Good morning, and welcome to the Serve Robotics fourth quarter 2025 financial results and conference call. I am Franz, and I'll be the operator assisting you today. All lines have been placed on mute to prevent any background noise. After the speaker's remarks, there will be a question and answer session. If you would like to ask a question during this time, simply press star one on your telephone keypad. If you would like to withdraw your question, please press star one again. Thank you. I would now like to turn the call over to Steve Webb. Please go ahead. Thank you, operator, and good afternoon, everyone. I'm Steve Webb, Serve's SVP of Marketing and Communications. Welcome to Serve Robotics' fourth quarter and full year 2025 earnings call. With me today are Serve's co-founder and CEO, Ali Kashani, and our CFO, Brian Read. During today's call, we may present both GAAP and non-GAAP financial measures. If needed, a reconciliation of GAAP to non-GAAP measures can be found in our earnings release filed earlier today. Certain statements in this call are forward-looking statements. You should not place undue reliance on forward-looking statements. Actual results may differ materially from these forward-looking statements. We do not want to undertake any obligation to update any forward-looking statements we make today, except as required by law. For more information about factors that may cause actual results to differ materially from forward-looking statements, please refer to the press release we issued today, as well as the risks and uncertainties described in our most recent annual report on Form 10-K and in other filings made with the SEC. We published our quarterly financial press release and our updated corporate presentation to our investor relations website earlier this morning, and we ask you to review those documents if you haven't already. With that, let me hand it over to Ali. Thanks, Steve. Good morning, everyone, and thank you all for joining us. A year ago, we told you that we would deploy 2,000 autonomous robots across the country by the end of 2025, that we would expand from a single city to a truly national footprint, and we would prove that this technology works, not just in a lab or a closed campus, but on open sidewalks in dense cities, navigating the full complexity of urban life. We did all of that and then some. Today, a fleet of 2,000 Serve robots have been activated across 20 distinct cities in six major metropolitan areas, from Los Angeles all the way to the Washington, D.C., corridor. We launched Atlanta, Dallas, Chicago, and Miami. We expanded aggressively in every existing market. We also added DoorDash alongside Uber Eats. This gives us access to over 80% of the U.S. food delivery market. We also completed four strategic acquisitions since early 2025. We met or exceeded our revenue guidance every single quarter, and through all of it, we maintained a 99.8% delivery completion rate and a proud safety record. Let me say it again. 20 times the fleet, national scale, four acquisitions, and near-perfect reliability. In Q4, we once again delivered revenue above guidance as we drove 400% year-over-year growth in the quarter. This is not incremental progress. This is a company that's defining a category in real time. Before I get into the quarter, let's look at the broader trends. We are living through one of the most consequential technology transitions of our lifetime. For the past few years, the world has marveled at what AI can do with words and images and code. The next frontier, the one that will reshape our physical world, is physical AI, machines that can see and think and act in real environments alongside people. As we try to anticipate what this future looks like, I find it really helpful to think about the evolution of computing so far. First, it was the personal computers. Then came the internet. It connected information. It connected people. Next, we put computing and connectivity in every pocket and in every device. We connected the physical things too. As a result of all this, all of the commerce and every industry is now digital and connected. Each leap along this path was worth trillions. Fast-forward to today, AI has taken over our digital lives over the past few years, arguably becoming the fastest rising ring of this evolution of computing. Physical AI is the natural next phase that's right around the corner. It's the moment when this intelligence leaves the digital realm and enters the streets. Like computing and the internet before it, the companies that build the platforms for physical AI will define this era. NVIDIA CEO Jensen Huang has called robotics and physical AI the next multi-trillion dollar industry. Every major AI company is racing to build models for the physical world. The investment thesis is pretty clear. The companies that build the platform and own the data will capture the value. Here's what's important. You can't build Physical AI from a research lab. You need robots in the real world gathering real data, encountering real edge cases, and at real scale. That's the flywheel, and it's exactly what Serve has built. Every transformative technology goes through the same arc. We are at a very familiar inflection point. Autonomous robots are here to fundamentally shift how we leverage technology in our lives. The question is no longer will this work, as we've seen by our progress last year. Now the question is: how fast can you scale? 