Ultra Robotics raises $62 million to sell a plug-in robot for mixed-item warehouse sorting

A carton arrives at receiving with parts for one laptop series packed together. Before any of it ships, someone on the dock splits the contents into bins, one component type to a bin ^[8]. Ultra Robotics says OP1 Operator can take over that split ^[9].

The robot does more than the split. By Ultra's account, the same unit also kits orders, pulling items from several containers into one parcel, and then packs them ^[9][10]. On the dock today, people split mixed containers by hand ^[8]. The three-job pitch asks a team to assume that one machine covers the inbound split and the outbound packing as well.

Each spot where the robot works needs two things. OP1 runs from a standard wall outlet, so nobody charges or swaps batteries, and a second cable connects it to the local network ^[1]. That cable also carries the robot's sensor data. OP1 shares the data with other robots, and Ultra uses it to improve the AI models behind its hardware ^[2]. A site running a pilot sends floor data to the vendor as part of normal operation. The arm reaches up to 10 feet, high enough to pull stock off shelves, and it can pick up an item it drops ^[7].

Ultra says its models need little configuration from customers, so a unit can be set up in a few hours ^[12]. The daily ceiling of 1,000 items comes to 125 an hour across one eight-hour shift, or about 42 an hour if the unit runs around the clock ^[14][17]. Both the setup time and the throughput are Ultra's figures. The announcement does not include a price or name a customer.

Y Combinator and NextView Ventures jointly led the seed round ^[5]. Subtracting the reported Series A from the $62 million total leaves about $12 million for the seed ^[3][4][16]. Ultra says the new capital will pay for engineering and for wider adoption of its robots ^[13]. "The robots that are actually changing the world are the ones doing valuable work for real people," said founder and chief executive Jon Miller Schwartz. "And this isn't some far-off dream. It's happening right now." ^[15]

Sort candidate stations on two lines. One is daily mixed-item volume, under or over the 1,000 items Ultra claims for one unit. The other is whether the work stays in one spot with an outlet and a network drop, or moves around the building. The cleanest pilot is a fixed spot under 1,000. It needs one unit, and the test is whether setup takes hours and the daily count holds. Put a fixed spot over 1,000 in the pilot and you learn whether one unit is enough or you are buying two. Low-volume work that moves means someone rolls the robot and re-plugs both cables at every move ^[11][1]. Leave the last box, high-volume work that moves, until the first pilot has produced its own numbers.

I'd start with the fixed inbound station under 1,000, measured on two numbers: items correctly sorted per day, and hours from delivery to the first correct bin. The tradeoff is that this pilot will not test kitting or packing, two of the three jobs Ultra is selling ^[9][10].