You watch a video of a robot packing an order. The scene is familiar: a flat table, a poly mailer, a tangle of items. The robot arm, a sleek column of grey and orange, moves with a kind of deliberate, unhurried grace. It picks up a shirt, folds it with a fluid tuck, places it in the bag. It adds a pair of socks, a sticker. Its gripper, a simple two-fingered claw, presses the adhesive strip closed. The motion is not the hyper-speed blur of a factory conveyor, but the methodical, adaptable pace of a human worker who never tires. This is Ultra’s OP1, and it is already clocking in at third-party logistics warehouses across the U.S., doing a job that has defied full automation for years.
A wedge into the warehouse
Ultra’s bet is not on building the warehouse of the future, with its miles of automated conveyors and monolithic storage systems. Its bet is on the warehouse of the present, the one already standing, with its existing workstations and persistent labor shortages. The company’s wedge is the Operator (OP1), a multi-purpose AI robot designed to be dropped in at a standard packing station, plugged into a 120V outlet, and trained on specific tasks in hours [Y Combinator, 2024]. It starts with e-commerce order packing, one of the most repetitive and yet surprisingly complex jobs in fulfillment, where items vary wildly in size, shape, and fragility. The robot uses neural networks trained on teleoperation data from RGB cameras to determine joint and gripper positions, allowing it to adapt to new objects and workflows without extensive reprogramming [Deep Tech Week, 2025]. The promise is not wholesale replacement of infrastructure, but a surgical insertion of automation where the pain is most acute.
The team behind the gripper
The founders are a reunion of sorts. CEO Jon Miller Schwartz, COO Max Friefeld, and CTO Oliver Ortlieb previously built Voodoo Manufacturing, a Y Combinator-backed startup that aimed to create a factory from hundreds of synchronized 3D printers [TechCrunch, 2017]. That experience in scaling a hardware-intensive, software-driven operation forms a throughline. Schwartz and Friefeld also co-founded the 3D printable design marketplace Layer By Layer (YC S13) [Deep Tech Week, 2025]. They are joined by Chief Scientist Chetan Parthiban, who holds a master’s in robotics from the University of Pennsylvania and brings the machine learning expertise needed to move from pre-programmed sequences to adaptive AI [Deep Tech Week, 2025]. This is a team that has navigated the physicality of manufacturing before, and is now applying that lens to the logistics floor.
| Founder | Role | Key Background |
|---|---|---|
| Jon Miller Schwartz | Co-founder & CEO | Co-founded Voodoo Manufacturing, Layer By Layer (YC S13) |
| Max Friefeld | Co-founder & COO | Ex-BCG logistics, co-founded Voodoo Manufacturing |
| Chetan Parthiban | Co-founder & Chief Scientist | MSE Robotics, University of Pennsylvania; ex-Arena ML |
| Oliver Ortlieb | Co-founder & CTO | Ex-Y Combinator product engineer, co-founded Voodoo Manufacturing |
Early traction and the roadmap
Ultra emerged from Y Combinator’s Summer 2024 batch with a $500,000 pre-seed round led by the accelerator and joined by Pioneer Fund [Deep Tech Week, 2025]. The company reports it is already generating revenue from deployments in U.S. warehouses, with a claimed annual recurring revenue of $900,000 (estimated) [GetLatka, 2025]. While specific customer names remain private, the company’s public narrative is focused on live pilots. The immediate product roadmap extends the robot’s utility beyond its initial packing use case. According to a January 2025 podcast, pilots are expanding into package sorting, secondary packaging, and order kitting, with box-packing capabilities slated for launch soon [The New Warehouse Podcast, Jan 2025]. This expansion follows a logical path within the same workstation environment, increasing the robot’s utilization and value proposition.
The competitive landscape
Ultra does not operate in a vacuum. The field of AI-powered warehouse robotics is crowded with well-funded players, most notably Covariant, which has raised hundreds of millions to develop universal AI for robotic picking. The competitive pressure is a function of the prize: automating the trillion-dollar logistics industry. Ultra’s answer is to avoid a head-on battle for the “universal brain” and instead focus on a specific, high-value workflow with a product optimized for ease of deployment.
- Deployment speed. Where major system integrations can take months, Ultra emphasizes a drop-in model that gets robots working in hours, requiring no fixed infrastructure [Ultra, retrieved 2024].
- The complexity wedge. By starting with e-commerce packing,a task that requires perception, dexterity, and decision-making,they aim to prove capability in a domain that has resisted simpler automation.
- Capital efficiency. With a comparatively small pre-seed round, the company is leveraging its founders’ hardware experience and off-the-shelf components to build and deploy quickly.
The risk, of course, is that the problem proves even more intractable than anticipated, or that larger competitors simply outspend and out-innovate in the same niche. The company’s near-term milestones,converting pilots to multi-unit deployments and successfully launching box packing,will be critical tests of both the technology and the business model.
The cultural question on the floor
The OP1’s design philosophy reveals a subtle but important cultural assumption. It is not a cage-bound speed demon meant to be sealed away from people. It operates at a human scale, in a human workspace, on a human timetable. Its value is measured not just in parcels per hour, but in the reallocation of human effort from monotonous, physically taxing work to more supervisory, cognitive, or customer-facing roles. The implicit question Ultra is asking, then, is not merely whether robots can pack boxes. It is whether automation can finally arrive not as a disruptive, top-down overhaul, but as a quiet, competent coworker that slots into the existing rhythms of work. The answer will determine if their orange-and-grey columns become as commonplace in fulfillment centers as the tape gun and the packing slip.
Sources
- [Y Combinator, 2024] Ultra: Practical, general-purpose robots for repetitive industrial tasks | https://www.ycombinator.com/companies/ultra
- [Deep Tech Week, 2025] Ultra - Deep Tech Week | https://www.deep-tech-week.com/organizations/ultra
- [The New Warehouse Podcast, Jan 2025] How Human-ish Robots are Packing Your E-commerce Orders with Ultra | https://www.youtube.com/watch?v=pXRI76DijiQ
- [Ultra, retrieved 2024] Ultra, https://www.ultra.tech/
- [GetLatka, 2025] GetLatka profile for Ultra | https://getlatka.com/company/ultra
- [TechCrunch, 2017] Voodoo Manufacturing raises $1.4 million to make a factory full of 3D printers | https://techcrunch.com/2017/01/24/voodoo-manufacturing-raises-1-4-million-to-make-a-factory-full-of-3d-printers/
- [TechCrunch, 2016] The future of 3D-printed prosthetics | https://techcrunch.com/2016/06/26/the-future-of-3d-printed-prosthetics/
- [Forbes, 2016] Max Friefeld - 2016 30 Under 30: Manufacturing | https://www.forbes.com/pictures/ehde45ememm/max-friefeld-24/