2025 was the year we proved the technology. Looking ahead, 2026 is the year we compound the business model. Last quarter, I said that beyond 1,000 robots, the system tips. Scale changes everything. The economics improve, the partners lean in, learning accelerates. At 2,000 robots, the system doesn't just tip, it compounds. We are now accelerating the flywheel. We discussed this concept last quarter when we described how more miles lead to more data, better models, and a more capable fleet. This is the flywheel that should be at the core of any Physical AI company, and we have really organized our strategy around it. Every investment we make, every acquisition or deployment or partnership, they're all designed to strengthen a specific step of that flywheel, and as a result, make the whole system spin faster. Let me walk you through it. There are four steps in the Serve flywheel. Step one is amassing data. Physical AI runs on large amounts of data. This isn't just some data you scrape off the internet. This is data collected in real environments. It's collected by robots at scale. Every mile our robots travel enriches our data set. Every edge case, every construction zone or rush hour or unmarked crosswalk, they all sharpen our models. This data is proprietary. You can't just download it on the internet or simulate it with the same depth and richness. You have to live on the sidewalks and no one is better positioned for it than Serve. What's new and exciting is that we are no longer just collecting data from a single environment. Today, our data spans multiple and distinct physical domains. On sidewalks, thousands of robots are mapping the world in 20 unique cities across the country. Every new neighborhood brings new edge cases and new pedestrian and traffic dynamics, new weather patterns. All of that enriches the models network-wide. In hospitals, our recent acquisition, Diligent Robotics, has a fleet of nearly 100 robots called Moxi, and they are navigating some of the most challenging indoor environments in robotics. These are multi-level facilities with tight corridors, constant foot traffic, high-pressure operations. Moxi robots have completed over 1 million deliveries across more than 25 hospitals and counting. Sidewalks and hospitals and beyond, multiple domains with wide-ranging geographies, all feeding a single robotics and autonomy platform. There is no one who is doing all this and realizing the value of the combination of indoor and outdoor data collection from commercial scale fleets. The second step of the Serve flywheel is the models. Data is a raw material, but step two is where we take everything our robots are seeing and experiencing, and we turn them into better AI models. This is where another recent acquisition comes into focus. Vayu Robotics brought us a specialized team that builds end-to-end models for Physical AI. We are building systems that empower us to train across all our operating domains, indoors and outdoors, so that what a robot learns in Los Angeles would help a robot in Dallas, or what a Moxi robot learns navigating a hospital corridor could improve a Serve robot that's navigating an obstructed city sidewalk. That kind of cross-domain learning is really significant, and it's a compounding advantage that will widen every quarter. Also, our acquisition of Phantom Auto bought us one of the most capable robot connectivity stacks with extremely low latency. This enables us to operate at a large scale and across a significant geographic region because we can reliably assist robots remotely in real time. What's underappreciated here is that every time a remote supervisor assists a robot anywhere in the country, we generate high-quality training data. Our operations, which is empowered by this connectivity stack we acquired, are a conduit to collecting more data and more edge cases, and it's paired with a considerable training data set, all collected at a faster rate than ever, feeding right back into our models. I should also mention the talented team of engineers that make all of this possible. When you have one of the largest autonomous robot fleets, plus data from multiple physical domains and the infrastructure to turn all of this into deployed AI, that's where the best people wanna work. Retention across our team has been really strong because people love building on real robots in the real world with significant unique data. The flywheel attracts talent, and talent accelerates the flywheel. The third step of the Serve flywheel, after you gather the data and develop the models, is to deploy those models into the real world. Better models only matters if you can actually get them onto live robots. That is pushing all that improved autonomy out to the fleet that's in the real world where the edge cases live. This is where our fleet scale and our partnerships become a strategic asset. Uber Eats and DoorDash combined serve over 80% of the U.S. food delivery market. We are now a multi-platform fleet. We see robots finishing a DoorDash delivery, then picking up an Uber Eats order on the way back. That kind of interoperability drives utilization, and of course, utilization is the key to both our economics and our data collection. Our merchant network has expanded to over 4,500 available restaurants and retail partners today. Just this morning, we announced a new partnership with White Castle, one of America's most iconic restaurant brands. Our geographic pipeline also continues to develop. We are in active discussions with city officials across the country, from New York to Boston to San Jose, and even internationally, Vancouver and Toronto and Sydney