GTC 2025: Vention unveils AI-powered bin picking system

Full-stack software and hardware automation platform provider Vention recently announced it will debut its AI-powered bin picking technology at GTC 2025 in booth 135.

The AI conference for developers is being hosted by NVIDIA March 17-21 at the McEnery Convention Center in San Jose, Calif.

While robots today can be programmed to pick and sort parts from unstructured bins, customized programming projects can become time and cost intensive, which can limit the adoption of technology. Vention said its proprietary system can reduce cost and complexity by enabling a robot work cell to autonomously pick and organize parts following a simple CAD file upload of each object.

“AI is evolving at an unprecedented pace, and it’s helping Vention make automation even more accessible to businesses of all sizes,” said Etienne Lacroix, Vention founder and CEO. “Our latest innovation, developed by integrating hardware and AI models from NVIDIA, pushes autonomous bin picking beyond simple identification and retrieval - it adds intelligence and organization that reduces the need for expensive integration and programming time. This will ultimately make bin-picking automation more affordable to manufacturers.”

Vention is slated to commercialize its AI-driven bin-picking technology later in 2025.

AI interprets pick and place, sort objects

During a live demonstration at GTC, an ABB GoFa CRB 15000 robot arm - equipped with computer vision, finger grippers, and connected to a Vention MachineMotion AI controller - will demonstrate its autonomous capabilities.

The robot will pick and organize complex plumbing components in the work cell from McAlpine & Co. Ltd, a U.K.-based plumbing manufacturer and one of Vention’s end users.

Using only an uploaded CAD file of the part being processed, the robot can understand how to detect, select, pick, and organize a part with speed and precision. This level of autonomy is made possible through NVIDIA’s foundation models.

“Generative AI and simulation technologies have reached a tipping point to accelerate physical AI deployments across manufacturing,” said Deepu Talla, NVIDIA VP of Robotics. “Using NVIDIA Isaac Robotics platform, Vention is bringing the latest advances in AI to factories of all sizes, addressing both high-mix and high-volume manufacturing.”

Plumbing components manufacturer seeks automation partner

The demonstration is part of an ongoing development project with McAlpine & Co., which has been seeking a bin-picking and machine tending automation system. McAlpine & Co. general manager John Gordon said adopting automation for bin-picking has been a challenge for the company, despite being a seemingly simple task.

“McAlpine & Co. Ltd has been seeking an automation partner we can trust to deliver our vision of using innovative solutions that align with our rigorous quality management standards,” Gordon said. “At the core of this vision is a collaborative environment where machines take on labour-intensive and highly repetitive tasks, enabling our people to focus on value-added activities. Having successfully worked with Vention on our initial automation project, we believe now is the right time to develop a state-of-the-art AI-powered bin-picking solution together.”

Motion control enabled by Jetson Orin

Vention said its MachineMotion AI controller, accelerated by NVIDIA Jetson Orin module-on-compute platform, is the backbone of this technology - delivering the real-time processing needed for autonomous bin picking at scale. This control system goes beyond traditional programmable logic controller (PLC)-based automation's deployment times and capabilities.

The proprietary intelligence and control technology relies on AI-ready hardware and foundational models from NVIDIA to process subtasks in this end-to-end process. Vention said the Jetson AI processor delivers the GPU-accelerated computing needed for reliable autonomy, allowing MachineMotion to orchestrate processes with precision and control.

“Our integration-first intelligence strategy ensures that as new AI models emerge, we can rapidly incorporate them into our turnkey and customizable work cells - keeping our valued manufacturing customers at the forefront,” said Francois Giguere, Vention CTO.

By integrating NVIDIA Isaac CUDA-accelerated libraries and models such as FoundationPose, Vention said it can help manufacturers benefit from AI advancements that are quickly entering the marketplace.