loss related to change in fair value of warrant liabilities this quarter, driven largely by the change in the price of our common stock and public warrants. In the year ago quarter, we saw a $29.2 million noncash gain from warrant liabilities, $1.2 million of stock-based compensation expense and $150,000 loss related to change in our contingent consideration liability. We believe excluding these items provides a clearer view of our underlying operating performance and cash usage. Turning to liquidity. As of March 31, we had $43.7 million in cash, cash equivalents and marketable securities.

During the quarter, we incurred minimal CapEx and used approximately $10.2 million in operating cash, partially offset by $6.5 million in net proceeds from our ATM program. Backlog as of quarter end was $17 million, up from $13.5 million at the end of 2025, reflecting gross additions of $7 million, offset by this quarter's recognized revenue of $3.5 million. Ben has already announced that we are reiterating our full year 2026 revenue guidance of $24 million to $27 million. This outlook reflects the contribution of the businesses acquired in November, and we continue to expect organic growth across each part of the company on a full year comparable basis.

We also continue to expect total CapEx and OpEx cash burn for the year to be in the range of $32 million to $36 million or approximately $8 million to $9 million per quarter on average for the remainder of the year. The increase from our 2025 run rate reflects ongoing OpEx investments in SwarmOS and IQ, bringing acquired programs to operational readiness and the incremental headcount costs I mentioned earlier. This also includes CapEx for our manufacturing business and the acquisition of several third-party drones to validate SwarmOS's collaborative swarming capabilities on new platforms. Based on our liquidity position and expected backlog conversion, we believe we are well positioned to execute our 2026 plan.

Operator, we're now ready to take questions.

Operator: [Operator Instructions] Our first question is coming from Michael Latimore of Northland Capital Markets.

Mike Latimore: Congrats on the start to the year here. So Ben, I think you mentioned -- I just want to clarify that a defense prime is integrated to ROS. I want to just clarify that you said that? And if so, can you elaborate a little bit? Are you exclusive? Is this related to UAVs or is it multi-domain? Any particular end programs you're dealing with?

Benjamin Wolff: So they have not actually done the integration yet. What I said was -- and the key takeaway is that on a major defense program where they are trying to become the prime on a contract award, they have included our autonomy software as an important element of that submission. So we would -- if they wind up winning that contract, we would wind up being a subcontractor to that prime. It does relate to machines that are flying as opposed to something that's in space or in -- on the sea or on land. So it is an aeronautical type of application.

Mike Latimore: Okay. Interesting. Okay. And then maybe talk a little bit about just your manufacturing operations. What is the capacity utilization now? Where might that go by year-end?

Benjamin Wolff: So right now, we think that we are roughly stated around 30% of our total utilization capability. So we have a lot of excess capacity that is not going to be able to produce significant more revenues and increase our margins. As Trevor referenced, we don't have to do a lot more in terms of additional investment to be able to drive a lot more revenue through those production facilities.

Mike Latimore: Great. And then just last on gross margin. It sound -- sort of sounds like you feel like gross margin probably improves by year-end? Or is that the takeaway?

Benjamin Wolff: No question. When you have these new start-up defense contracts where we're producing -- expecting to produce large volumes of particular components for aircraft and missile systems, one of the important milestones to unlock go-forward revenue is to produce a first article that gets evaluated for tolerance and precision, et cetera, and we have to get the government to approve that first article so then we can open up the gates on the high-volume manufacturing. That has been delayed on some of our key contracts. So when you look at our margins, as Trevor referenced, we have all of the costs incurred to develop that first article, but none of the revenue that associates with it.

So when you look at our overall margins across manufacturing, it looks depressed this quarter because of that investment in getting to first article.

Operator: The next question is coming from Greg Konrad of Jefferies.

Greg Konrad: Maybe just to go back to a couple of things that you talked about. Just on the SwarmOS, I mean, you talked about that being on 4 platforms and kind of laid out the crawl walk scenario for the next 2 years. Can you maybe talk about just kind of next steps? And if you can just remind us on when you think about like autonomy software, like what is the monetization? How do you get paid and how we think about that kind of going forward?

Benjamin Wolff: Sure. So we have -- I'm really delighted with the progress that we've had. When you and I have talked before, we've talked about the fact that our primary goal for '26 is to get different customers within the Pentagon to be able to understand that this technology exists that it works, that it's not just on PowerPoint and how differentiated it is from everything else that people talk about in terms of swarming and autonomy. I think we're hitting all the marks on that. And frankly, we're doing it even earlier in the year than I had expected us to do. So I think we're seeing great traction.

