Serve Robotics - Earnings Call - Q4 2024

Executive Summary

• Q4 revenue was $175,842, down sequentially vs Q3 ($221,555) but up sharply YoY vs Q4 2023 ($42,719); FY 2024 revenue reached $1,812,483 (+773% YoY), driven by software services and expanding delivery/branding revenue.
• GAAP net loss in Q4 was $(13,119,495) with GAAP EPS of $(0.36); non-GAAP net loss was $(8,495,049) and adjusted EBITDA was $(8,679,137), reflecting higher R&D and G&A as SERV ramps its 2,000-robot plan.
• Management reaffirmed plans to deploy 2,000 robots by year-end 2025, accelerated Miami launch (Q1), added Atlanta (Q2), and expects 250 Gen3 robots fully deployed by end of Q2, followed by ~700 lower-cost units in Q3 and the remainder in Q4; Gen3 cost down ~65% vs Gen2 following an additional 30% reduction post-year-end.
• Liquidity strengthened: $123.3M cash at 12/31/24 and debt-free; additional $80M raised post-year-end; CFO now expects to avoid equipment financing, saving ~$20M over two years; ~57M shares outstanding as of Mar 6, 2025.
• Potential stock catalysts: accelerated multi-city deployments (Miami, Dallas, Atlanta), third-gen robot cost reductions, strong cash runway and PwC auditor transition signaling governance upgrades; multi-modal Wing partnership could extend delivery radius and monetization opportunities.

What Went Well and What Went Wrong

• What Went Well

  • “Transformational year”: revenue +~700% YoY to $1.8M; reach expanded to 1,000+ restaurants and 300,000+ households; first 75 Gen3 robots delivered ahead of schedule.
  • Fleet and operations scaling: daily supply hours +94% YoY and daily active robots +81% YoY; Q4 delivery volume +20% QoQ despite late-quarter fleet expansion, driven by operational optimization.
  • Cost-down and manufacturing: completed Gen3 design with faster, longer-range robots and 5x compute at ~50% of prior manufacturing cost; an additional ~30% cost reduction post-year-end (total ~65% vs Gen2) and Magna production ramp.

• What Went Wrong

  • Sequential revenue decline: Q4 total revenue ($175,842) fell vs Q3 ($221,555) primarily due to reduced software services in Q4 relative to Q2/Q3 and a higher fixed cost base impacting COGS during fleet ramp preparation.
  • Elevated operating expenses: Q4 GAAP opex increased to $12,920,693 (vs $8,289,324 in Q3), with R&D and G&A up as SERV invested in fleet scaling, software development, and corporate infrastructure.
  • Profitability still distant: Q4 adjusted EBITDA of $(8,679,137) and non-GAAP net loss of $(8,495,049), reflecting early-stage scaling; management emphasized non-recurring nature of software revenue and utilization ramp required for margin improvement.

Transcript

Operator (participant)

Thank you for standing by. My name is Danielle, and I will be your conference operator today. At this time, I would like to welcome everyone to the fourth quarter and full year 2024 Result Conference Call. 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, followed by the number one on your telephone keypad. If you would like to withdraw your question, press Star one again. Thank you. I would now like to turn the call over to Aduke Thelwell. Please go ahead.

Aduke Thelwell (Head of Investor Relations)

Thank you, Operator, and good afternoon, everyone. Welcome to Serve Robotics' fourth quarter and full year 2024 Conference Call. With me today are Serve CEO and co-founder 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 risks may differ materially from these forward-looking statements, and we do not 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 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 website earlier this afternoon. We ask you to review those documents if you haven't already. With that, let me hand it over to Ali.

Ali Kashani (Co-Founder and CEO)

Thank you, Aduke. We got a lot to cover today. I'm going to start by giving you an update about the past quarter and put that in the context of our progress over the year 2024. We will talk about the year ahead, give you some additional details about our plans to deploy 2,000 robots by the end of 2025, which remains on track. Lastly, I want to talk a little bit about the recent progress in AI and how Serve fits into that. Let me start with five quick highlights before we dive into each one of them. First, revenue. We wrapped up the year with $1.8 million in revenue, which is a 700% increase year over year. Second, our reach.