and Melbourne. As we evaluate all these new wave of market launches, each market will represent a natural extension to our existing footprint, and we are really excited to share more about our plans throughout 2026 as these initiatives progress. This is a critical point. Every deployment across every domain into every new city, it generates new unique data that feeds directly back into step one, and the cycle continues. Finally, the fourth step of the Serve flywheel is monetization. This is the step that makes the whole flywheel self-sustaining. When you monetize your fleet, you fund the next turn of the cycle and make the flywheel accelerate much faster. The companies that figure this out early and can get paid to collect their proprietary data, they have a real advantage over those who have to pay for their data. Tesla is the obvious example. They collect massive amounts of road data to train their models by simply selling cars to consumers. One way we are really advancing our monetization is by increasing our revenue sources rapidly. Delivery fees are, of course, our core business. It's continuing to accelerate as we scale geographies. Branding and advertising saw a 50% increase in Q4 year-over-year. With 2,000 robots moving through high-density neighborhoods, we have effectively built a neighborhood-level media network on wheels. Advertisers' response has been exceptional, and we are building a robust bookings pipeline. Over time, we believe advertising and branding can represent as much as 50% of our fleet revenues. Think about what that means. It monetizes miles that are already being driven at nearly zero marginal cost. Also, data and platform revenues are emerging. In 2026, we plan to further invest in our data and platform capabilities to strengthen the foundation of our robotic solution offering. By offering the platform that powers our deployed robots to external partners and other robot operators, we expect this new revenue base to mature and become a meaningful high-margin contributor. Also, going forward, healthcare revenue from Diligent Robotics will be another meaningful contributor. Nearly 100 Moxi robots across over 25 hospital facilities, with each facility generating over $200,000 in annual revenue. This is already a fully functional business unit that's generating both meaningful data and meaningful revenues. Here's what ties everything together. Every dollar of revenue funds more robots, which leads to more data, which helps us create better models, which leads to even more deployments and more revenues, and the cycle repeats. The monetization doesn't just sustain the flywheel, it accelerates it. I think that our acquisition strategy also deserves a moment of its own. We've completed four acquisitions in the last 12 months. Every acquisition we've made maps directly to a step or two of the Serve flywheel. Phantom Auto strengthens our data collection and our deployment scale as well. Vayu Robotics strengthens our models creation. Diligent Robotics further strengthens our data gathering by introducing a new operating domain, and also it boosts our monetization through recurring revenues with compelling economics. Last but not least, Weebo strengthens our delivery robot monetization by boosting our partnerships with restaurants and major QSRs. This is all deliberate. It's a flywheel-driven strategy. Each deal is designed to make the flywheel stronger. Now let me bring this back to our 2025 progress and specifically our Q4 results. In Q4, we exceeded our revenue guidance once again. Total revenue for the fourth quarter was $0.9 million, representing nearly 400% growth year-over-year and also meaningful sequential acceleration. Full-year 2025 revenue came in above our $2.5 million guidance at $2.7 million. We completed the deployment of our 2,000th robot in mid-December on time and on plan. In Q4 alone, we deployed nearly 1,000 robots. That's in a single quarter. That is more than many robotic companies' entire fleet size. Delivery volume grew 53% quarter-over-quarter in Q4 and roughly 270% for the full year versus 2024. This is a compounding effect of fleet at scale. Also, it's the geographic expansion and the deepening platform partnerships, all of which are working in concert. We expect this growth to continue as we start to see the benefits of new robots and also optimize their operations and utilization. Our merchant base has also expanded to over 4,500 restaurants and retail partners today. This is a more than 10x increase from roughly 400 a year ago. We now reach over 1.7 million households in our metro areas. This covers a population of over 3.75 million people. We did all of this while maintaining our 99.8% delivery reliability and also our strong safety record. This is the part that I'm most proud of. Scaling fast is hard, but scaling fast while maintaining quality and safety is what really separates us. Okay, I want to close with where all of this leads to. A year ago, we had roughly 100 robots. Today we have 2,000. The path is clear from here to 10,000 robots and well beyond. This would be across more cities, more verticals, even internationally. We have the engineering and operations roadmap and also a track record of execution. The hardest part, building the platform and proving the technology, earning the trust of our partners and cities and consumers, these are all tailwinds now. What excites me most is that each additional robot we deploy would make the entire system more valuable. The data gets richer, the models get sharper, the economics improve, the partnerships deepen. This is the nature of a platform business with a flywheel at the core. We are just entering that phase where the compounding effects and the acceleration of the flywheel become visible. With the Diligent Robotics acquisition, we have extended this platform beyond the sidewalk and into hospitals. That's not a one-off. It's a signal of where things are heading. The robotics platform we are building will be general enough to operate where very intelligent machines are needed to move safely among people and mature enough to deliver real commercial value right away. We are not building a delivery company. We are building the operating layer for how robots integrate into our lives. That's the long game, and we are playing it from a position of strength. 