We're actually out on an exercise right now, where we are operating in a real battlefield environment. And soldiers are telling us that this should be the standard platform going forward for collaborative autonomy and swarming. So we're getting great feedback. It's going really well. In terms of what our business model looks like, our expectation and what we've talked with government customers about is that our software will cost the government about 10% of the overall UAV platform cost. So if we're talking about $40,000 drone, our cost will be $4,000. If we're talking about $1 million drone, the cost for our software will be $100,000. And if you're talking about a $4,000 drone, the cost will be $400.

That makes sense to the government, and it makes sense to us because when we talk about these larger, more exquisite drones that cost more, they have more sensor capabilities on them. They have longer duration in the air. They are far more capable. And the more capable the platform, the more value and utility our software brings to the battlefield. So it is really a value pricing proposition. It is a onetime upfront license fee. Most drones aren't expected to survive in the battlefield for longer than 1 year. So this is almost like a razor blade business in that we're continually selling more software on more drones.

They get used, they get expended and the government buys more of them.

Greg Konrad: And then just on the full year guidance, I mean, you called out some of the issues, including the government shutdown in Q1 and kind of how you expect that to ramp through the year. Thinking about like backlog and what's maybe not in backlog with expected awards, I mean, how do you think about visibility into year-end and some of the expected awards? How much is competitive versus just follow-ons and just kind of how you think about visibility for the rest of the year?

Benjamin Wolff: So when you take a look at the $7 million of new contracts in the first quarter, obviously, if we just did that every quarter and if all of the revenue was coming in on kind of a very scheduled basis, 4 quarters' time $7 million, you've got $28 million. We believe that every quarter will increase and that -- consistent with our internal plan, we knew first quarter was going to be lower. We expect second quarter to be larger than first and so on and so forth throughout the rest of the year as we continue to build the business. And we executed on

Trevor Thatcher: all 3 aspects of the business, the software side, the manufacturing side, the engineering services side and the drone hardware side included in that. So we believe that with what we have in backlog and with the go-gets that we have that are in the pipeline, we are highly confident at this point with where we -- with being able to hit that $24 million to $27 million guidance.

Greg Konrad: And then maybe just last one for me. I mean you called out the award with Portal. How are you thinking about the space opportunity in general? I mean, how is that emerging and just kind of how you're thinking about that going forward?

Benjamin Wolff: I think volume in space is going to be far more limited than what we look at terrestrially. But there's obviously much larger dollars on individual discrete efforts going into it. And so I think it is a very potentially large opportunity set for us, potentially lucrative. But to be candid, it is an area that we see as kind of a growth opportunity for us, but not something that we're putting anywhere near the kind of investment of effort and resources into the way we are on kind of terrestrial UAV efforts. So we are taking those opportunity sets and pursuing them on a more discrete basis.

I'd say that's more of a rifle-shot approach, whereas what we're doing with trying to get our software anywhere and everywhere that it can be relevant on UAV, that's more of a scatter gun or a shotgun approach. In both cases, I think we're applying the appropriate amount of resources to realize the opportunity. I think space can be big. I think it will be longer duration to get too big than it is near term the way we are with terrestrial drones.

Operator: Our next question is coming from Max Michaelis of Lake Street Capital Partners.

Maxwell Michaelis: First one, I just want to go back to Draganfly. So you guys finished up a few successful flight simulations in the quarter. I think you mentioned you're going to be moving on to live flight tests. Kind of help us out with sort of a timeline, what it looks like in 2026, when you guys are going to start these flight tests and kind of when this becomes more of a -- I guess, not meaningful partnership, but when does this start to kind of turn into some revenue?

Benjamin Wolff: So we're expecting to have some of our demonstrations on the Draganfly platform, I think, starting in this current quarter, in the second quarter. Certainly, I know we've got some plan for the third quarter where we're actually out with the government customers. I mean you know the drill with the defense contract. You have to show them that it works, that it exists. They then decide to go allocate dollars towards it, you negotiate a contract and you get an award. So I can't give you a prediction on when this becomes -- when that specific partnership results in revenue for us and Draganfly.

But what I can tell you is our partnership with them is very important because they are one of the core platforms that we've identified has a unique capability set that with our software on it can show increased value to the government customer. So it's an important partnership for us, and we expect to be getting that integration done in this quarter, the second quarter and doing demonstrations for the government as soon as that integration is done.

Maxwell Michaelis: Okay. And then last one for me. I don't think anybody has touched on sort of the commercial side of business with the IQ 2.0. I mean you talked about this customer that you're actively deploying with. I mean I'm assuming this initial deployment, you have a few systems in there, but does this customer have sort of the capacity to bring on -- to become a meaningful customer sometime in the future?

