Retail Robotics
Store Automation & Last-Mile Delivery

1. Executive Summary

The global retail robotics market is projected to reach $31.3 billion by 2028, expanding at a compound annual growth rate (CAGR) of 29.6%. This staggering growth is being driven by converging forces: chronic labor shortages in retail sectors worldwide, rapidly rising consumer expectations for seamless omnichannel experiences, and the relentless pressure on margins that demands operational efficiency at every touchpoint from shelf to doorstep.

Retail robotics is no longer limited to backroom operations. Today, robots scan inventory in store aisles, fulfill online grocery orders in micro-fulfillment centers beneath supermarket floors, deliver packages autonomously to suburban doorsteps, serve food in restaurants, scrub floors overnight, and even operate entire stores without human cashiers. The retail floor has become one of the most diverse and rapidly evolving arenas for commercial robotics deployment.

This guide provides a comprehensive analysis of every major category of retail robotics, from the computer vision systems powering inventory scanning to the navigation stacks driving last-mile delivery vehicles. We examine the leading vendors, deployment architectures, integration patterns, and ROI frameworks with particular attention to the APAC retail market, where adoption is accelerating faster than any other region globally. For retailers operating in Vietnam, Thailand, Singapore, and across Southeast Asia, the question is no longer whether to deploy retail robotics but which categories to prioritize and how quickly to scale.

2. Retail Robotics Market Landscape

2.1 Market Segmentation by Robot Category

The retail robotics ecosystem spans a remarkably diverse set of applications, each addressing distinct operational pain points within the retail value chain. Unlike warehouse robotics, where the environment is controlled and largely predictable, retail robots must operate in semi-structured or fully unstructured environments alongside shoppers, diners, and pedestrians. This fundamental distinction shapes every aspect of robot design from navigation and safety systems to human-robot interaction paradigms.

• Inventory Scanning & Data Capture: Autonomous robots that patrol store aisles capturing shelf-level data using RFID, computer vision, and LiDAR. Market leaders include Simbe Robotics (Tally), Zebra Technologies (SmartSight), and Brain Corp-powered platforms. These systems detect out-of-stock conditions, price discrepancies, and planogram violations with 95-99% accuracy.
• Micro-Fulfillment & Automated Picking: Compact automated systems built inside or adjacent to retail stores for rapid online order fulfillment. Key vendors include Takeoff Technologies, Fabric (formerly CommonSense Robotics), AutoStore, and Dematic. These systems can pick a 50-item grocery order in under 5 minutes versus 45-60 minutes manually.
• Last-Mile Delivery: Autonomous ground vehicles and drones delivering orders from stores or local hubs to customer doorsteps. Starship Technologies leads with 6+ million commercial deliveries completed, followed by Nuro (permitted for on-road autonomous delivery in multiple US states) and Amazon Scout (now integrated into Amazon delivery network).
• Food Service & Restaurant: Delivery robots, cooking robots, and automated food preparation systems. Pudu Robotics (BellaBot, KettyBot) and Bear Robotics (Servi) dominate the Asian restaurant delivery robot segment with over 70,000 units deployed globally.
• Commercial Cleaning: Autonomous floor scrubbers, vacuum systems, and UV disinfection robots. Brain Corp's BrainOS platform powers over 30,000 cleaning robots across major retailers including Walmart, Kroger, and Sam's Club.
• Customer Service & Concierge: Interactive robots providing wayfinding, product information, and customer engagement. SoftBank Robotics (Pepper), LG (CLOi), and Hanwha lead this segment, particularly in APAC department stores and airports.

2.2 Investment & Funding Landscape

Venture capital and strategic investment in retail robotics reached $4.8 billion in 2025, with significant concentration in micro-fulfillment and last-mile delivery categories. Notable funding rounds include Nuro's $600M Series D, Fabric's $200M Series C, and Starship Technologies' $100M Series C. Strategic acquirers have been equally active: Amazon's continued investment in autonomous delivery and Just Walk Out technology, Walmart's deployment of Symbotic across all 42 regional distribution centers, and Shopify's integration of 6 River Systems into its fulfillment network.

