Locus Array - Long Runtime

Autonomous picking speed of the best AI model benchmarked on a real warehouse task is ~64 picks/hour, versus ~1,300 picks/hour for a human

An independently conducted community benchmark [18] tested four robot AI models on a real industrial picking task and found the best model achieved ~64 picks/hour; the ~1,300 picks/hour human baseline was also measured in the same study — though the benchmark does not specifically test the Locus Array hardware, leaving a gap in direct product-level validation.

GEODIS deployed 1,000 LocusBots across 14 global warehouse sites (U.S. and Europe) over 24 months

A GEODIS press release [11] — an independent customer announcement — directly confirms the expanded agreement to deploy 1,000 LocusBots across 14 sites in the U.S. and Europe; this is a customer-issued statement, though actual operational outcomes (throughput, labor savings) at those sites are not independently audited.

The legacy Origin/Vector AMR model requires human associates to perform the physical pick; robots navigate and carry totes but do not manipulate items

The MWPVL independent consultant review [9] and the Forrester TEI report [5] both explicitly describe the operational model as robots guiding human associates to pick locations while workers perform the physical pick, with a bot-to-picker ratio of ~3.5:1; this is further corroborated by the official picking page [4].

Locus Robotics has raised over $330M in venture funding at a ~$2B valuation (as of Series F, November 2022)

The Series F press release [10] and Nasdaq Private Market listing [8] independently confirm the $117M Series F round bringing total funding over $330M at a ~$2B valuation; however, this reflects a 2022 snapshot and no subsequent funding round or updated valuation has been reported in the dossier.