Germany's Polybot secures SPRIND funding to commercialize AI-powered harvesting robot

The research team behind Polybot, a modular AI-driven agricultural robot developed at Germany's University of Hohenheim, has secured a €450K validation grant from SPRIND, the Federal Agency for Breakthrough Innovation. The funding supports the team's transition from academia to a commercial startup planned for Q4 2026. Polybot's architecture is genuinely novel: a single wheeled platform swaps end-effectors between weeding, thinning, and selective harvesting tasks via a quick-change mechanism. The AI stack uses a two-stage pipeline - a lightweight YOLOv9 detector for coarse localization, followed by a fine-grained 3D point-cloud model for millimeter-accurate manipulation. The labor economics drive the business case. Germany's horticultural sector faces a 38,000-worker seasonal shortage, according to ZVG industry data. Polybot targets asparagus, strawberry, and leafy-greens operations where labor represents 55-65% of variable costs. Early field trials at the university's research farm achieved 89% harvest accuracy on white asparagus - a notoriously difficult crop due to sub-surface detection requirements. The team projects 94% accuracy with a larger training dataset from commercial partner farms joining in summer 2026. The SPRIND grant is explicitly designed to bridge the 'valley of death' between lab prototype and investable product. Polybot's founding team is already in conversations with Lemken and CLAAS for potential distribution partnerships across the DACH region.