一致する結果は見つかりませんでした。

別のキーワードでお試しください。

Innodisk, RealSense, and EverestLabs Partner to Enable Real-Time AI Sorting for Industrial Recycling

2026/03/11

Innodisk, a leading global AI solution provider, today announced a strategic collaboration with RealSense and EverestLabs—a leader in AI-powered recycling automation—to deliver a scalable, real-time AI vision solution engineered for demanding conditions of industrial recycling environments.

The joint solution integrates Innodisk’s rugged APEX-P200 edge AI system with RealSense’s depth camera and EverestLabs’s RecycleOS platform, enabling accurate material identification and sorting directly on the plant floor—where split-second decisions, environmental resilience, and continuous uptime are critical.
 

Addressing Critical Industry Challenges 

Material Recovery Facilities (MRFs), recycling operators, and waste management companies face mounting pressure to improve recovery rates, material quality, and labor efficiency while meeting stringent regulatory requirements. Traditional sorting methods struggle with the volume and complexity of mixed waste streams, resulting in contamination, lower material value, and operational inefficiencies.
 

Industrial-Grade Edge AI System at the Core

At the heart of the solution is the Innodisk APEX-P200, an industrial-grade edge AI platform engineered to withstand harsh recycling conditions, including constant vibration, dust, variable lighting, and temperature fluctuations. Powered by a 13th-Gen Intel® Core™ i7 processor and an NVIDIA RTX 2000 Ada GPU delivering up to 120 INT8 TOPS, the platform provides the high-performance compute required for real-time, multi-camera 3D analysis. By processing AI inference at the edge, the solution enables immediate decision-making without cloud dependency, helping facilities maintain throughput while improving material recovery and purity.

"Advancing sustainable recycling requires an edge AI solution that provides reliability in real-world industrial conditions," said Victor Le, President of Innodisk USA Corporation. "Working with EverestLabs and RealSense, we're delivering platforms that run continuously on production lines, helping facilities improve both recovery rates and environmental impact."
 

Precision 3D Sensing Meets Advanced Processing

RealSense’s depth camera provides factory-calibrated 3D depth with aligned color imaging, enabling precise measurement and segmentation of mixed, overlapping materials moving at conveyor speeds. As these high-precision imaging data streams require high-performance and reliable computing, Innodisk’s APEX-P200 serves as the ideal engine for the task. This combination of spatially accurate data and robust processing significantly enhances AI model performance, supporting consistent, real-time sorting decisions.

“Getting recycling right comes down to reliable perception in harsh, high-speed conditions,” said Nadav Orbach, CEO of RealSense. “Together with Innodisk and EverestLabs, we’re bringing production-grade 3D sensing to the line so operators can increase purity and recovery without slowing throughput.”
 

Proven Performance in the Field

EverestLabs integrates the APEX-P200 and RealSense technology into its AI-powered recycling systems, deploying advanced Vision AI models that help Material Recovery Facilities improve operational efficiency, reduce contamination, and scale AI-driven automation across facilities.

“Our systems operate in some of the world’s toughest industrial environments and require embedded-grade reliability, high-accuracy Physical AI, and real-time responsiveness,” said Ravichandra Malapati, Principal Engineer at EverestLabs. “Innodisk’s APEX-P200 meets all three requirements.”
 

Expanding Edge AI Through Vertical Collaboration

This collaboration reflects Innodisk's strategy of partnering across the value chain—from vertical market specialists to advanced sensing innovators. With flexible edge AI platforms and deep integration expertise, Innodisk enables partners to rapidly deploy production-ready solutions tailored to industry-specific requirements at scale.

 

読み込み中...