Benjamin Wolff: So the important thing about this customer is, yes, to answer your question, they have the ability to scale to more machines with our software on it once they become delighted with what they see from the first installation.

More importantly is this is the first time that we've had an active partnership with systems integrator, an indirect channel partner, if you will, and I mentioned on our prior call last quarter that one of the things we needed to figure out was how we create an attractive economic value proposition, both for ourselves, for our end customer, but also for the systems integrator because there -- as you know, there are some 1,800 systems integrators and maybe even more at this point across the United States. Getting them to be out selling for us was kind of a holy grail moment for us, and that happened in the first quarter.

So we've got this partnership now with the first systems integrator. We've got several other discussions underway where we've been able to figure out how we talk about this product and the economic opportunity associated with it in a way that works for both the systems integrator and for us and the end customer. So the reason that's important is systems integrators obviously deal with a lot of different customers, and they can be an indirect channel to get our message out there more broadly. So that's what we're really excited about. Yes, the first deployment is going very well. I mentioned in my prepared remarks that it is about surface preparation.

So think about doing things like sand blasting, sanding, grinding, those kinds of applications, even paint application on surfaces. Those are all the kinds of jobs that have historically required a human to do the job because of the variability associated with that kind of task. And we're showing the end user and the systems integrator how it can be done now with a robot on an automated basis using our IQ 2.0 software.

Operator: The next question is coming from Brian Kinstlinger of Alliance Global Partners.

Brian Kinstlinger: Can you quantify how much revenue was delayed due to the government shutdown? How would you characterize the procurement environment now? And then can you quantify TCV or bids outstanding in pipeline?

Benjamin Wolff: So, I can't quantify what was the amount associated with the delay because it's frankly difficult, Brian, to know had we gotten first articles approved, how much would have actually been taken based -- in the quarter based on when that approval had happened. So I can't really quantify that. Again, I'll tell you, reiterate that even with the government shutdown, we actually achieved what our internal projections called for. And so I'd ask you to take that plus the guidance that we've given and assume we're going to ramp over the remaining 3 quarters to achieve what we expect to achieve. In terms of the pipeline, I can't quantify pipeline.

I mean we obviously have internal numbers, but I don't want to throw that out there. The thing that we feel very solid about is backlog, which means it's contractually committed. I don't want to speculate beyond what's contractually committed.

Brian Kinstlinger: Well, in one of the prior questions, you talked about needing something like $7 million of bookings per quarter. Maybe talk about what the sales cycle generally you're seeing right now? Is it months? Is it things you've been bidding on for a very long time? Just kind of help us understand what that sales cycle looks like?

Benjamin Wolff: Yes, sure. So on the software side, we're seeing stuff that we had expected was a 12- to 18-month sales cycle, and we're seeing things now coming in, in less than 6 weeks. That might be an anomaly. It might just be the particular circumstances of those handful of different engagements that we've had. But we've landed some things that frankly surprised us with how quickly they came in. There are some other opportunities that we've been picking away at for 12 months now, and they still haven't come to fruition.

I think the thing that is encouraging to me, though, without kind of changing my own expectations about the sales cycle is that we are definitely, particularly on the defense side, seeing money come in faster than what we -- or at least contracts come in faster than what we would have expected 3, 4 months ago. So I think it's very bullish. It doesn't have to take as long as it has historically taken. And stay tuned. We'll hopefully be announcing some additional contract wins that have come in faster than what we would have expected.

Brian Kinstlinger: Great. Can you provide an update on the Red Cat partnership testing and integration? I think last quarter, it sounded like you were very close to signing a production agreement, but we didn't hear anything about that this -- on this call? I think I didn't.

Benjamin Wolff: So -- Yes. So we have a solid partnership with them. We have -- I think we inked the new expanded partnership agreement, and we are out doing demonstrations with their drones. In some instances, Brian, those demonstrations are being done jointly with the Red Cat team. In other cases, we take their drones out and we are demonstrating our software on their drones and their team isn't necessarily needed to be there. So -- but the Red Cat drones are kind of a cornerstone of the demonstrations that we're doing for government customers. Sometimes, as I said, it's in collaboration and sometimes it's just we've been invited to something and we go do it.

Brian Kinstlinger: So just to be clear, the economics are in place and the contracts in place for...

Benjamin Wolff: Yes.

Brian Kinstlinger: That determines your piece of the Red Cat sale and you can go-to-market now?

Benjamin Wolff: That's correct.

Brian Kinstlinger: Okay. Lastly, can you tell us how many shares did you sell in the ATM in the first quarter? It looks like you sold -- you raised $6.5 million. What was the average price of that? Or what were the shares?

Benjamin Wolff: Trevor, do you have that information?