We increased the number of restaurants we served by 3X last year and also increased the households we reach by over 2X just in the last four months. We did all this by expanding to new neighborhoods in Los Angeles and expanding to our first East Coast city, which is Miami. Third, the fleet. We are about to complete building 250 3rd-gen robots to add to our fleet. Fourth, an engineering update. We have designed further cost reductions into our 3rd generation robots, effectively cutting the cost of building robots by two-thirds in just a year. Last but not least, our balance sheet. We massively improved our cash position, became debt-free, and increased our capital efficiency by saving $20 million in future capital costs, which Brian will discuss later. Now let's dive in.

As you can see, we've been growing fast on all fronts, expanding into new neighborhoods and cities, onboarding more restaurants, scaling our fleet. In Los Angeles, we launched in Downtown LA, Sawtelle, and Westwood just in November, instantly increasing our household reach by 50% and tripling the number of restaurants we served compared to the start of the year. We also didn't stop there. Since January, we added Glendale and Long Beach. Plus, we also expanded beyond California for the first time to Miami. Today, we support over 1,000 restaurants and reach over 300,000 households. That's a 2X expansion in market reach in just a few months. We are continuing to work with our delivery partners and local merchants and municipalities to unlock even more neighborhoods and cities throughout the year.

On the fleet side, in Q4, we completed the design of our 3rd generation robots and even manufactured 75 units in the final days of December ahead of schedule. These are the first production batch of third-generation robots, and we've been ramping them up into the delivery fleet since the new year. We expect their impact on delivery volume to really start showing in Q1 and Q2. Despite the fact that our delivery fleet didn't expand until the very end of Q4, our Gen 2 robots have actually delivered 20% more orders quarter over quarter. This is primarily because of optimizing our operation and expanding our reach. This is true even though we actually had to use some of the existing fleet to support the expansion activities, like sending robots to new neighborhoods and cities to map and collect data.

We were actually reducing our delivery fleet size, and yet we achieved a 20% increase in our delivery output quarter over quarter. If you look at our daily active robots and daily supply hours, the numbers have remained roughly the same as Q3 since we did not expand the fleet until the very end of Q4 last year. We had 57 daily active robots in Q4, up 81% from 29 robots in Q4 of 2023, and comparable to Q3's 59 robots, the slight difference here being because we were using some of our robots for expansion activities rather than deliveries. Similarly, we had 455 daily supply hours, which was in line with 465 hours from Q3 and nearly double Q4 2023. Let's take a step back and look at the entire 2024, which was a truly transformational year for us.

We listed on NASDAQ, raised over $250 million from the beginning of 2024 to today to really strengthen our balance sheet. We partnered with Magna International to scale our fleet manufacturing reliably, and we also formed new partnerships, including Shake Shack and Wing. We doubled our supply hours and active robot capacity and volume. We completed our Gen 3 robot design, which can move twice as fast, go twice as far, have five times more brain power, and cost about one-third of our previous generation robots. We built the first batch of those Gen 3 robots ahead of schedule at the end of last year. Now, behind the scene, there was even more going on. We have actually boosted the speed of our robots by 30% last year, and we cut our missed delivery deadline rate by 75%.

We did that all the while doubling our delivery volume. I'm proud of what our team has accomplished over the last year, and I think it's positioning us really well to achieve what we are set out to do in 2025, which is deploy 2,000 of our robots. We are really entering this year with strong momentum. Let's talk about 2025, starting with the first half of the year. We've started by accelerating our expansion timeline by entering Miami ahead of schedule in Q1. Dallas remains on track for Q2, which is what we promised. I can announce that we have a plan to launch a third market in Q2. Let's go through this one by one. First, Miami. In February, we launched in two neighborhoods, Brickell and Miami Beach, in collaboration with Uber Eats. We are now serving 50 local restaurants already.

