Developer of AI-powered sidewalk delivery robots, Serve Robotics, has a question for the American food delivery industry: “Why move 2 lb burritos…in 2 ton cars?”
It is seeking to provide a solution with thousands of AI-powered, low-emission sidewalk delivery robots, with a long term aim to bring down delivery cost to about $1, and is now, as the company says, a national leader in doing so.
“The future of cities is autonomous, and we are at the forefront of this,” company co-founder and CEO Ali Kashani said on its recent earnings call.
Serve, which spun off from Uber in 2021, boasts of partners like Uber Eats and DoorDash, and had Nvidia as an investor until early this year, is not alone in doing so.
Its own partner DoorDash is developing its robot- ‘Dot’, while drone delivery services are also a rising competition. But Serve remains unfazed.
“What sets Serve apart is that we’ve built cutting-edge robots that are purpose-built from the ground up for dense urban delivery, with industry-leading autonomy and a success and safety record we’re extremely proud of,” Kashani tells Invezz in an interview.
As the company leans heavily on AI for its robots, Kashani also shares his views on the raging AI-bubble debate: “The difference between hype and value in AI comes down to real-world application…the winners will be those who turn AI into real productivity gains.”
Serve has crossed the crucial 1,000-robot deployment mark and says it is on track to its goal of deploying 2,000 robots by year’s end, even as it reported a 210% increase in revenues in Q3 at $687,000. While losses have widened, the company says it has a “clear path to breakeven”.
Kashani also shares the company’s partnership strategy, how it ensures efficient autonomous navigation, and its 5-year plan of becoming an integral player in the short-distance delivery and local logistics game.
Excerpts:
On partnership strategy and market penetration
Invezz: Partnerships will play a major role in the expansion of your reach and in driving demand for your robots. After Uber, you recently partnered with DoorDash. What is your strategy when it comes to partnerships and choosing markets you want to enter?
Our strategy is to partner with the largest and most trusted delivery platforms — companies that already have strong restaurant relationships, high delivery volumes, and established customer trust.
By integrating our autonomous delivery robots into these ecosystems, we can scale faster, optimise utilisation, and deliver meaningful economic value to both partners and merchants.
When it comes to markets, we’re very data-driven. We look at delivery density, regulatory readiness, and consumer receptivity.
We have naturally started with some of the most dense, populated markets in the US (Los Angeles, Miami, Chicago, Atlanta and Dallas), but in the next few years, we hope to introduce our robots in a wide range of cities across the US.
On navigating rising competition from other delivery robots and drone services
Invezz: Some companies like DoorDash are also building their own robots like Dot, while there are also drone services like those provided by Wing and others, which are running on a similar model to yours. How do you view this competition, and how do you plan to evolve as the space gets more crowded?
Competition is healthy. What sets Serve apart is that we’ve built cutting-edge robots that are purpose-built from the ground up for dense urban delivery, with industry-leading autonomy and a success and safety record we’re extremely proud of.
As the market matures, our focus is on deepening integration with partners, reducing cost per delivery, and leveraging AI to continuously improve routing, perception, and fleet optimisation.
Source: Serve Robotics
Autonomy as a solution to rising costs for US restaurants
Invezz: The American restaurant industry is currently grappling with high costs, including high labour costs, as well as fewer people wanting to eat out and preferring to order in. Do autonomous deliveries then offer a more cost-effective resource to the industry, or do they add to its costs?
Autonomy reduces costs. Serve aims to handle short, repeatable trips at a fraction of today’s delivery costs, while improving reliability and speed.
As we scale, unit economics are expected to become even more compelling. Our long-term goal is to bring delivery costs down to $1.
Real world application is a differentiator between hype and value in AI
Invezz: There is a spirited debate going on, especially in the wake of major tech companies announcing their results, about the magnitude of investments being poured into AI and whether these investments are leading to meaningful revenue generation. Michael Burry has bet against Nvidia and Palantir- two major AI companies. Being an AI play company yourself, how do you view this debate?
The difference between hype and value in AI comes down to real-world application.
AI isn’t a buzzword at Serve — it’s the foundation of how our robots perceive, plan, and navigate safely in complex urban environments.
Every delivery generates data that improves our models, which in turn improves performance and efficiency.
Across industries, the winners will be those who turn AI it into real productivity gains.
We’re applying AI to one of the hardest and most economically meaningful problems out there: last-mile delivery.
Ensuring efficient autonomous navigation
Invezz: One of the factors limiting the widespread use of humanoid robots is the challenge faced in navigating the real world autonomously. How are you tackling this challenge to ensure minimum intervention?
Real-world navigation is hard — and that’s where Serve excels.
Our robots are designed for sidewalk-level autonomy, which is arguably one of the most complex environments to master due to pedestrians, pets, and unpredictable movement.
We’ve invested heavily in multi-sensor fusion, advanced computer vision, and proprietary localisation techniques that allow our robots to operate safely and reliably with minimal human oversight.
Continuous learning from fleet data ensures every robot gets smarter over time, reducing intervention and improving scalability.
On aim to deploy 2,000 robots by year-end
Invezz: How on track are you to deploy 2000 robots by year’s end, and what are your estimates for the no. of robots that you aim to add over the next year, as well as the number of deliveries that you aim to achieve?
As we shared during our recent earnings commentary, we’re on track to hit our deployment goals by year end.
Scaling to 2,000 robots represents a major operational milestone, and our partnerships with Uber and DoorDash give us the demand to support that growth.
Looking ahead, we expect to continue ramping our deployments in both existing markets and new ones.
Just this past quarter, in Q3 2025, our delivery volume grew 66% quarter over quarter, representing a 300% increase from the third quarter of 2024.
On offsetting losses and the road to breakeven
Invezz: While there is operational momentum, analysts have questioned Serve’s persistent net losses and increasing operating costs. How will you achieve cost discipline, and when can you be expected to break even?
Scaling any frontier technology requires upfront investment. What’s important is that our unit economics are improving steadily as we deploy more robots and drive utilisation higher.
Each new market we enter benefits from lessons learned and efficiencies gained in previous ones.
Hardware costs are coming down, autonomy is reducing the need for human oversight, and partnerships are increasing delivery density.
Those levers move us toward profitability. We have a clear path to breakeven driven by scaling efficiencies, and we’re confident in our trajectory.
‘Expect Serve robots to become ubiquitous part of local logistics in 5 years’
Invezz: Looking ahead five years, what new use-cases or adjacent businesses do you foresee Serve entering (e.g., grocery delivery, pharmaceuticals, non-food logistics) and how do you see the robotics space evolving in the times to come?
The opportunity beyond restaurants is huge. Groceries, convenience, small parcels, and return logistics are all a natural fit.
Anywhere that there’s frequent, short-distance delivery, autonomy can create real value.
Imagine a situation where you can order two pairs of shoes and send back the pair that doesn’t fit via a Serve robot.
Over the next five years, we expect Serve robots to become a ubiquitous part of local logistics, powering a wide range of delivery types as cities rethink how goods move.
Sidewalk robots will be a normal part of city life.
The same autonomy stack that powers Serve today will enable broader logistics applications: from groceries to local parcel delivery, to other urban mobility services.
As the cost of autonomy falls and reliability rises, robots will become an integral layer of local infrastructure — just as essential as delivery vans or e-bikes are today.
Serve will continue leading that transition, expanding both our capabilities and partnerships to define the future of autonomous local delivery.
The post Interview: AI’s value lies in real-world use, says Serve Robotics’ Ali Kashani as he pushes to make sidewalk robots mainstream appeared first on Invezz
