Maya Labs vs Locus Robotics

Locus Robotics ships hardware and the software that coordinates it. The feature list is physical-operations shaped: AMRs for picking, putaway and transport; adaptive point-to-point transport with no fixed paths or permanent infrastructure; multi-level and mezzanine fulfillment across floors; continuous putaway and replenishment with intelligent task interleaving. LocusONE balances work between associates and robots in real time as volume, labor and SKU mix shift. Locus Array, launched as a Physical AI class for Robots-to-Goods, extends autonomy to picking, putaway, induction, drop-off, slotting and replenishment — Locus claims up to 90% fulfillment labor reduction with Array, a vendor figure worth pressure-testing in a pilot. The integration surface is enterprise WMS: SAP EWM, Manhattan, Blue Yonder, Oracle, Dynamics 365, NetSuite, Infor and Körber.

Maya Labs is a software-research artifact. PAC-1 induces interpretable programs from natural-language scripts supporting conditionals, loops and recursion, branching, custom functions, repeating workflows, web navigation, dashboards and UI, platform bots, data pipelines, and native memory for persistent context. The pitch is machine behaviour as readable programs assembled and deployed on the fly — a program-synthesis and interpretability thesis, not an operations tool. There is no published integration catalog, no robot, no WMS handshake. The two feature sets share only the word 'autonomous'.

Neither vendor publishes a number, and the two 'contact sales' motions mean very different things. Locus Robotics is explicitly a Robots-as-a-Service subscription: you are committing to ongoing RaaS fees in exchange for robots, LocusONE and support, deployed into a facility you already operate. The economics are throughput-based — the vendor's own 'not for' list names small warehouses with low daily order volumes where RaaS fees won't pay back, and flags facilities that can't commit to ongoing subscription costs. So the real evaluation is payback: baseline picks per hour and labor cost against subscription plus deployment effort, plus WMS integration work across the systems you already run (SAP EWM, Manhattan, Blue Yonder, Oracle, Dynamics 365, NetSuite, Infor, Körber). Expect a procurement cycle with an ops pilot, not a card swipe.

Maya Labs lists pricing as contact with no tiers, no RaaS analogue and no published commercial packaging — consistent with an applied research lab whose audience is researchers, academics and developers rather than a buyer signing a multi-year automation contract. If you're comparing these as purchases, you're comparing a funded operational program against an engagement with a lab. Budget accordingly: one is capex-plus-subscription with a payback model, the other is a research conversation with no stated SLA or integration commitment.