This includes two important brands who signed with us right at launch: our existing partner, Shake Shack, and our new partner, the beloved local pizzeria, Mister O1. This news may come as a bit of a surprise since we had not shared any plans to launch Miami in Q1, but we did see a demand from our partners and decided to take the opportunity and move quickly. Next up is Dallas. We are on track to launch in Q2, bringing our robots to neighborhoods like Uptown and St. Thomas. We also expect the first deployment of our Wing partnership for multimodal robot-to-drone deliveries to take place once we launch there. Last but not least, I am excited to share that we are bringing Serve robots to Atlanta in Q2.

This is another great city because it has dense commercial and residential neighborhoods, really good walkability and mixed-use developments, and a thriving restaurant scene. We're really hopeful that as we scale in Atlanta, our robots are going to actually help reduce the grid-like traffic there and improve congestion by taking these short-distance deliveries off the street. Beyond these new three cities, we plan to bring Serve robots to more cities by the end of the year, and we will announce additional details and dates as we work with our partners and local officials to finalize the plan. In terms of the fleet size, I can confirm that we are working to build the 250 Gen 3 robots that we had shared earlier by the end of this quarter. We've already built 75, more of them are now in process.

We're also currently ramping up those 75 robots into our delivery fleet, and we expect that all these 250 robots would be fully deployed and in deliveries by the end of Q2. The real volume impact that we expect to see from these robots is going to be seen mostly in the second quarter. As is typical with new hardware rollouts, we will have to deal with hardware and software adjustments, any kinks that we need to figure out. To really bring these robots to true productivity is going to take a few weeks of effort. These 250 robots are really meant to provide us valuable insights before we get to the larger scale-up that we are going to have in the second half of the year. Beyond Q1, I have another important update for the additional rollouts.

I'm very excited to share that we've achieved another 30% cost reduction in our next batch of Gen 3 robots compared to the current batch we're building. If you recall, our current batch had a 50% cost reduction compared to the previous generation robots. This is something we announced back in October. The following batches are going to be even 30% cheaper than the current batch. In total, that means that our Gen 3 robots that we are building later this year are going to be 35% of the cost of our Gen 2 robots, almost a third. This is a result of a very aggressive focus on our profitability goals, and I'm very proud of what the team has achieved in a few short months as we prepare to scale up.

This basically means that the rest of the 2,000 robots after the 250 we are building this quarter are going to be the cheaper new robots. We are on track to build them in the second half of the year at a cost that would be millions of dollars less as a result of this cost down. Our plan is to build 700 of these new lower-cost robots in Q3 and then the rest of them in Q4. We are going to share more updates, obviously, as we go through and get closer to these dates. A lot of the work we are doing this quarter and next quarter is really preparation for a big scale-up in the second half of the year.

I mentioned the additional robot cost reduction as an example, but there is a lot more effort that's going into hiring and training staff, preparing internal tools and operating procedures, working with our partners and municipalities, sending robots and staff to map and collect data and run tests in new cities and neighborhoods, etc. Our number one priority in the first half of the year is getting ready for the second half of the year rather than any short-term targets. I expect we see volume growth and revenue growth in Q1 and Q2, but mainly as an outcome of the work we are doing to prepare for that big growth push for the second half. Now, on top of those efforts I mentioned, we are also trying to build redundancy in our plans.

We want to make sure that we have layers of contingency in place just in case unexpected things happen, whether it's government dynamics changing or the shipping delays or any factors outside of our control that could threaten our timelines in a given market. We want to have additional options, such as going deeper in given cities like L.A. and Miami, both of which have significant demand for us to meet. The plan for 2025 is really to build the rest of the 250 robots in this quarter, have them all fully deployed by the end of next quarter so that we can learn from them, build the rest of the 2,000 robots in the second half of the year at a cost that would be about 30% less than the current robot cost, and simultaneously launch Dallas in Q2, followed by Atlanta, and add additional cities after that.

In order to make sure we successfully accomplish all this, spend the first half of the year really preparing for the second half scale-up and design redundancy into our plans so that we are ready for the inevitability of unknown unknowns. Okay, with 2024 and 2025 plans out of the way, I want to spend a few minutes talking about the recent advances in AI and how that impacts Serve. AI is really advancing at a breakneck speed: cheaper training, faster inference, open-source models like DeepSeek, you name it. For us, this is all tailwind. Right now, everyone and their grandma is racing to build AI models and infrastructure, which naturally results in commoditization. Our view is that a lot of the value from AI will accrue to the application layer, where the same kind of race to the bottom is less common.

