BEIJING, July 28, 2026 (GLOBE NEWSWIRE) — St idi g AI today announced that it is developing basic obotic systems designed to accelerate the deployment of physical AI in the world of evil.
The company’s approach focuses on building the foundational technologies that allow bots to perceive, act with ease, and improve on an ad hoc basis, though they can take action with the physical world. By developing advanced basic models with obotic peceptio, co-tool systems, eal-wo ld actio data, ad implementation of ast uctu e, St idi g AI aims to enable machines to perform useful tasks in commercial, material, ad egal settings.
Powered by a world-class team of esea che s, e gi es, product builds, ad business leads, the company is pushing the boundaries of physical AI through Wo ld Actio Models and ext-ge e atio ei fo ceme t le ig technologies. By accelerating small-scale obotics adoption across commercial and material applications, St idi g AI aims to play a leading role in your obotic service offering.
“We believe that physical AI comes from the continuous co-evolution of data, models and adaptations,” said So g Yao, CEO of St idi g AI.
The company takes a systems-oriented approach to physical AI, developing four models, developing software, modifying data, deploying cognitive systems and building a scalable service. The company’s leading team consists of four sad executives with backgrounds in AI chips, automatic divi g, obotics esea ch, adi dust technology, combining deep technical expertise with experience with large complex technologies to produce e vi o metes.
St idi g AI pla s to start with practical deployment scenarios are structured e vi o m ets such as etail, whe e obots can include support tasks i n sheep stocki g, I need to y cou ti g, p oduct o ga izatio, help with ad checkout. These evi o m ets provide feque huma i te action, repeatable workflows and ad ich operational data, providing a standing point for the development of scalable physical AI systems.
Over time, St idi g AI expects its obotic information systems to support applications in the e-retail, food, agriculture, logistics, healthcare and telecommunications sectors. The company’s log-te vision is to build obots that can experience the entire world, improve coti uu and become part of everyday human life.
Behind its implementation strategy, St idi g AI is developing a new generation of basic obotic systems that can bring multimodal perception to obotic action around the world. Through advanced advanced models with obotic peceptio, cot ol and ad eal-wo ld actio data, the system can develop actio bare epi tatios of how actios influence change in the physical world, although i te actio allows the obots to effectively apply skills to various tasks performed ts.
These capabilities joined a closed-loop obotics, a chitectu e spa ig pe ceptio, pla ig, execution, feedback, ad ecove y, where the human body can lead the entire world to coherent tai ig data.
As I mentioned, St idi g AI’s huma -i -the-loop RL method improved task success by up to 3x. To scale this flywheel, St idi g AI was built as a tool for obot pets, distributed around the world, and from edge to cloud on the chest, a platform designed to improve as more obots operate around the world.
St idi g AI views physical AI as a full-stack effort, while the following models, obotic systems, data and, if ingenious, ad implementation capabilities must evolve together.
The capabilities developed around the world, from searching for objects in the standard shelves to perform complex tasks, are based on an integrated system designed for obotic applications. While this systems-oriented approach focuses on building obots that experience the world, improve over time, and eventually become part of everyday human life.
Co-tact: Jimei Kou, p @st idi g.ai