EverestLabs Navigator & AI Recycling Robotics

EverestLabs is a Fremont, California company that applies vision AI, AI agents and robotics to recycling and resource recovery. It serves material and plastics recovery facilities, plastics reclaimers, metal and e-waste recyclers, construction and demolition sorters and packaging brands that need data on what is recovered.

The portfolio has two platforms. EverestLabs Robotics provides compact, turnkey sorting cells that work over existing conveyor belts and are run 24/7 from a remote operations centre. EverestLabs Navigator turns RGB-depth video from the belts into real-time data and lets operators ask questions in a chat; data, vision, automation and fleet agents recommend fixes and, after approval, apply equipment presets through SCADA. The vision model is trained on more than 100 billion real-world objects.

• Edge vision AI: identifies hundreds of objects per frame in under 15 ms from RGB-depth video.
• Navigator chat: natural-language questions about losses, throughput and composition grounded in line data.
• Agents with operator approval: data, vision, automation and fleet agents recommend and apply SCADA presets.
• Compact sorting cell: 3x3 ft footprint, works on inclined or cleated belts, up to 6-way sort.
• Remote operations centre: AI and human staff monitor, tune and recalibrate robots 24/7.
• Commodity audit: continuous classification of every item creates a system of record for each belt.

Vietnamese PET and HDPE recyclers can add robotic quality control before baling to meet exporter purity specs. Material recovery facilities serving cities can recover cans and bottles from residue lines. Packaging brands and EPR organisations can use belt-level data to verify how much of their packaging is actually recovered.

Robots recover material that would otherwise go to landfill; the site cites $100,000 to $800,000 a year of recoverable material leaving a typical facility. Quality-control picking raises bale purity for PET, HDPE, aluminium cans, PP and fibre. The 3x3 ft cell installs without line redesign and reaches over 90% typical pick success. Navigator flags equipment faults early and audits every item on the belt instead of manual sampling.

Labour shortages, worker safety risks and stricter quality requirements for recycled bales are pushing recyclers to automate. EPR schemes increase the need for item-level data on recovered packaging. In Vietnam, sorting is still largely manual, so modular robots that fit existing lines suit investors upgrading recycling facilities.

Headquartered in Fremont, California. In August 2025 it announced its first international deployment, with Veolia, and in March 2026 a partnership with Innodisk and RealSense. Customers named in its news include Caglia Environmental, LRS and California Waste Solutions. The company says it has the largest installed base of AI-powered high-speed sorting systems in the industry.