Also, thanks to this proprietary data and domain expertise that you have in the application layer, you can create more efficient modes. If, as part of running your business, you accumulate unique data, you can train your AI to get better, which makes your products better, which helps you scale more, which means you get even more data. You can collect this data at lower and lower costs, eventually even while making a profit. This flywheel really gets going, and it becomes harder and harder for new entrants to compete with you. Serve is a great example of this. Our robots, they're an application layer on top of AI. They're really AI-embodied. They collect a terabyte of data each per day. That data helps make our AI and autonomy better, which leads to even better robots, which leads to even more scale and therefore more data.

You see where this goes. This flywheel really gets going. I do not really see any evidence today of a plateau in AI progress, but I do expect that it is inevitable that at some point we would follow the classic hype cycle, which is folks get ahead of their skis and then pull back. I have always had this aversion to running a business on hype. The way we like to operate in this kind of cyclical environment is by really focusing on first principles. What do we believe to be true regardless of what the current thing is? I believe there are three fundamental truths that are relevant to Serve. First, machines can now comprehend and communicate with us. It used to be that you had to say specific magic words to get machines to react, but now you can hold natural conversations in almost any language.

Second, machines can now generalize. We used to train AI for very narrow tasks. Any problem had to be broken down to a small list of tasks to train very specific AI models for. This really limited how much we can bring AI into the physical world because it has so much complexity and randomness. Now, thanks to transformers, we can train machines that develop their own complex internal world models. This leads them to exhibit emergent behaviors that you were not expecting when you were actually training them, and that makes them a lot more suitable for this unpredictability of the physical world. Third and last is the inevitable fact that AI is going to enter our physical spaces. Limiting AI to the digital realm really limits its potential.

We are going to see these new thinking machines of ours enter our physical spaces as well to make our lives better. I believe the time for this transition is now. Now that the machines can finally talk to us and now that they can understand our messy physical world, we are finally in a position to really realize our robot dreams. Robotics is really one of the most important natural endpoints of AI progress. What's unique about robots as an application of AI is that while software gets commoditized, hardware is hardware. Robots are a great example of how value can accrue to the application layer because it takes years to build that hardware, and it's expensive. It scares a lot of new entrants. Existing players enjoy the benefits of economy of scale as they progress.

Robots organically collect a ton of application-specific and proprietary data, which is otherwise very hard to obtain. This really feeds that data flywheel. Needless to say, this is exactly the thesis that we built Serve around from the beginning. We've been engineering our unique domain-specific hardware and software, and we've been collecting petabytes of data for years already. We are in a really fortunate position right now to take advantage of this very historic moment that's taking place in front of us. With that, I'll hand over to Brian to walk you through our financials.

Brian Read (CFO)

Thanks, Ali, and good afternoon, everyone. 2024 was a defining year for Serve. Over the past 12 months, we strengthened our financial position, scaled our operations, and set the stage for a transformational 2025.

We are working hard to support the planned expansion of our business to 2,000 robots before the end of the year, while at the same time remaining disciplined with our cash to improve financial flexibility. Today, I'll walk you through the financial results and provide insight into the year ahead. As you know, our business is still in its early innings. Revenue for 2024 reached $1.8 million, with Q4 contributing $176,000. This marks over 700% growth year-over-year, demonstrating the increasing adoption of our technology and services. A significant driver of this growth was the addition of $1.2 million in software services booked in 2024, alongside a 227% increase in annual delivery and branding revenue, which, despite no new robots being deployed until the final days of the fourth quarter, grew $435,000 to reach $627,000 in 2024.

