The most expensive part of a package's journey is the final few feet from the van to your door. It is also, for now, the most stubbornly human part. A new Berkeley-based startup, RAY Robotics, is betting it can automate both the inside of the van and the sidewalk outside, turning the last mile into a two-robot relay race [F6S, retrieved 2024].
The company, founded in 2025, describes a system where an autonomous sorting arm inside the delivery vehicle preps parcels on the move, handing them off to a fleet of small, wheeled robots that fan out to cover the final stretch [F6S, retrieved 2024]. The idea is to parallelize the slow, linear work of a human driver walking back and forth, a unit economics problem that gets worse as e-commerce volumes climb.
The Two-Part Bet
RAY's proposed architecture tackles two distinct inefficiencies. First, the in-van sorting system is meant to eliminate the minutes a driver spends rummaging for the right box in the back. Second, the autonomous mobile robots are designed to handle the "last 100 yards," each following its own miniature delivery route from a parked van to multiple doorsteps [F6S, retrieved 2024]. If it works, the van becomes a mobile hub, and the driver becomes a supervisor, or perhaps is removed from the equation entirely.
This puts RAY into a crowded field of last-mile automation, though most competitors focus on one piece of the puzzle. Sidewalk delivery bots from Starship Technologies and others handle the final leg but still require human loading. Autonomous vehicle companies like Nuro aim to replace the van itself. RAY's bet is that automating the handoff between vehicle and curb is the more immediate unlock for legacy logistics operators, who are measured on parcels per hour.
The Stealth-Mode Questions
The ambition is clear, but the company's public footprint is remarkably light. Beyond its F6S profile, there is no verifiable record of founders, funding, or pilot customers. The competitive landscape, however, is well-defined and capital-intensive.
- Sidewalk specialists. Companies like Starship have deployed thousands of robots on college campuses and in some suburbs, proving the model for small, predictable routes.
- Autonomous vehicles. Nuro and others are developing purpose-built delivery vehicles, a heavier but more comprehensive solution.
- Industrial automation. Firms like Boston Dynamics are creating mobile manipulation robots for warehouses, a technology that could eventually migrate to vans.
For RAY to move from concept to curb, it must answer several foundational questions. Who is building the hardware, and what is the unit cost of a sorting arm plus a fleet of delivery bots? Can the system operate reliably in the chaotic, unstructured environment of a residential street? And which logistics partner would be first to test a completely unmanned delivery cell?
The math, on paper, is compelling. If a human driver spends 30 seconds finding a package and 90 seconds walking it to a door, that's two minutes of labor per stop. At a fully loaded cost of, say, $30 an hour, that's $1 per parcel just for the final handoff. A robot system that cuts that time in half starts to pencil out quickly at scale, especially for dense urban routes with high stop density. The company it must ultimately beat isn't another robot startup, but the entrenched economics of the gig-economy driver,a cost center that has proven frustratingly resistant to automation.
Sources
- [F6S, retrieved 2024] RAY Robotics profile | https://www.f6s.com/company/ray-robotics