The investment thesis is clear: retail labor costs are rising globally at 5-12% annually while robotics costs are declining 15-20% per year on a capability-adjusted basis. The crossover point, where robotic retail labor becomes cheaper than human labor for specific tasks, has already been reached for inventory scanning, floor cleaning, and food delivery in high-cost markets. Southeast Asian markets are approaching this crossover for an expanding set of retail applications.

3. In-Store Inventory Scanning Robots

3.1 The Inventory Visibility Problem

Out-of-stock items represent a $1.1 trillion annual problem for global retail. Studies consistently show that 8-12% of items displayed in planograms are out of stock at any given time, with the figure rising to 15-20% during promotional periods. Traditional inventory management relies on periodic manual audits, which are expensive ($500-700 per full-store audit), time-consuming (8-12 hours for a large-format store), and immediately outdated as soon as the count is complete. Scanning robots fundamentally change this equation by providing continuous, real-time inventory visibility.

3.2 Leading Inventory Scanning Platforms

3.3 Computer Vision Architecture for Shelf Scanning

Modern inventory scanning robots employ multi-layer computer vision pipelines to extract actionable data from shelf imagery. The technical architecture typically consists of four processing stages that run in near real-time on edge hardware mounted on the robot.

3.4 Price Verification & Planogram Compliance

Beyond out-of-stock detection, scanning robots deliver significant value through automated price verification and planogram compliance monitoring. Price discrepancies between shelf tags and POS systems cost retailers an estimated $40 billion annually in the US alone, factoring in regulatory fines, customer dissatisfaction, and lost margin from incorrect promotions. Scanning robots with OCR capabilities can audit every visible price tag in the store during each pass, flagging discrepancies for immediate correction.

Planogram compliance, the degree to which actual shelf layouts match corporate merchandising plans, directly impacts sales performance. Research by the Grocery Manufacturers Association found that improving planogram compliance from the typical 50-60% to 90%+ increases category sales by 7.8% on average. Scanning robots achieve this by comparing detected product positions against master planogram data using spatial matching algorithms, generating exception reports that store associates can action within minutes.

4. Shelf-Stocking & Replenishment Robots

While inventory scanning identifies gaps, shelf-stocking robots aim to close them. This is one of the most technically challenging applications in retail robotics because it requires dexterous manipulation of diverse product geometries, packaging types, and shelf configurations. The problem is substantially harder than warehouse picking because retail shelving was designed for human hands and aesthetic presentation, not robotic actuators.

4.1 Current Capabilities and Limitations

Tote-to-Shelf Systems: Companies like Berkshire Grey and Dexterity are developing systems that can unpack totes from the backroom and place items on shelves. Current systems handle rigid, uniformly shaped products (canned goods, boxed items) with 90-95% success rates but struggle with flexible packaging (bags, pouches), fragile items, and products requiring specific orientation for display.

Collaborative Restocking Carts: A more immediately practical approach uses autonomous carts that transport restocking inventory from the backroom to the correct aisle, where human associates perform the final shelf placement. This eliminates 60-70% of the unproductive walking time in traditional restocking workflows while avoiding the manipulation challenge entirely.

Backroom Automation: Several retailers have deployed automated backroom storage systems that organize incoming inventory and present items for restocking in planogram sequence. This approach, used by Walmart in partnership with Symbiotic and Alert Innovation, bridges warehouse-style automation with the retail store environment.

5. Autonomous Checkout & Cashierless Stores

5.1 Amazon Just Walk Out Technology

Amazon's Just Walk Out (JWO) technology represents the most ambitious deployment of autonomous checkout at scale. The system uses a combination of overhead cameras, weight sensors on shelves, and deep learning algorithms to track every item a customer picks up (or puts back) and automatically charges their account upon exit. Originally deployed in Amazon Go convenience stores, the technology has been licensed to third-party retailers including Hudson Nonstop airport stores and select Whole Foods locations.