On a sequential basis, when adjusting for $26,000 in non-recurring pilot program delivery revenue reported in Q3, our quarter-over-quarter delivery and branding revenue grew by 12%. As Ali mentioned, last quarter's volume growth reflects higher utilization across our fleet and improved operational efficiencies. 2024 gross margin improved significantly from -700% in 2023 to -4% in 2024, reflecting increased fleet efficiency and a more favorable revenue mix from the high-margin nature of our software services. Within delivery and branding, gross margins expanded 76% year over year, driven by fleet utilization improvements and ongoing progress in the unit economics per delivery. In terms of Q4 gross margin, cost of sales increased in Q4 as the fixed cost base expanded from the ramp-up of our 2,000-unit fleet. As fleet utilization and delivery hours improve, the portion of cost allocated into COGS in our P&L will adjust accordingly.

Total GAAP operating expenses for 2024 were $38.2 million, increasing from $19.2 million in the prior year. This reflects increased personnel, software development, and fleet expansion costs. On a non-GAAP basis, 2024 operating expenses were $23.7 million compared to $18.7 million in the prior year. Breaking this down further for 2024, R&D expenses were the largest driver of our cost structure, totaling $24.3 million on a GAAP basis, up from $9.9 million in the prior year, driven by $11.5 million of stock-based compensation. On a non-GAAP basis, R&D spend was $12.8 million compared to $9.5 million for the prior year. This increase aligns with our ongoing investment in fleet scaling, next-gen hardware, and autonomous capabilities. General and administrative expenses total $10.1 million, increasing from $4.6 million in the prior year.

The primary driver here was stock-based compensation expense, infrastructure investments as we strengthen our internal systems and governance, and headcount expansion. On a non-GAAP basis, G&A expense was $7.3 million. For our fourth quarter, total GAAP operating expenses were $12.9 million, increasing from $8.3 million in Q3 and $6.5 million in Q4 the prior year. On a non-GAAP basis, Q4 operating expenses were $8.4 million compared to $6.1 million in Q3 and $6.2 million in Q4 last year, which, compared with a doubling of delivery volume, demonstrates our continued discipline in managing core expenses. Breaking down Q4 further, R&D expenses remain the largest driver, totaling $6.8 million on a GAAP basis, up from $5 million in Q3 and $2.8 million in Q4 last year.

On a non-GAAP basis, R&D spend was $4.4 million compared to $3.3 million in Q3 and $2.6 million in Q4 last year, reflecting controlled resource growth while ensuring continued innovation. General and administrative expenses total $5.2 million, increasing from $2 million in Q3 and $1.2 million in Q4 last year. On a non-GAAP basis, G&A expense was $3 million compared to $1.6 million in Q3 and $1.2 million in Q4 last year, reflecting a more normalized run rate. Our GAAP net loss for 2024 was $39.2 million compared to $24.9 million in 2023. Q4 GAAP net loss was $13.1 million compared to $8 million in Q3 and $7.1 million in the same period last year. The sequential and year-over-year increase reflects higher operating expenses tied to investments in R&D and corporate infrastructure as we scale for broader commercialization.

On a non-GAAP basis, our net loss for 2024 was $24.6 million, with the fourth quarter being $8.6 million. We ended the fourth quarter with a robust cash position of $123 million. Since January 2024, we've raised over $250 million, significantly improving our financial flexibility. During the fourth quarter of 2024, we received $80 million in proceeds under a previously filed ATM facility at a weighted average price of $14.04. Following the end of the year, we completed an $80 million registered direct offering at a $19 price per share with institutional investors. Due to our strong cash position, we no longer anticipate funding of our 2,000 robot fleet through equipment financing. This results in cash savings over the next two years of approximately $20 million due to the removal of interest and buyout purchase options.

Earlier today, as part of our disciplined approach to capital management, we established a new shelf registration and an at-the-market equity program, which remains subject to SEC review. As we just reviewed, Serve is fortunate to have a strong cash balance to support our growth, so we do not have plans to raise additional capital in the near term. This new structure is meant to provide us with continued flexibility to act opportunistically, whether it is to fund additional fleet expansion or capitalize if and when opportunities are presented. We believe we successfully utilized this approach in December to strengthen our balance sheet while carefully managing dilution. As a result, our enhanced liquidity position provides us the ability to operate the business from a position of strength. Looking forward to 2025, we are focused on scaling our fleet to expand top-line revenue and strengthen our margins.

