From Novelty to Infrastructure: What Physical AI Really Means
Physical AI refers to robots that can decide, move and act in the real world with minimal human micromanagement. Unlike traditional automation locked inside factories or scripted workflows, these systems navigate messy, dynamic environments such as hotels, warehouses, roads and hospitals. A recent industry report describes physical AI as an inflection point, enabled by cheaper hardware, better batteries, foundation models and powerful simulation and edge computing. Executives across sectors like logistics, agriculture and high tech see it as a way to tackle labour shortages, reshoring and reindustrialisation challenges while improving safety and flexibility. Autonomous mobile robots, cobots and robotic arms are among the fastest-growing categories. As more robots are deployed, they generate real-world data that feeds back into models, making subsequent generations smarter and more adaptable. This feedback loop is what turns isolated pilots into full-scale physical AI service robots that quietly become part of everyday infrastructure.
Pudu Robotics: Service Robots Go Mainstream in Restaurants and Hospitality
Pudu Robotics illustrates how physical AI is already embedded in hospitality and commercial spaces. The company has inaugurated a new U.S. headquarters in Dallas, Texas, transforming it into a central hub for operations across North and South America. From this base, Pudu coordinates logistics, warehousing and customer support, signalling a mature phase of commercial robotics expansion rather than small-scale pilots. Since entering the U.S. market in 2018, nearly 15,000 Pudu robots have been deployed across the Americas, with regional revenue reportedly growing 285% year on year. Its portfolio spans four core categories: service delivery robots such as the BellaBot and BellaBot Pro in restaurants and retail; commercial cleaning robots like the PUDU CC1 and BG1 series; industrial delivery robots in the T-series for warehouses and factories; and the PUDU D5 line focused on general embodied AI. Customers include major names in retail, logistics, travel and technology, highlighting how a Pudu delivery robot is increasingly a standard tool in modern service operations.
Nexar and the Rise of Real-World Intelligence for Mobility Fleets
While Pudu’s robots serve tables and clean floors, Nexar focuses on the roads that move people and goods. The company positions itself as a real-world intelligence platform for the physical AI era, targeting commercial mobility operations such as rideshare networks, commercial fleets, autonomous vehicle developers, insurers and public agencies. Nexar argues that capability is no longer the bottleneck; distribution is. To make AI mobility platforms work at scale, predictive safety and real-world intelligence must be embedded directly into the vehicles and fleet tools where millions of decisions are made daily. The appointment of experienced mobility executive Jen Vescio to its board underlines a shift from experimentation to commercial infrastructure: building data moats, partnerships and go-to-market systems that let physical AI operate reliably in production environments. Nexar’s work shows how the same intelligence layer that anticipates collisions or maps roads can power future autonomous delivery, taxi and logistics services.
From Service Bots to Robot Dogs: One Technology Stack, Many Forms
Across restaurants, warehouses and roads, a common physical AI stack is emerging: sensors, mapping, autonomous navigation, predictive models and cloud-based coordination. Service robots such as Pudu’s BellaBot or industrial T-series use this stack to move safely through crowded dining rooms or factory aisles. Nexar’s real-world intelligence similarly relies on distributed sensing, data collection and predictive algorithms embedded in vehicles. Quadruped “robot dogs” for security patrols, last‑metre delivery and infrastructure inspection draw on the same foundations. They must perceive obstacles, plan routes and coordinate with human supervisors or control centres. This convergence means that as platforms like Pudu delivery robots and AI mobility platforms mature, they effectively de-risk deployment of robot dogs in malls, hotels and smart cities. Once you can trust robots to share space with people in restaurants and on busy roads, it becomes technically and operationally easier to deploy four-legged patrol or inspection robots in equally complex environments.
What Malaysian Businesses Should Prepare For
For Malaysian F&B operators, mall managers and logistics firms, these developments are an early preview of how physical AI service robots may arrive locally. International deployments show that robots can already handle repetitive tasks such as table delivery, corridor cleaning, basic warehousing transport and safety monitoring. This offers potential cost savings, more consistent service quality and better use of human staff for higher-value roles like customer engagement and exception handling. However, the workforce impact must be managed: roles will shift from purely manual tasks to supervising, maintaining and integrating robots with existing systems. Malls might experiment with robot dogs in malls for after-hours security patrols, while hotels and hospitals deploy service robots for room deliveries or disinfection. Logistics providers can study how Pudu and Nexar scale their fleets and data platforms to understand what local AI mobility platforms may require in terms of connectivity, safety standards and staff training before physical AI becomes mainstream in Southeast Asia.