The technical architecture relies on three sensing modalities working in concert:

• Computer vision (primary): Dense arrays of ceiling-mounted cameras (typically 100-300 per store depending on size) track customer identities and hand-item interactions. Pose estimation models identify reaching and grasping actions, while object detection models classify items being taken from or returned to shelves.
• Shelf weight sensors (secondary): Load cells integrated into shelf fixtures detect weight changes, providing confirmation of item removal or return events. This modality is especially important for visually similar items (e.g., different flavors of the same product in identical packaging).
• Sensor fusion engine: A central processing system fuses camera and weight data in real-time to maintain a virtual shopping cart for each customer in the store. The system must handle occlusion (customers blocking camera views), multi-person interactions at the same shelf, and edge cases like customers handing items to each other.

5.2 Alternative Autonomous Checkout Approaches

6. Micro-Fulfillment Centers (MFC)

6.1 The MFC Revolution in Grocery Retail

Micro-fulfillment centers represent one of the most transformative applications of robotics in retail. An MFC is a compact automated fulfillment system, typically 8,000-15,000 square feet, built inside, beneath, or immediately adjacent to an existing retail store. By co-locating fulfillment automation with store inventory, MFCs enable retailers to offer online grocery fulfillment that is both profitable and fast, addressing the two fundamental challenges that have made e-grocery economically unsustainable with manual picking.

Manual grocery picking in a supermarket costs $12-15 per order in labor alone, making it nearly impossible to fulfill online orders profitably at typical grocery margins of 2-4%. An MFC reduces picking labor to $2-4 per order while achieving 10-minute order completion times versus 45-60 minutes for manual picking. This transforms the economics of grocery e-commerce from a loss leader into a sustainable, scalable channel.

6.2 Leading MFC Platforms

6.3 MFC Integration Architecture

A successful MFC deployment requires tight integration between the store's existing systems and the MFC automation platform. The integration layer must synchronize real-time inventory between the store floor and MFC storage, route online orders to the MFC while maintaining store replenishment schedules, and coordinate the handoff of completed orders to delivery or curbside pickup.

7. Last-Mile Delivery Robots

7.1 Market Overview

Last-mile delivery accounts for 53% of total shipping costs, making it the single most expensive segment of the logistics chain. Autonomous delivery robots address this challenge by eliminating the driver, the largest cost component, while operating at speeds and on routes that avoid the regulatory complexity of autonomous passenger vehicles. The market for last-mile delivery robots is projected to reach $5.4 billion by 2028.

7.2 Leading Delivery Robot Platforms

7.3 Navigation & Safety Architecture

Last-mile delivery robots face uniquely challenging navigation requirements. Unlike warehouse robots operating in controlled environments, delivery robots must navigate shared public spaces with unpredictable pedestrians, cyclists, vehicles, animals, and constantly changing environmental conditions including weather, construction zones, and seasonal obstacles.

Starship Technologies, the deployment leader with over 6 million completed deliveries, uses a nine-camera array providing 360-degree visual coverage combined with ultrasonic sensors, an IMU, and GPS for localization. The navigation stack processes data in three layers: a global path planner using pre-mapped routes, a local planner for real-time obstacle avoidance, and a behavioral layer that encodes social navigation norms such as yielding to pedestrians, staying on designated sidewalk paths, and waiting at curb cuts for safe road crossing.

Nuro operates at the opposite end of the design spectrum as a purpose-built on-road vehicle that never carries passengers. This unique category allowed Nuro to obtain the first fully autonomous delivery vehicle exemption from NHTSA, permitting a vehicle without side mirrors, windshield, or steering wheel. Nuro's R3 vehicle uses a comprehensive sensor suite of LiDAR, cameras, radar, and thermal sensors to operate on public roads at speeds up to 72 km/h.

