Amazon's AI/robotics systems are speeding up delivery (www.aboutamazon.com)

Amazon announced a suite of AI and robotics deployments aimed at speeding delivery while reducing repetitive work and improving safety. Key systems include Blue Jay, a multi-arm robotics platform that collapses picking, stowing, and consolidating into one compact line (now testing in South Carolina); Project Eluna, an agentic AI that ingests real-time and historical facility data to generate natural‑language operational guidance (piloted in a Tennessee fulfillment center for holiday optimization); and smart delivery glasses that use computer vision for hands‑free scanning, turn‑by‑turn walking directions, and hazard detection. Complementary programs include immersive VR training (iLMDA/EVOLVE) that has trained 300,000+ drivers since 2022 and an EVOLVE driving simulator used by 6,000+ new drivers. Amazon is also investing $4B to triple its rural delivery network by 2026 and expanding same‑day capabilities (including pharmacy and perishable items) with specialized vehicles and drones. For the AI/ML community the announcements signal a shift from standalone models to integrated, human‑centric systems: agentic models for real‑time operations, edge CV in wearables, and robotics that augment rather than replace workers. Technical implications include scalable orchestration of multi‑arm manipulators, real‑time multimodal facility telemetry for decision agents, and production CV for safety and proof‑of‑delivery. Amazon also emphasized sustainable AI practices—its Packaging Decision Engine has avoided 4.2M metric tons of waste since 2015, Project P.I. reduces return waste, and investments in SMRs and water‑efficient cooling aim to decarbonize compute—pointing to a production‑scale balance of performance, workforce impact, and environmental cost.