Ultra raised $62M betting warehouses will lease packing robots by the month

Ultra builds box-packing robots and leases them to third-party logistics warehouses under a robots-as-a-service model.

That warehouses will rent packing robots as a service — integration fee plus monthly fee, no capex — and that non-humanoid robots pay off years before humanoids do.

$62M announced 2026-10-09: a $50M Series A led by Framework Ventures with Y Combinator participating, on top of an earlier $12M seed led by Y Combinator and Next View.

CEO Jon Miller Schwartz, a mechanical engineer rather than a Silicon Valley software founder, started Ultra in Brooklyn to put packing robots to work in third-party logistics warehouses, choosing a robots-as-a-service model: clients pay an up-front integration fee, then an ongoing monthly fee for hardware and software support.

The model produced significant revenue (undisclosed) and enough pricing power that Ultra has raised prices, Schwartz says. The company deepened its partnership with Physical Intelligence — the $5.6B 'robot brains' firm founded by ex-Google DeepMind researchers — as a division of labor: Ultra focuses on building and installing robots, PI gets out-of-the-lab data to train its models, and the robots learn and improve inside each customer's setup.

RaaS removes the capex objection that slows warehouse automation sales, and turns Ultra's revenue into a recurring stream it can raise prices on.

The body-and-brains split lets a hardware startup use frontier robot intelligence without building it, while PI gets deployment data outside the lab — each side funds its own core competency.

It is contrarian positioning in a humanoid funding frenzy: ship the boring robot that works now, and let rivals spend years teaching humanoids not to fall over.

In a hype cycle, the unglamorous form factor with a billing model that removes the customer's capex can book real revenue long before the headliner ships.

As of 2026-10-09 Ultra is installing robots at 3PL sites across the US, says its machines have packed over 500,000 orders, has raised prices on the strength of demand, and is deepening its Physical Intelligence tie-up. Revenue is significant but undisclosed. Schwartz puts scaled humanoid deployment about five years out.