2025 was the year of proof. 2026 is the year of compounding returns. I have never been more energized about what's ahead. With that, let me hand it over back to Brian. Thank you, Ali. Good morning, everyone. Entering 2025, we set explicit operating targets around fleet expansion, revenue growth, and geographic scale, and we delivered against each one of them. More importantly, we strengthened the economic foundation of our business while doing so. That operating discipline will continue to define Serve into 2026. Let's walk through the details. Total revenue for Q4 2025 increased over 400% year-over-year to $0.9 million. Full year 2025 revenue was $2.7 million, exceeding our guidance of $2.5 million and representing growth of 46% over the prior year. Fleet revenue was $0.7 million for the quarter, growing 50% sequentially. Branding saw record bookings during the quarter as our expanded fleet attracted larger advertising commitments. We also recorded our first revenues related to data monetization in the quarter, an early signal of the data and platform opportunity ahead. As Ali and I have mentioned, these opportunities will continue to evolve through 2026. Software revenues were over $200,000 in the quarter. Our transition to recurring software revenue continues to progress, with our recurring software base now representing approximately 70% of software revenues. More broadly than software, we noticed a shift in revenue quality during the year. Our underlying recurring revenues, defined as revenue excluding one-time agreements, grew over 3x during the year. That shift increases revenue visibility while reducing volatility as we scale. Beneath the top line, Q4 margins reflect the largest single quarter deployment in our history with nearly 1,000 new robots. When deployments occur at this scale, newly introduced cohorts initially operate below steady-state efficiency. That's expected and by design. What matters is the trajectory as that fleet matures. This past year, we observed average daily operating hours per robot climb 56% to over 12 hours compared to Q4 last year. Cost per delivery trended down quarter-over-quarter during the year as our operations team gained experience and our systems continued to mature. Collectively, along with other metrics, these trends give us confidence in continued margin improvement moving into 2026. As reflected in our 2025 results, the operational infrastructure required to support our larger fleet was established this past year. We expanded market operations, built fleet maintenance capabilities, remote supervision systems, and deployment capacity ahead of 2026 revenue, and of course, the achievement of our 2,000 robot deployment milestone. As we move through 2026, we expect a growing portion of that infrastructure to be absorbed across a larger and more productive fleet. GAAP operating expenses for Q4 were $34.3 million, reflecting the cost of deploying nearly 1,000 new robots and expanded operational capacity across new cities within Alexandria, Virginia, and Fort Lauderdale, Florida. On a non-GAAP basis, excluding stock-based compensation of $6.3 million, operating expenses were $25.2 million. R&D remains our largest investment area at $15.9 million on a GAAP basis or $12.8 million on a non-GAAP basis. This is directed towards advancing our AI stack, integrating capabilities from the Vayu and Phantom Auto acquisitions, and building the data infrastructure for our growing fleet. G&A spending stayed lean and purposeful, decreasing from the prior quarter by $2 million to $11.1 million on a GAAP basis, and $6 million on a non-GAAP basis. We expanded to one new metro area and nine new cities during Q4 and anticipate a flattening of our G&A expense growth even as we continue to scale through 2026. As I mentioned, we will continue to manage operating expenses with discipline, aligning investment with measurable deployment milestones. Interest income generated in the quarter was nearly $2 million. Additionally, Q4 reflects a $3.8 million tax benefit related to deferred tax liabilities associated with the Vayu acquisition, resulting in a partial release of our valuation allowance. Turning to the balance sheet. We closed the year with $260 million in cash and marketable securities. Capital expenditures for the quarter were $16.5 million, representing the tail end of our costs for the 2,000-unit build. Our liquidity position provides strategic flexibility in a capital-intensive industry where balance sheet strength is a competitive advantage. We continue to evaluate additional funding opportunities opportunistically. Adjusted EBITDA was negative $28 million. As revenue scales and per unit economics improve, we expect sequential improvement in adjusted EBITDA margins throughout 2026. Turning