A wave of embodied intelligence companies from Pudong, including Ant Lingbo, Keennon Robotics, and Kepler, recently showcased their latest breakthroughs at the 2026 Inclusion·Bund Summit, demonstrating how artificial intelligence is stepping out of the digital realm and into physical reality.
These intelligent systems are now leaving the exhibition halls and entering real-world environments such as pharmacies, warehouses, and factories, where they are reliably performing complex tasks in unpredictable settings—a clear sign that Pudong's embodied intelligence industry is accelerating from technological breakthroughs to practical deployment, injecting fresh momentum into the AI-driven economy.
This year, more than 40 embodied intelligence companies gathered at the event. The focus has shifted dramatically: robots are no longer just dancing or doing flips on stage—they are now genuinely working in pharmacies, warehouses, and factories. This transition brings far greater challenges. A robot must not only execute a single motion but also understand its surroundings, interpret tasks, anticipate changes, and maintain stable performance in complex, unscripted environments.
A prime example is the "Robot Smart Pharmacy" developed jointly by Ant Lingbo and Shanghai Guoda Pharmacy. Inside the exhibition, a mock pharmacy was set up where a robot receives an order, swiftly moves to the shelf, extends its arm with precision to grip a medicine box, and delivers it to the pickup counter. According to staff, real-world pharmacies are far more demanding: medicine boxes come in similar sizes, packaging is densely packed, corridors between shelves can be as narrow as 80 centimeters, and visual interference from transparent wrapping, reflections, and shelving arrangements is constant. Equipped with the Lingbo Brain, the robot can autonomously receive orders, identify target medications, and complete grasping, picking, and delivery tasks—even assisting pharmacists with nighttime sorting, all without requiring any modifications to the store itself.
For years, the robotics industry has grappled with the "specialized machine for specialized tasks" problem: whenever the task or robotic arm changes, the algorithms must be rewritten from scratch. Ant Lingbo solution is an "embodied-native general-purpose brain." At the exhibition, different robot bodies were used for logistics and industrial scenarios, yet all were driven by the same brain—the self-developed LingBot-VLA 2.0 model from Lingbo. During its pre-training phase, LingBot-VLA 2.0 has already adapted to 17 leading robot brands and 20 robot configurations, covering single-arm, dual-arm, bipedal, and wheeled platforms.
In the life services sector, Keennon Robotics humanoid robots are collaborating in laundromats to handle long-sequence tasks such as retrieving clothes, operating machines, and folding. Kepler robots have entered coffee shops, independently handling cup retrieval, beverage preparation, and cleaning. In manufacturing, robots are moving from fixed workstations to flexible operations, autonomously executing grasping, positioning, insertion, and tightening tasks. For high-risk environments, aerial intelligent agents and quadruped robots are entering confined spaces to replace human workers in long-distance, high-danger inspection duties.
The entry points for human-machine interaction are also expanding rapidly. From grabbing a coffee to commuting, job hunting, seeing a doctor, or elder care, AI is steadily weaving itself into daily life. Ant Abao has already adapted over 10,000 services for AI integration, covering restaurant ordering, ride-hailing, tourism, government services, and more. It connects seamlessly with smartphones, vehicle systems, smart glasses, and large-model applications. Users simply state their needs in natural language, and the intelligent agent understands the intent, invokes the service, and handles the entire process from reservation to ordering and payment. Even more thoughtful, the AI remembers habits: telling Abao "buy me an iced Americano from Starbucks every day at 10 AM" enables it to set up an asynchronous "coffee alarm" task that runs automatically, long-term.
These services may seem ordinary, but they illustrate how AI is becoming everyday infrastructure for more people. At the "Future Health Center," health management is shifting from passive treatment to proactive wellness. Take a photo of a meal before eating, and AI combines your weight goals, recent exercise, and medical records to offer personalized dietary advice. Lie down on the OVAL smart temperature-controlled mattress cover, and AI independently adjusts four temperature zones based on sleep and environmental data. Millimeter-wave radar discreetly monitors seniors' vital signs, alerting family members to anomalies in real time. The Qiyuan Q1 mini-robot can complete health checks—measuring heart rate, blood oxygen, and body temperature—through simple voice commands.
Questions naturally arise: "How safe must AI be before we truly trust it with our medical care and our driving?" and "When asking AI becomes the default way to find answers, how do we teach children to think?" There may be no standard answers, but they all point in the same direction: the AI-driven economy ultimately must answer how technology can better serve people, society, and the future.