We remain on track to deploy 2,000 robots by year-end 2025. After working hard to lower the cost of our robot by another 30% over and above our previous 50% cost reduction, the fleet rollout will accelerate in the second half of the year with at least 700 reduced-cost Gen 3 robots built in Q3 and the remainder built in Q4. First quarter 2025 deliveries are on pace to exceed Q4, even as we focus on operational readiness for the second-half scale-up. Software revenue remains non-recurring, and our long-term revenue model remains focused on building and deploying 2,000 robots, targeting $60 million-$80 million in annualized revenue once the fleet reaches full utilization estimated in 2026. As you would expect, we are planning for increased capital expenditures during 2025, with investments in tooling, expansion costs, and fleet build-out being thoughtfully sequenced.

Our current cash position, following the January offering and debt-free balance sheet, gives us the ability to execute our business plan from a position of strength. Estimated shares outstanding as of today are approximately $57 million shares. Last, I am pleased to announce that our audit committee has approved the transition to PricewaterhouseCoopers as our independent auditor, replacing DBB McKinnon following the completion of our 2024 audit. The audit committee's appointment of PwC, a firm with a global reputation for auditing tax and financial services, reflects the company's commitment to maintaining robust financial oversight and corporate governance. We sincerely thank DBB McKinnon for their partnership and contributions to Serve over the past several years. During 2025, we will be investing in ERP upgrades and data transformation initiatives to further strengthen internal controls and enhance operational efficiency, further reinforcing our commitment to financial discipline and scalability.

In closing, we ended 2024 with our strongest cash position ever, and we will continue to remain disciplined in managing core expenses. 2025 will be focused on executing our expansion into new neighborhoods and cities, onboarding more restaurants, and scaling our fleet. With that, let me turn it back to Aduke for Q&A.

Aduke Thelwell (Head of Investor Relations)

Thank you, Ali and Brian, for those detailed updates. We'll now transition to the Q&A session. First, I'd like to express our gratitude to all the investors and analysts who submitted questions over email. We really appreciate your engagement. Our first question, I think this might be appropriate for Ali. You mentioned that robot costs have been lowered. Does this mean you removed technology or made major component changes? How was this cost reduction achieved?

Ali Kashani (Co-Founder and CEO)

Thank you. That's a great question. No, we did not remove any technologies or capabilities or components.

This was primarily done because of improvements to our supply chain, an area that we've been investing in a lot internally. There are benefits from scale that we are now enjoying. There are, for example, suppliers, tier-two suppliers. We've now upgraded to tier one. We're getting better cost components. I can think of components that are 70% cheaper now than they used to be. I generally expect we'll continue this process of making design improvements, improving our supply chain, and by increasing scale, getting additional benefits from that to see the cost of the robots come down over time. Okay, thanks. This next question was submitted via email. Can you say why you focused on the second half for the robot rollout? Yeah, absolutely. We want to do the scale-up thoughtfully and in a measured way where we are cost-efficient.

This quarter, building 250 robots effectively means that we are tripling our fleet size. Again, in Q3, when we build another 700, another tripling of the fleet size. The point of this process is to do it gradually so we learn, we fix things along the way, we reduce costs, just as I mentioned with the hardware, for example, so that we can, again, as I said at the beginning, do this in a cost-efficient and thoughtful way as we scale.

Aduke Thelwell (Head of Investor Relations)

Perfect. This next question we received from a number of people, but most notably Glenn Madison at Oppenheimer. Can you comment on recent developments with NVIDIA? Ali, do you want to take that one?

Ali Kashani (Co-Founder and CEO)

Yeah, absolutely. What developments? We have really no ability to comment on NVIDIA's behalf. I want to be clear on that. But I do want to emphasize a couple of things.

NVIDIA isn't really privy to any material confidential information about Serve or vice versa. They invested in Serve in 2021, 2022, 2023, and 2024, right before our IPO. That is going to be a very private company. They exited that position when we became a public company. What's most important, what I can comment on, is that they remain a key partner for us on the technology side. That partnership has not changed. Our robots continue to use their technology. Even the robots we are building later this year are using those NVIDIA chips. We are continuing to work together, and that hasn't changed.