to our outlook. Today, we are raising 2026 revenue guidance to approximately $26 million. The improved outlook is primarily driven by the acquisition of Diligent Robotics, which we believe represents a high return use of capital while broadening our platform, expanding our addressable market, and increasing the proportion of revenue derived from durable recurring contracts. To fund that acquisition, we moderated our planned 2026 capital expenditures. As a result, we redirected a portion of planned near term fleet investment toward a significant new market opportunity that is expected to contribute roughly $7 million of revenue during 2026, primarily through recurring healthcare contracts. Looking beyond 2026, we continue to expect this newly combined core business to deliver sustained accelerating growth. We have previously discussed a $60 million-$80 million of annualized revenue run rate associated with that full utilization of our fleet. Internally, we view that level less as an endpoint and more as an intermediate milestone as our business continues to scale exponentially. Our growth is expected to be driven in part by disciplined geographic expansion. As Ali touched on earlier, we are in productive discussions to extend our footprint across additional U.S. markets and over time, pursue selective expansion into major international cities like Toronto, Sydney, Tokyo, Madrid, and London, among many others. We expect 2026 capital expenditures of approximately $25 million associated with the production and deployment of additional robots as we continue expanding the fleet and increasing volume of real-world operating data that strengthens the flywheel. Recent acquisitions are expected to increase our 2026 operating base by approximately $20-$30 million. Non-GAAP operating expenses in 2026 are expected to be approximately $160 million-$170 million, reflecting continued investment in autonomy development, fleet scale, platform capabilities across both delivery and healthcare robotics. Let me close with this. The investments we are making in 2026 are specifically designed to strengthen the plan Ali described. We are expanding the fleet, improving the autonomy stack, and increasing monetization opportunities across the platform as the flywheel accelerates. Serve has evolved into a diversified robotics platform with multiple revenue streams spanning delivery, advertising, data services, software, and now healthcare automation. In the age of Physical AI, we're using our strength in autonomous robotic delivery to build a generational robotics company that would define this era. With that, I'll hand it back to Steve for Q&A. Thank you, Ali and Brian. We'll now move into the Q&A session, but first I'd like to say a big thank you to all the investors and analysts who submitted questions via email. Thank you so much for your engagement. The first question we have is related to new robots. Serve deployed 2,000 robots last year. What's the goal from a unit deployment perspective in 2026 and beyond that? Thank you. I'm happy to take this one. Over the next few years, we expect to deploy thousands more robots. In the short term, as we've shared in the past, before we go on and share a detailed plan, we want to really let the recent growth settle in, and we wanna gather all the data and learnings from last year's 20x fleet growth. We have the capacity to continue growing our revenue right now. On the other hand, manufacturing and supply chain, as we all know, they require certain lead time. We are already working on the supply chain for the next batch of robots so that we can expand to new major markets as they become available quickly. The time between now and when the the supply chain and manufacturing of the robots would be available, it's a good time for us to really hone in on our playbooks and get them refined based on the existing growth. We don't really want to be deploying more robots until we get all the current ones fully activated on a daily basis. Yeah, if I can wrap up on that, Ali. In the prepared remarks, right, we talked about CapEx guidance being approximately $25 million during 2026. A significant majority of that will be for the Serve fleet expansion. We're going to continue to invest, not only in Serve, but for additional Moxi robots and look to accelerate their growth as well. I think, Ali, exactly as you summed up, Q1 in this time period, Q1 2026, we're looking to optimize the performance of the full fleet, and most importantly, we retain control over that CapEx timing, and also the OpEx deployment costs as that fleet continues to grow. Great. Thanks, Brian. On to our next question. What percentage of the 2,000 deployed robots should be daily active by the end of first quarter? I'm happy to take this one as well. From manufacturing and deploying robots to reaching full utilization of the fleet, as we've discussed this in the past, there are several steps that take place. You start with obviously creating the depots in each new market, building the operational footprint, which includes hiring and training staff. This is a lot of the work we've already done. The next step after that is getting any requirements by local municipalities, any stage gates, all of that addressed. We need to activate neighborhoods with our delivery partners, onboard local merchants, and then once all of that is done, we can have robots at full operational hour every day. We would focus on operational efficiency. It's a question of where to put the robots, how to move them around, all of that, so that