8. Restaurant & Hospitality Robots

8.1 Food Delivery Robots

Restaurant delivery robots, specifically autonomous serving robots that transport food from kitchen to table, represent one of the fastest-growing and most visible categories of retail robotics. The segment exploded during the COVID-19 pandemic when contactless service became a priority, and has continued growing post-pandemic driven by persistent staffing challenges in food service across all major markets.

Pudu Robotics dominates the global restaurant delivery robot market with its BellaBot and KettyBot platforms. BellaBot, recognizable by its cat-like face design and interactive expressions, uses multi-floor navigation with elevator integration and can carry up to four trays simultaneously. Over 70,000 Pudu units are deployed across 60+ countries, with particularly dense penetration in China, Japan, South Korea, and Southeast Asia. A single BellaBot replaces approximately 1.5-2 FTE of bussing and delivery labor, with a unit price of $12,000-18,000 providing sub-12-month payback in most APAC markets.

Bear Robotics (Servi platform) has gained significant traction in the US market with deployments at Denny's, Chili's, and hundreds of independent restaurants. The Servi robot emphasizes reliability and simplicity, with a teach-and-repeat navigation system that restaurant managers can configure without technical expertise.

8.2 Kitchen Automation Robots

Beyond delivery, robotic systems are increasingly handling food preparation tasks:

• Miso Robotics (Flippy): AI-powered fry cook deployed at White Castle, CaliBurger, and Chipotle. The latest Flippy 2 system mounts on an overhead rail and manages entire fryer stations autonomously, handling 30% more food volume than a human cook with consistent quality.
• Picnic (Pizza Assembly): Automated pizza assembly line that applies sauce, cheese, and toppings with precise portioning. Deployed at stadiums, college dining halls, and high-volume pizza operations.
• Sweetgreen (Infinite Kitchen): Custom-built robotic assembly line for salad preparation, where ingredients are dispensed from overhead hoppers into bowls on a conveyor system. Increases throughput by 50% while reducing food waste by 25%.
• RoboBurger: Fully autonomous burger vending machine that grinds, grills, assembles, and serves burgers in approximately 6 minutes with no human intervention.

9. Commercial Cleaning Robots

9.1 The Brain Corp Ecosystem

Brain Corp has established itself as the dominant software platform for commercial cleaning robots, powering over 30,000 autonomous floor care machines worldwide through its BrainOS operating system. Unlike consumer Roombas, commercial cleaning robots are industrial-grade machines weighing 200-500+ kg that scrub, squeegee, and vacuum large-format retail floors, warehouses, and airports. Brain Corp's BrainOS converts conventional ride-on scrubbers from manufacturers like Tennant, Nilfisk, and ICE Cobotics into autonomous systems through a retrofit kit.

Walmart is the single largest deployer of Brain Corp-powered cleaning robots, operating autonomous scrubbers in over 4,000 US stores. Each unit operates during overnight hours, completing full-store floor cleaning in 2-3 hours versus 4-5 hours for manual operation, while generating detailed cleanliness maps and compliance reports for store management.

9.2 Avidbots Neo

Avidbots takes a different approach with its Neo platform, offering a purpose-built autonomous floor scrubber rather than a retrofit solution. Neo uses proprietary AI navigation with 3D obstacle detection and has completed over 10 million kilometers of autonomous cleaning across airports, shopping malls, hospitals, and convention centers. The Neo platform achieves a 72% reduction in labor costs for floor care operations, with a typical lease price of $2,500-3,500 per month versus $4,000-6,000 for equivalent manual labor in US markets.

10. Customer Service & Concierge Robots

Customer-facing service robots serve dual purposes in retail environments: providing practical assistance (wayfinding, product information, multilingual support) and generating marketing value through novelty and brand differentiation. While the ROI calculation is less straightforward than for operational robots, the category has found sustainable niches in airports, large shopping malls, hotels, and department stores across Asia.