Trevor Thatcher: Yes. We sold just under 890,000 shares. So if you do the math, it ends up being about $7.35 a share.

Operator: [Operator Instructions] We're showing no additional phone questions at this time.

Brian Siegel: We've got some questions from online. First one, there have been numerous expanded contracts with Air Force and other Department of War initiatives. These have been smaller scale thus far. How is the company positioning at this time to ramp production if a large purchase order is received?

Benjamin Wolff: So yes, I think that the correct way to characterize the contracts that we're seeing with DoD both directly and through primes is they start small and then we do everything we can to try and expand them. The key is getting those first contracts inked and demonstrating what we can do and then having it grow from there. The question about ability to scale, I'll take that in 2 different parts. One is on the software side. Software is easy to scale. We've got the code. The product is locked, and it's just about pushing software onto whatever hardware platform is being used, and we can do that quickly.

So there's a lot of opportunity to scale the software side. Responding to the manufacturing side of the business, I think in response to a question I was asked earlier, we're only at about 30% capacity on our manufacturing facilities. So there's a lot of room to scale there. And we rely on a lot of automation and advanced technologies in that business. So it does not require a lot of ramp-up of human personnel to be able to leverage that capacity that we have, that isn't being used at this moment. So I think we've got a ton of capacity without additional costs that gets incurred to be able to ramp up revenues significantly.

So we feel like we're in very good shape on that front.

Brian Siegel: Next question for IQ 2.0, can you help us understand what's behind the customer and the land-and-expand opportunity? And then in general, how many qualified opportunities or kind of what does the pipeline sit at today for this product? And how are some of those conversations progressing?

Benjamin Wolff: So the -- I think I mentioned before that the surface prepped use case is what this first customer and the systems integrator is focused on. There are -- there is a massive amount of market need for automating that kind of task, whether you're talking about stripping corrosion or paint off of a part, whether you're talking about applying new surface treatment to the part to be ready for delivery to an end customer, whether it's painting or other types of prep. This is historically an area that has been very labor-intensive. And what we're seeing is that, that is a greenfield opportunity for automation, and just a tremendous amount of interest and therefore, demand in what we're doing.

The pipeline, not backlog, but pipeline is filled with dozens of conversations that we are having. But again, we believe that on the IQ side of our house, historically, it has been a 12- to 18-month sales process. If you want to say historically, that's what we've historically thought it would be. The one that we're deploying now, I think, came together in about 8 weeks. So that's on the short side. But if our 12- to 18-month expectation is correct, and we just launched IQ 2.0 at the beginning of this year, we got some ground to cover. So our expectations are modest for '26. That's part of the crawl, walk, run approach that we've talked about.

But we could be pleasantly surprised. I hope we're pleasantly surprised, and we'll see more like this first one that come in, in shorter than the 12- to 18-month time frame.

Brian Siegel: Next question, and this kind of relates more to some of the things you said in the white papers. Recently, there was a Bloomberg article stating that Google is dropping out of the $100 million Pentagon prize challenge to create tech for voice controlled autonomous drone swarms. The article says OpenAI, Palantir and xAI are still competing. There's no mention of Palladyne for SwarmOS. Does this mean others have equally as good swarming technology and don't have to use Palladyne? Or should we be expecting some future growth as those companies need our SwarmOS?

Benjamin Wolff: Yes, I think it's the latter. So let's break this down. The ability to give voice commands to a drone, while that certainly is easier than using joysticks, it is still a one soldier to one drone operating environment, and there's nothing collaborative about it between drones and there's nothing swarming about it. It certainly makes giving direction and manually managing multiple drones easier if you can just have a soldier say, do X, Y or Z without having to have your fingers and thumbs on joysticks and controllers. But that is -- that's like a Band-Aid on the problem.

And so the great part about that program is the Pentagon is saying, "Hey, we want to put money and resources into trying to make this ease of operation a real focus of ours." They want to lighten the cognitive load on the operator. That's all great tailwinds for us because we have the ultimate solution to that. It's not about -- I mean, whether you're typing in a command or giving a verbal command, there's certainly some efficiency there. But that's just like the -- that's step one.

We're at step 5 or 10 already where we are able to very, very easily with a very small amount of input, get a whole swarm of drones working collaboratively to achieve an objective or a mission. So I think back to the question, we are the endpoint that, that program ultimately wants to get to, and they're kind of looking at stop-gap measures. So that's very exciting to us.

Brian Siegel: So that concludes the question-and-answer session from online. Operator, we can close out.

Operator: Thank you. Ladies and gentlemen, this brings us to the end of today's teleconference. We would like to thank you for your participation and interest in Palladyne AI. You may disconnect your lines or log off the webcast at this time, and enjoy the rest of your day.