Aduke Thelwell (Head of Investor Relations)

Okay, thank you. Next question. In recent weeks, we've learned more about planned policy changes and tariff introductions. Do you anticipate any impact on your operations? Brian, do you want to take this one?

Brian Read (CFO)

Sure. Thanks for the question.

Obviously, we're monitoring developments on a daily basis to understand what the impact's going to be. The bottom line is, right now, we don't anticipate any material impact to our operations. We have a global supply chain and hardware team who are working to diversify the supply chain. We have some exposure in China, but that's going to be an immaterial impact when we think about our country of origin for the parts that we're getting. I think, really importantly, like we talked about on the call, we're seeing further cost reductions throughout this year for the remainder of the fleet. Hopefully, should policy changes come in that will impact us, we would be able to offset those with some of the additional reductions in this year.

Aduke Thelwell (Head of Investor Relations)

Okay, perfect. Next question comes from Mike Latimore at Northland. We've seen that there were recent wildfires in L.A.

Any impact on Serve? Will this slow usage or rollout plans in new neighborhoods? Ali, can you take this one?

Ali Kashani (Co-Founder and CEO)

Yeah. No, there was not any impact to our rollout plan. We had a few team members who had to, unfortunately, deal with evacuations. We experienced a few days of lower volume than usual. Compared to the impact to the broader L.A. community, the impact to us was very minimal and brief.

Aduke Thelwell (Head of Investor Relations)

Okay, thank you. Next question is from Aaron Kessler at Covert Edge. Can you give an update on the Vayu acquisition?

Ali Kashani (Co-Founder and CEO)

I can do that. Yes. We have had some issues on the closing conditions right now. The deal has been on somewhat of a pause until we work that out. The discussions continue. Hard to say where it would end up, but it is not a material transaction.

If any updates come up, we would share that with everyone.

Aduke Thelwell (Head of Investor Relations)

Okay, thank you. Follow-up question, Ali. How are things going in Miami?

Ali Kashani (Co-Founder and CEO)

Things are going great. Actually, we are seeing good utilization of the robots. We have 50 restaurants already onboarded. The growth is looking to be on the right track. We are excited to expand further there. In fact, in some of our key delivery metrics, we are ahead of schedule. Overall, I would say we are very happy with what we are seeing. We are going to continue to collect data there to improve our models and our operation overall. As we do, we are going to scale the size of the robots further.

Aduke Thelwell (Head of Investor Relations)

Okay, thank you. Our next question is from Mike Latimore at Northland. Can you provide any updates on the Gen 3 robots' performance?

Are there any early insights available on how well the new hardware is working in the new market?

Ali Kashani (Co-Founder and CEO)

Yes, I'm happy to do that. The robots are looking great. Compared to a similar timeframe of the previous generation robots, they're performing much better. We are obviously rolling them out in stages so that we can learn and detect any problems and put them really to their paces before we scale even further. The most important achievement we needed in Gen 3 robots was the ability to scale them efficiently and quickly. The manufacturing had to be scaled efficiently and quickly. We are very happy with that manufacturing process. The early results are positive, and we are going to keep learning and improving.

Aduke Thelwell (Head of Investor Relations)

Okay, and this is our final question. It comes from Mike Latimore at Northland. What are your expectations now for 2026 and beyond?

Do you still think the market can absorb as many of your robots as you can produce and optimize? Ali, do you want to take it home?

Ali Kashani (Co-Founder and CEO)

Yes. Absolutely. Yes, we do believe we see a strong demand for the robots, both in the existing markets as well as in new cities and even new countries. Ultimately, we really believe that the significant cost reduction in last-mile delivery will keep even accelerating the rate of growth that already exists in demand for last mile. That means there is going to always be demand for these robots.

Aduke Thelwell (Head of Investor Relations)

Okay, that is all our questions for today. Thank you for participating, and I turn it back to the operator.

Operator (participant)

This concludes today's conference. You may now disconnect.