The most important thing to know about Max-AI is not its neural networks, but its shipping address. The company's headquarters in Eugene, Oregon, is shared with its parent, Bulk Handling Systems (BHS), a firm that has been building the physical guts of recycling plants since 1976 [Max-AI / BHS, April 2017]. It is an incumbent's bet that the brains of a sorting line should be software-defined, and it is executing that bet from inside the factory floor.
Max-AI's product is a vision system and AI that identifies material on a fast-moving conveyor belt, directing either a robotic arm or an air jet to pluck specific items from the stream. The company claims it is the world's most installed AI-powered recycling solution [NRT, 2017]. Every new BHS or NRT optical sorter or robotic line can now come with Max-AI's intelligence baked in, turning what was a hardware sale into a combined hardware-and-software platform [Max-AI, retrieved 2024].
The Wedge of Incumbent Distribution
Max-AI's market entry reads less like a disruptive assault and more like a strategic product line extension. BHS and NRT launched the technology in 2017 as a new capability for their existing machinery [Max-AI / BHS, April 2017]. The wedge was distribution. Instead of convincing a facility manager to rip out a working line to install an unproven robot, Max-AI could be sold as an upgrade to the BHS equipment already in the plant. The $20 million in funding from True West, reported in 2026, likely fuels scaling [Tracxn, retrieved 2026].
Traction followed the path you would expect. High-profile recyclers like Recology in San Francisco added Max-AI Autonomous Quality Control (AQC) units to flagship facilities [Bulk Handling Systems (BHS), October 2019]. In Europe, regional sales manager Remi Le Grand leads the push, and the technology has landed in metal scrap yards in Norway [Bulk Handling Systems (BHS), October 2025] [Max-AI, retrieved 2026].
A Crowded Field of Vision
| Company | Primary Approach | Key Differentiator |
|---|---|---|
| AMP Robotics | Standalone AI-guided robots | Deep focus on AI training data |
| Greyparrot | AI vision software analytics | Selling the data layer and purity insights |
| TOMRA Systems ASA | Sensor-based sorting + AI | Global leader in sensor-based sorting hardware |
| ZenRobotics | Heavy-duty robotic sorting | Focus on construction and demolition waste |
| Everest Labs, Glacier AI | Robotic cell solutions | Newer entrants targeting the retrofit market |
Max-AI's rebuttal is its deep integration. Its VIS (Visual Identification System) is designed to work seamlessly with collaborative robots on BHS lines, and its AQC units are built as complete sorting stations [Max-AI, May 2019].
The Unit Economics of a Ton
Max-AI attacks the cost side by reducing reliance on manual sorters and increasing the purity of output streams. A typical single-stream MRF might process 25 tons of material per hour. If an AI system can improve recovery of high-value PET plastic by 1% and reduce labor costs by $10 per ton, the annual impact for a medium-sized facility is meaningful. Over a year, that's an extra 12.5 tons of PET and $1.25 million in saved labor. The system pays for itself by making the existing operation more profitable.
For Max-AI to fully win its bet, it must prove its AI is not just an added feature, but the primary reason a recycler chooses a BHS system over a TOMRA one. It must become the defensible software moat for its parent's hardware empire.