Autonomous delivery is gaining traction as companies seek faster drop times and lower labor costs. Serve Robotics Inc. (NASDAQ:SERV) is the clearest public market expression of the trend. The company had deployed 2,000 robots in its fleet as of year-end last year. Nebius’ Avride unit is another player in this field with a fleet size of several hundred units. Avride is scheduled to cross the 1000 mark over the coming months.
In 2025, SERV grew the number of merchants it serves, reaching 4,500 restaurant and retail partners, ten times higher than the previous year, supported by its partnerships with DoorDash and Uber Eats.
The market has not rewarded that growth.
As of this writing on August 24, 2026, Serve Robotics traded at approximately $4.98 per share, up about 0.6% on the day. That places the company’s market capitalization in the range of $380 million, down from levels near $900 million in late 2025, with shares trading near the lower end of their 52-week range of $4.32 to $18.64.
The gap between operational scale and market valuation is the part worth examining.
Why the Robot Stops at the Lobby and What That Limitation Does to the Economics
Autonomous robots can navigate efficiently to a destination, but they can’t enter apartment buildings, operate elevators, or move through secured spaces.
In dense urban markets, that is not a rounding error. A meaningful share of deliveries end at a lobby door rather than a recipient’s door, meaning a person still completes the final segment.
The labor cost the technology was meant to eliminate reappears at the last step.
This is where solutions offered by QuikBot Technologies PTE Ltd. come into play.
Building the Layer That Lets a Robot Through the Door
QuikBot, a Singapore-based company, builds trust infrastructure allowing robots to interact with buildings and urban systems. Its Autonomous Final-mile Delivery platform, deployed as a Platform-as-a-Service (AFMD PaaS) and running on QuikBot’s QuikSync platform, combines indoor robots, delivery vehicles, and smart lockers, enabling robots to move through lobbies, elevators, and corridors to deliver directly to recipients without human help.
Several global logistics leaders are already collaborating with QuikBot. DHL Express, FedEx Corp. (NYSE:FDX), and United Parcel Service Inc. (NYSE:UPS) have each signed commercial agreements for deployment of the AFMD PaaS in their Singapore operations.
Beyond the carriers, the deployment footprint already crosses sectors: campus (Republic Polytechnic), healthcare (National University Hospital), hospitality (Resorts World Sentosa, including Universal Studios Singapore), commercial and industry (CapitaLand, Mapletree, and JTC), and elevator OEM partners (Mitsubishi Elevator, Otis, KONE, and Hitachi).
For a company at this stage, the customer list is the signal worth noting.
Elevator Access Is a Software Problem That Requires Hardware Relationships
Physical Infrastructure Access Is a Software Problem That Requires Hardware Relationships
Underpinning AFMD PaaS is QuikSync, QuikBot’s platform in the Ambient Permission Plane (AmbPP), a category the company coined for trust infrastructure in the Physical AI economy. QuikSync connects robots from any vendor to elevators, access control, security, and building management systems, letting them move through a building under enforceable permissions with an audit trail, for deliveries, cleaning, and security patrols alongside the human teams that run the building.
Building that capability required working with the manufacturers of that physical infrastructure themselves. In the elevator layer, QuikBot has collaborated with global elevator manufacturers, including Mitsubishi Elevator, Otis, and KONE, to name a few, to support this technology.
Elevator manufacturers are also stepping up, launching their own robot-integration solutions. Otis Worldwide Corp. (NYSE: OTIS), for example, offers Otis Integrated Dispatch, a cloud-based or on-premises API that links service robots to Otis elevators while providing automated monitoring and diagnostics to resolve connectivity issues faster.
The Neolix Partnership and the Case for an Unbroken Chain
As it continues to innovate the autonomous delivery process, QuikBot has partnered with Chinese autonomous logistics company Neolix to develop an end-to-end autonomous delivery solution spanning public roads to individual doorsteps.
The partnership integrates Neolix’s autonomous mobility capabilities and fleet operations with QuikBot’s AFMD PaaS to create a continuous delivery chain.
The mechanics are straightforward. A parcel picked up by a Neolix vehicle will be transferred to a QuikBot robot at the building entrance, which then completes the delivery process within the interiors of the building all the way to the recipient’s door.
