On August 19, the 2026 World Robot Conference (WRC) kicked off in Beijing, bringing together leading global robotics companies and their latest innovations to explore new paths for commercial deployment. BLACK SESAME (02533) made a major appearance at the event, showcasing its full SesameX multi-dimensional embodied intelligence computing platform, highlighting the company's advanced positioning and commercialization progress in the embodied AI sector while demonstrating its technical strength in applying automotive-grade AI computing power to the robotics industry.

Fully self-developed technology builds the brain for intelligent robots

As the industry's first full-stack, self-evolving commercial deployment platform supporting holistic robot intelligence, the SesameX multi-dimensional embodied intelligence computing platform represents BLACK SESAME's core strategic expansion from smart vehicles into embodied intelligence. The company is extending its automotive-grade AI computing capabilities into robotics, positioning itself as the optimal enabler for robot commercialization across service, industrial, humanoid, and medical robot categories, delivering high-compute, low-power, and highly reliable "whole-brain intelligence" solutions.

The SesameX platform embodies the product philosophy of "deploy and evolve, perceive and understand, control and react, safety as a boundary," built upon a four-layer full-stack software architecture. The foundational computing layer is equipped with a proprietary operating system and ROS2 ecosystem, while the middleware layer enables coordinated scheduling across multiple intelligent units. A proprietary atomic application layer breaks down complex tasks into reusable skills, and the top-tier X-Safety system, benchmarked against ASIL-D automotive safety standards, internalizes safety as a system instinct through a four-domain isolation architecture, transforming robots from mere mechanical executors into true embodied intelligence partners.

Three-tier computing ladder covers robot scenarios across all categories

At the expo, BLACK SESAME unveiled its two core computing modules and development kits, Kalos and Aura, designed to flexibly adapt to the development and deployment of robots with varying forms and complexities.

Kalos, a dedicated platform for commercial service robots, integrates vision perception with motion control in a single unit, delivering up to 48 TOPS of computing power. It features a dual-core lockstep MCU motion control architecture, precisely addressing the three major pain points of commercial devices—poor stability, high cost, and difficult mass production. It is well-suited for low-speed wheeled scenarios such as delivery, cleaning, inspection, and reception, making it the most mature and cost-effective universal robot hub.

Aura, a general-purpose multi-task execution platform, is equipped with 70 TOPS of heterogeneous computing power, supporting synchronized multimodal perception and millisecond-level real-time motion control. Designed for complex operational scenarios involving multi-legged robots, industrial inspection robots, and collaborative robotic arms, it empowers robots to "understand the world and act with precision."

Liora, the flagship embodied intelligence computing platform, delivers up to 700 TOPS of peak computing power and supports world models and end-to-end control. Targeting advanced scenarios such as humanoid robots and multi-robot collaborative systems, it enables robots to achieve autonomous thinking, prediction, and decision-making.

Benchmark partner products on display demonstrate tangible strength

Beyond the core modules and development kits, BLACK SESAME also presented several benchmark partner terminal products and scenario-based solutions, showcasing the mass-production readiness of the SesameX platform.

A wheel-legged quadruped robot dog, powered by BLACK SESAME's Aura computing module, was among the highlights. A single module simultaneously handles both the "brain" and the "cerebellum": the brain manages multimodal perception, decision-making, and planning, enabling real-time understanding of complex terrain and obstacles, while the cerebellum takes charge of real-time gait and balance control, ensuring steady foot placement. With both functions coordinated within the same module, the robot can instantly avoid obstacles and land precisely on stairs, rubble, and slopes, operating stably and autonomously in high-risk environments such as power inspection and campus security.

Additionally, the Liora edge-side multimodal VLA model, built on a unified computing architecture and task scheduling mechanism, provides robots with a continuous evolution path from task execution to whole-brain collaboration. With measured VLA frame rates exceeding 10Hz, leading the industry, this platform product is purpose-built for next-generation flagship humanoid robots, endowing them with real-time environmental awareness and precise action response capabilities.

Looking ahead, BLACK SESAME will continue to broaden the boundaries of computing power applications, leveraging its full-stack proprietary technology and open, collaborative ecosystem model to drive the robotics industry from technical demonstrations toward large-scale commercial deployment, fostering high-quality development in the embodied intelligence sector.