San Francisco-based startup Tuurny has announced Nantul, an AI-integrated robotic system designed to automate the recovery of functional RAM chips from old printed circuit boards (PCBs). This solution aims to address the global electronic waste (e-waste) crisis and secure a supply of increasingly scarce legacy chips.
Developments
The Nantul system utilizes neural networks to identify components and then automatically searches the internet for temperature specifications to guide a thermal suction nozzle, removing the chips without causing physical damage. Tuurny states that each machine can harvest up to 300 RAM chips per hour. The company has signed an agreement to deploy dozens of these robots to a UK-based recycling firm in early 2027 to process massive volumes of old televisions.
Background
The world is projected to generate 82 million metric tons of e-waste annually by 2030. However, traditional recycling methods typically shred entire circuit boards, wasting valuable intact components. Manual extraction for chip reuse is too costly and difficult to scale. Tuurny originally received funding from NASA to develop circuit board repair tools, but pivoted to recycling after recognizing a high demand for "legacy chips" across various industries.
Why It Matters
This project showcases the potential of "physical AI" in addressing environmental and supply chain issues. For the Vietnamese tech community, Nantul offers a valuable lesson on applying AI to specific industrial processes rather than focusing solely on software. However, challenges remain, as the robots must handle dirty or warped circuit boards in real-world scenarios without damaging the microchips.