The 2026 IEEE International Conference on Robotics and Automation (ICRA 2026) is underway in Vienna, Austria, running from June 1 to June 5 at the Vienna Congress and Convention Center (VIECON). For researchers, engineers, startups, and robotics companies, ICRA remains one of the industry’s most influential gatherings, offering an early look at technologies that could shape the next generation of real-world robotic systems.
This year’s conference covers a broad range of topics, including humanoid robots, robot learning, autonomous systems, medical robotics, perception, manipulation, human-robot interaction, and foundation models for robotics.
ICRA 2026 Program Highlights
ICRA 2026 opened on June 1 with workshops, tutorials, technical tours, and the official conference opening. The main conference program runs from June 2 to June 4, featuring keynote presentations, technical papers, panel discussions, exhibitions, robot competitions, and industry-focused sessions.
The program spans a wide range of research areas, from field robotics and aerial systems to soft robotics, robot learning, and multi-robot coordination. On June 5, the conference concludes with additional workshops and tutorials that provide deeper technical discussions on specialized topics.
Vision-Language-Action Models Become a Major Focus
One of the most discussed research directions at ICRA 2026 is the rise of Vision-Language-Action (VLA) models. These systems aim to connect visual perception, language understanding, and physical actions, allowing robots to operate more effectively in dynamic and unstructured environments.
As the robotics industry moves beyond narrowly trained systems, researchers are increasingly focused on building robots that can understand instructions, interpret their surroundings, and adapt their behavior when conditions change. The growing number of VLA-related papers at ICRA reflects the importance of this approach within robot learning and embodied AI research.
From Dexterous Manipulation to General-Purpose Robotics
Another key trend is dexterous manipulation. Rather than demonstrating isolated tasks in controlled settings, researchers and companies are working toward robots capable of performing longer sequences of actions in unfamiliar environments.
Physical Intelligence’s pi-0.5 is one example attracting attention in this area. The system is designed to handle household tasks across different environments, highlighting ongoing efforts to improve robotic generalization beyond pre-programmed scenarios.
NVIDIA’s GR00T N1 points toward a similar objective. The platform combines low-level motor control with higher-level planning and reasoning capabilities. Its reported ability to transfer learned behaviors across different humanoid robot platforms could help accelerate development across the industry.
Simulation and 3D Perception Continue to Advance
Simulation remains a critical tool for robotics development. Platforms such as NVIDIA Isaac Lab allow researchers to train robotic policies in virtual environments before deploying them on physical hardware. This approach reduces development costs, increases training efficiency, and enables testing across a wider range of scenarios.
At the same time, advances in 3D perception are improving how robots understand and interact with the physical world. Technologies including LiDAR, depth cameras, semantic segmentation, 6D pose estimation, and real-time 3D reconstruction continue to play an important role in navigation and manipulation tasks.
As robots become more capable of operating outside structured environments, robust perception systems are becoming just as important as advances in AI models and robot hardware.
Chinese Robotics Companies Increase Their Presence at ICRA 2026
Chinese robotics companies have established a significant presence at ICRA 2026, reflecting the country’s growing role in the global robotics ecosystem.
Among the most visible participants is AgiBot, which appears as both a Platinum Partner and an exhibitor. Other notable companies include Unitree Robotics, RobotEra, LimX Dynamics, and Booster Robotics, all of which are actively developing humanoid robots, legged robots, and embodied AI platforms.
The conference also highlights the broader robotics supply chain. Companies such as RoboSense, CubeMars, Mech-Mind Robotics, and Noitom Robotics are showcasing sensors, actuators, machine vision technologies, and motion capture systems that support next-generation robotic platforms.
The growing visibility of these suppliers underscores a broader industry reality: progress in robotics depends not only on the robots themselves, but also on the supporting hardware and software ecosystem behind them.
What ICRA 2026 Reveals About the Future of Robotics
While humanoid robots continue to attract much of the public attention, ICRA 2026 suggests that the industry’s most important advances may lie beneath the surface. Vision-language-action models, improved manipulation capabilities, simulation-based training, and increasingly sophisticated perception systems are all contributing to more capable and adaptable robots.
The conference reflects a robotics industry that is becoming more focused on integration, scalability, and practical deployment. Whether these technologies can achieve widespread commercial adoption remains an open question, but the direction is becoming increasingly clear: future robotic systems will rely on the combination of intelligent software, advanced perception, and reliable hardware working together as a unified platform.
