PomegraBrooklyn warehouse robot renter Ultra raised $62M, led by Framework Ventures, and partnered with Physical Intelligence to supply the AI behind its robots.
Key Takeaways
- Ultra raised $62M in total: a $50M Series A led by Framework Ventures and a $12M seed round.
- Y Combinator backed both rounds; NextView co-led the seed. Valuation was not disclosed.
- Ultra builds the robot bodies, and Physical Intelligence supplies the AI models.
Lead
Ultra, a Brooklyn company that rents packing robots to warehouses, announced on October 9, 2026 that it has raised $62 million across two rounds. The larger piece is a $50 million Series A led by Framework Ventures, with Y Combinator participating. It follows a $12 million seed round co-led by Y Combinator and NextView Ventures.
Alongside the financing, Ultra said it is deepening its relationship with Physical Intelligence, the robot foundation model developer. The company did not disclose a valuation, and it gave no revenue figure beyond saying it has booked significant revenue.
What Does Ultra Actually Sell?
Ultra sells a service, not a machine. Customers, mostly third-party logistics warehouses, pay an upfront integration fee to have a robot installed. They then pay a monthly subscription covering hardware and software support, which is the "robots as a service" model the company uses to describe itself.
The robots pack goods for shipment. Ultra says its devices have packed more than 500,000 orders in warehouses across the US. That is a modest number by the standards of large fulfillment networks, but it is real production volume, which is rare among robotics startups that raise on the strength of demonstrations.
CEO and co-founder Jon Miller Schwartz has argued that robots doing useful work for paying customers matter more than flashier humanoid projects. He has said that humanoid robots operating at scale are about five years away.
Why Did Ultra Pair With Physical Intelligence?
Ultra paired with Physical Intelligence because each company holds the piece the other lacks. Ultra makes the bodies and runs them in live warehouses, while Physical Intelligence builds the general-purpose models that control robots. Physical Intelligence, founded by researchers with Google DeepMind backgrounds, has been valued at $5.6 billion.
For Ultra, the arrangement means it does not have to build a frontier robotics model itself. For Physical Intelligence, the benefit is data. Each deployed robot generates footage and sensor logs from real packing work, and that kind of training data is expensive to collect in a lab.
The commercial terms were not disclosed. It is also unclear whether the tie-up is exclusive, or how revenue and data rights are divided. Those details will determine whether this is a deep integration or a supplier relationship.
How Should the Round Be Read?
The round reads as a bet on deployment over spectacle. A $50 million Series A is large for a company with no disclosed valuation and no public customer list, and it comes only after a $12 million seed. The jump suggests investors saw enough repeat demand in the packing business to fund a scale-up.
The skeptical case is simple. Subscription robotics carries heavy capital costs, because the vendor owns or finances hardware that customers can cancel. Margins depend on how many orders each unit handles and how little human intervention it needs. Ultra has not published either figure.
Framework Ventures is better known for crypto and gaming investments than for industrial automation. Its lead role is notable, and Y Combinator appearing in both rounds shows continued support from the earliest backer.
What Comes Next for Ultra?
Ultra says the money will go toward engineering and wider adoption of its robots. The near-term test is whether it can add warehouse customers faster than it adds hardware obligations.
Two outcomes are plausible. If Physical Intelligence's models improve the robots' reliability on messy, varied packing tasks, Ultra could widen its range of jobs and push per-unit economics up. If the models plateau, Ultra will have to rely on its own engineering and the integration fee model, which scales more slowly.
Competition is also a factor. Warehouse automation vendors with established customer bases are adding AI-driven picking and packing, and larger robotics startups are pursuing similar foundation-model partnerships.
Outlook
Ultra now has $62 million, a lead investor from outside traditional industrial automation, and a partnership with one of the most closely watched robot-AI developers. What it lacks, publicly, is a valuation, a revenue number, and terms for the Physical Intelligence deal. The next signal will be customer growth beyond the 500,000 orders packed so far.