Submitted 06/2026
Oscorp Autonomous Battery Sorting System
Short Description
The Oscorp Autonomous Battery Sorting System is a retrofit, AI-driven robotic system that identifies and removes batteries from mixed waste streams on a facility's existing conveyor lines. It fuses computer vision, multi-energy X-ray, and infrared to separate batteries by form factor and lithium-ion chemistry (NMC, LFP, LCO). It is used in incoming inspection & storage at battery recyclers and material recovery facilities, replacing manual sorting.
Inputs
Outputs
Ani GoswamiFounder CEO at Oscorp Energy“Battery sorting is the first critical decision point in the recycling value chain - getting chemistry identification right at incoming inspection is what makes downstream recovery both safe, scalable, and importantly, profitable. At Oscorp Energy, we're building the autonomous layer that removes manual sorting from the equation entirely. Happy to connect with anyone working in battery recycling or material recovery.”
Image Gallery(5)
Specifications & Service
The Oscorp Autonomous Battery Sorting System is a retrofit, AI-driven robotic system that detects, classifies, and removes batteries from mixed waste and e-waste streams on a facility's existing conveyor lines. Deployed at battery recyclers and material recovery facilities (MRFs), it replaces dangerous manual sorting — reducing fire risk, lowering insurance premiums, and segregating lithium-ion chemistries to feed downstream black-mass and hydrometallurgical processes.
- Sensor fusion: computer vision (object class, brand, form factor, likely chemistry) + multi-energy X-ray (chemistry confirmation) + OCR (brand/text) — vision narrows priors, X-ray confirms, OCR reinforces.
- Sorting: robotic arms pick into up to 7 output streams, segregated by form factor and lithium-ion chemistry (NMC, LFP, LCO), with analytics on incoming waste composition.
- Training data: ~140,000 labelled battery and electronics objects to date, via an MOU with Livium, Australia's market-leading battery recycler.
- Throughput: continuous, fully autonomous operation at ~3 items/sec per arm, targeting 200+ items/min.
- Accuracy: ~90% with vision alone today (X-ray fusion in integration); roadmap target 99.5% by end of 2026, versus ~95% for manual sorters.
- Scope: current prototype handles payloads up to ~5 kg per pick and small-format / consumer batteries; larger EV modules/packs are out of scope for now.
As an early-stage startup, Oscorp Energy is at approximately TRL 6 (proof of concept deployed; prototype in assembly) and MRL 3–4 (prototype manufacturing underway). The platform is software-defined and modular: AI models update for new chemistries and labels without hardware changes, and capacity scales by adding pick arms and sensors as throughput needs grow.