we capture the maximum demand. We expect that by the middle of this year, as I said before, if we have any, additional robots to manufacture, we would get all of the existing robots on a fully active, daily basis, and shift our focus to that operational optimization. We are timing everything again so that we have the full activation of these robots before manufacturing new ones, given the lead times for manufacturing. Thanks, Ali. On to the next question. We received this one about the acquisition of Diligent Robotics. How are the integration efforts going, and what are your plans for growing the healthcare business? That's a great question. We covered some of this earlier, but I can dig in a bit more. We have always intended for our autonomy platform to extend beyond just food delivery and into many other environments, including, in this case, hospital and healthcare. As we looked at Diligent Robotics during our acquisition process. It became pretty clear very quickly that it's the right time and right company for us to expand our scope. This acquisition actually strengthens our flywheel, as I mentioned earlier, by really enriching our data further. It also creates a more balanced and resilient revenue base for us, and it opens up, obviously, new markets, opportunities and a new growth engine. We are already starting to integrate our platform capabilities with Moxi robots, but this will take some time. As we do this integration work, we are creating a repeatable playbook for expanding into new verticals and operating in multiple different domains. On the second part of that question for the revenue, and just to give a little bit more color, right? These are existing, established, recurring revenue contracts that we were able to acquire through Diligent. These are different than our demand cycle for current food delivery. The $7 million number we referenced in the pre-prepared reports is for revenue here in 2026. I think it's important, right from an integration standpoint, we're gonna continue to focus on additional investments into the healthcare business around engineering headcount, infrastructure to support that team through their next phase of growth. Our business development team, sales teams are looking at other opportunities in the pipeline. Several of those are currently being evaluated, and we're gonna make the best decisions to drive long-term revenue growth. Great. On to the next question. Is optimization of the fleet a linear process, or are there step functions? And if so, what would cause that? Yeah. You know, we touched on the steps earlier. Of course, we are pushing a lot of these steps at the same time, but you're never going to get everything done at the same time. I think going from that deployment to full utilization steps are pretty important. There are many factors that affect that utilization and those steps kinda outline, as I mentioned earlier. Overall, though, we are seeing that our more mature markets are further along on that optimization curve. We mentioned this earlier, but 2026 is really about compounding returns for us. 2025 was all about building that infrastructure. In 2026 we are going to be really laser-focused on optimization and efficiency of the fleets, both on the sidewalk and on the hospitals. We have enough time for one more question. Can you speak more about your plans to expand internationally? What's the timeframe for those city launches? Yeah, that's an exciting one to end on. Let me maybe give some context on our thinking here. We have really built a great foundation for expansion. We are now in 20 cities, six major metros. We have really proven the tech at scale, built the operational playbook, the way to launch new markets efficiently. This work really supports that international expansion well. We are now in active discussions with city officials and partners in multiple international markets, from Canada to Australia, Japan, Spain, the many other countries. We are considering major cities, dense urban environments, strong delivery markets and municipal governments that are really leaning into autonomous robots on sidewalks. I wanna emphasize that we are going to be disciplined and intentional about these expansions, especially weighing our growth opportunities here in the U.S. versus markets abroad. We have learned from our U.S. expansions to date that the right way to go to a new market is methodical, and we wanna really be measured as we identify the right partners and the right expansion cities. We do get a ton of inbound interest to consider, but we wanna be very selective. We see this ultimate growth opportunity internationally as a 2027 opportunity. 2026 is for us to lay the groundwork for it, just as we laid the groundwork for this year last year by expanding to new cities. In the meantime, our robots obviously will continue and collect more data in, you know, 20 cities today and expanding by the end of the year. We'll keep making that Serve fly. We'll move faster and become more durable so that we can enable even further rapid growth and expansion. I'll just end by saying this again, I've never been more energized and excited about what's ahead for Serve, and I can't wait to see Serve robots operating in cities across the globe. Great. Thanks, Ali, and thanks, Brian. That's all the time we have for today. I'd like to thank everyone for joining us again. Thank you for joining us on the call today. Ladies and gentlemen, thank you all for joining, and that concludes today's conference call. All participants may now disconnect.