10.1 Notable Platforms

• SoftBank Robotics Pepper: The most widely recognized humanoid service robot, with over 27,000 units produced. Pepper provides customer greeting, product recommendations, and wayfinding in retail stores, banks, and hotels. Particularly popular in Japan (SoftBank mobile stores) and Europe (HSBC bank branches, Carrefour stores).
• LG CLOi GuideBot: Deployed in airports (Incheon, Seoul) and retail spaces, CLOi provides autonomous guided tours, product information, and multilingual assistance. The platform integrates with building management systems for elevator control and access management.
• Keenon Robotics (T5, T8): Chinese-manufactured service robots with strong APAC presence, deployed across hotels, restaurants, and retail stores for delivery and guest interaction tasks. Over 40,000 units deployed globally with significant presence in Southeast Asian hospitality.
• OrionStar Mini: Compact customer service robot by Cheetah Mobile, focused on retail greeting, temperature screening, and product promotion. Especially popular in Chinese and Southeast Asian convenience stores.

11. APAC Retail Automation Landscape

11.1 China: The Global Leader in Retail Robotics Deployment

China is the world's largest and most advanced market for retail robotics, driven by technology giants that have integrated automation into every layer of the retail experience. Understanding China's retail robotics ecosystem is essential for APAC operators because Chinese vendors and models are increasingly available throughout Southeast Asia at price points 40-60% below Western equivalents.

Alibaba Hema (Freshippo): Alibaba's new retail concept stores combine physical shopping with automated fulfillment. Ceiling-mounted conveyor systems transport online orders from shelf picking zones to packing stations, while in-store customers use the Hema app for checkout. Each store serves as both a retail destination and a delivery hub, fulfilling online orders within 30 minutes for customers within a 3 km radius. Over 300 Hema stores operate across China, with the model being replicated by competitors.

JD.com: JD operates fully unmanned convenience stores and has deployed autonomous delivery vehicles (JD Delivery Robots) on public roads in over 25 Chinese cities. The JD delivery robot has completed over 1 million commercial deliveries. JD's 7Fresh supermarket chain incorporates autonomous shopping carts and robotic fish tank displays that use computer vision to provide customers with information about seafood products by pointing at them.

Meituan: China's leading food delivery platform operates autonomous delivery drones and ground robots across multiple cities. Meituan's drone delivery service completes deliveries in 12 minutes on average, serving as a blueprint for how delivery robotics integrates with super-app ecosystems.

11.2 Japan & South Korea

Japan's extreme labor shortage (1.3 million unfilled retail positions) has made it the most receptive market globally for customer-facing retail robots. FamilyMart, Lawson, and 7-Eleven Japan are all piloting various levels of store automation, from autonomous checkout to robotic shelf stocking. South Korea's CU and GS25 convenience chains operate several unmanned store formats. Woowa Brothers (operator of Baedal Minjok, South Korea's largest food delivery platform) has deployed indoor delivery robots in apartment complexes and office buildings, with Dilly Drive robots completing over 200,000 deliveries.

11.3 Southeast Asia

Grab: Southeast Asia's super-app has invested in last-mile delivery automation research, partnering with autonomous delivery companies for pilot programs in Singapore. Grab's cloud kitchen network, GrabKitchen, represents a natural deployment environment for food preparation and delivery robots.

Singapore: The most advanced Southeast Asian market for delivery robotics. Starship Technologies' sidewalk robots are operating at the National University of Singapore campus. The government's Smart Nation initiative provides grants covering up to 70% of qualifying automation costs for retail operators through the Productivity Solutions Grant (PSG).

12. Vietnam Retail Robotics Opportunity

12.1 Market Context

Vietnam's retail sector is at an inflection point that creates significant opportunities for early robotics adoption. Modern trade (supermarkets, convenience stores, mini-marts) now accounts for 28% of total retail sales, up from 12% in 2018, and is projected to reach 40% by 2030. This rapid modernization of retail formats creates a natural onramp for automation technologies that were designed for organized retail environments.