Neither company can deliver a complete autonomous chain alone. The value sits in the handoff.
Insurance Is the Adoption Constraint Nobody Puts in the Pitch Deck
As Physical AI deployments continue to grow, so too do the challenges around liability, safety, and accountability.
In response, QuikBot has signed a publicly announced memorandum of understanding with financial infrastructure company Embed Financial Group Holdings (EFGH) to develop insurance infrastructure for autonomous systems operating through the Ambient Permission Plane, so every approved action is authorized, governed, and insured.
EFGH will develop modular risk coverage for public liability, cyber risk, product defects, business interruption, and goods in transit, tailored to commercial buildings, residential developments, retail spaces, and logistics hubs.
Most Physical AI narratives skip this part entirely.
Building owners do not adopt autonomous systems because the technology is impressive. They adopt when the liability exposure is defined, insurable, and carried by someone other than the building owner.
Where the Model Goes Next and Whether It Travels
Having established its initial deployments in Singapore, QuikBot is poised for entry into additional markets.
Government validation is reinforcing that base. As announced at ATxSummit 2026, Singapore’s IMDA, JTC, and the Singapore Institute of Technology are launching a living testbed for autonomous robots at Punggol Digital District, the country’s first to deploy multi-operator robots in a mixed-use public area. Certis, DHL, Grab, and QuikBot are the initial Design Partners, and of the eight industry leaders involved, QuikBot is the only one also named in the embodied AI collaboration alongside FieldAI, Thoughtworks, Slamtec, and Unitree.
In the United Arab Emirates, QuikBot has formed a strategic partnership with the Dubai Integrated Economic Zones Authority (DIEZ). Having completed trials, the company is now conducting pilot runs at Dubai Silicon Oasis and Dubai CommerCity, with more venues to be added later this year. A Japan initiative is also in motion, and the United States has been identified as a subsequent market, with early groundwork underway.
Whether the model translates is the open question for investors watching this category.
Singapore offers unusual building density and a regulatory posture that has actively supported autonomous systems in occupied buildings. Dubai, Tokyo, and New York each present different building codes, elevator standards, and liability regimes.
Geographic expansion is where infrastructure theses tend to prove or break.
Two Companies, Two Layers, Two Different Ways to Be Valued
While both Serve Robotics and QuikBot operate in the autonomous delivery industry, they have different strategic priorities.
Serve Robotics is primarily focused on delivering autonomous mobility in public environments. Its technology demonstrates how robots can travel safely and efficiently through city sidewalks to deliver goods.
In contrast, QuikBot addresses what happens after the robot arrives. QuikSync now orchestrates not only QuikBot’s own fleet but also Neolix’s autonomous vehicles and third-party robots across cleaning, security, concierge, and even humanoid functions, a multi-actor ecosystem in which carriers, buildings, and insurers each play a governed role.
Their functions can be considered complementary rather than competitive, as they represent different layers of the same ecosystem.
A robot must be intelligent enough to know where it needs to go. It must also be authorized to get to the end destination.
That distinction shapes how each company eventually gets valued.
From an investment standpoint, Serve Robotics offers direct exposure to the firm’s scaling of autonomous robot fleets. Every incremental delivery requires an incremental robot, and the current market capitalization reflects investors waiting on evidence of operating leverage rather than deployment volume.
QuikBot’s AFMD PaaS partnerships suggest early commercial validation, though it remains an early-stage private company whose financials are not public, and pilots do not always convert into scaled contracts. Its investment case ultimately depends on whether those integrations become repeatable, high-margin infrastructure rather than customized robotics projects.
That is the whole question.
If the permission layer becomes standardized software across many buildings, the economics look nothing like a robotics company. If every building requires bespoke engineering, they look exactly like one.
For additional information about QuikBot Technologies, readers can visit the company’s official website at www.QuikBot.ai.
Disclosure: Kirsten Co currently advises QuikBot Technologies on U.S. go-to-market strategy and does not receive compensation from the company or in connection with this article. QuikBot Technologies is privately held and is not publicly traded. References to the company are provided for informational purposes only and do not constitute an endorsement. Market data is as of August 24, 2026, and is subject to change. This content is for informational and educational purposes only and should not be construed as investment advice or a recommendation to buy or sell any security.