Key factors shaping Vietnam's retail robotics opportunity:

• Convenience store explosion: Vietnam's convenience store count has grown from 1,500 in 2018 to over 7,500 in 2025, led by GS25 (South Korean), Circle K, 7-Eleven, and domestic chains like Vinmart+ (now WinCommerce). This dense network of small-format stores is ideal for restaurant delivery robots (BellaBot-style platforms for in-store service) and cleaning robots.
• E-grocery growth: Online grocery penetration in Vietnam reached 8% in 2025, still far below China (25%) and South Korea (30%), but growing at 45% annually. As e-grocery scales, the unit economics will demand micro-fulfillment automation for major players like WinMart, Big C (Central Retail), and Emart (Thaco).
• Rising labor costs: Vietnamese retail wages in Ho Chi Minh City have increased from $220/month in 2019 to $340/month in 2025, with continued 8-10% annual growth projected. While still significantly below developed markets, the trajectory is closing the gap with automation ROI thresholds for cleaning, inventory scanning, and food delivery applications.
• Young, tech-forward consumers: Vietnam's median age of 30, combined with 78% smartphone penetration and high social media engagement, creates a consumer base that is receptive to robotic interaction and automated retail experiences rather than resistant to them.

12.2 Recommended Deployment Priorities for Vietnam

13. Implementation & ROI Framework

13.1 Total Cost of Ownership Model

The economic case for retail robotics depends on accurately modeling total cost of ownership (TCO) against the labor and opportunity costs being displaced. The following framework provides a structured approach to evaluating retail robotics investments.

13.2 Phased Deployment Strategy

We recommend a three-phase implementation strategy that balances quick wins with strategic, longer-term automation investments:

1. Phase 1 -- Quick Wins (Months 1-3): Deploy restaurant delivery robots and commercial cleaning robots. These categories offer the fastest payback, lowest integration complexity, and most visible impact on operations. Start with 2-3 units per location to validate operational workflows and staff acceptance. Measure labor hours saved, customer feedback, and unit utilization rates.
2. Phase 2 -- Data-Driven Operations (Months 4-9): Introduce inventory scanning robots in high-volume stores. Use the shelf data to optimize planogram compliance, improve in-stock rates, and generate analytics that justify Phase 3 investment. Pilot autonomous checkout in one high-traffic convenience format location to test consumer acceptance.
3. Phase 3 -- Strategic Transformation (Months 10-18): Evaluate micro-fulfillment center deployment for the highest-volume e-grocery locations. Investigate last-mile delivery pilots in controlled environments (gated communities, commercial campuses). Build the data infrastructure to connect all retail robotics systems into a unified operational dashboard.

13.3 Critical Success Factors

• WiFi infrastructure: Retail robotics requires consistent connectivity. Conduct a wireless site survey before deployment and invest in enterprise-grade access points (Cisco Meraki, Aruba, Ruckus) with a minimum of -65 dBm signal strength throughout the operating area. Budget $5,000-15,000 per store for WiFi upgrades.
• Floor surface quality: Autonomous navigation systems, particularly cleaning robots and delivery robots, are sensitive to floor conditions. Cracks, significant slope changes, transitions between flooring types, and wet surfaces can impair robot performance. Audit floor conditions before procurement.
• Staff change management: Retail employees often fear automation will eliminate their jobs. Successful deployments frame robots as tools that eliminate tedious tasks (mopping floors, walking to find stock) while freeing staff for higher-value customer interaction. Invest in comprehensive training programs and designate robot champions on each shift.
• Customer communication: Proactively communicate the presence of robots to customers through signage, staff explanations, and social media. In APAC markets, our experience shows that the novelty of retail robots generates significant organic social media content that provides free marketing value during the initial 3-6 month deployment period.