The most expensive part of a robot arm is the time it spends not moving. For an engineer tasked with programming one to pack boxes in a warehouse, that time is measured in weeks of painstakingly coding trajectories, testing for collisions, and fine-tuning for speed. Jacobi Robotics, a startup from a cluster of UC Berkeley labs, sells a way to shrink that timeline to a single day. Their bet is that the real bottleneck in industrial automation isn't the hardware, but the missing layer of software that makes it useful.
Founded in 2022, the company has raised a $5 million seed round led by Moxxie Ventures to commercialize a suite of software tools, anchored by a proprietary motion planning library. The pitch is straightforward: give robotics engineers a modern development platform so they can stop reinventing the wheel for every new deployment and start shipping robots that work. In early production tests, the company claims its partners have seen a 95% reduction in deployment time and a 24% savings on project costs [Jacobi seed announcement]. The numbers are bold, but they point at a real and expensive problem. A warehouse manager doesn't buy a robot arm, she buys a palletizing solution. The gap between those two things is where Jacobi is trying to live.
The missing middle layer
Jacobi's core product is the Jacobi Motion Library, a collection of algorithms that compute time-optimized, jerk-limited trajectories for robot arms in milliseconds. It's available via a simple pip install, through a web application called Jacobi Studio, or via cloud API. The library is hardware-agnostic, designed to slot into existing workflows built on the Robot Operating System (ROS). This is the foundational wedge: a better, faster planning engine that promises to increase robot throughput by up to 30% [Jacobi company page].
But the company isn't stopping at a planning library. It's building out what it calls "the missing middle layer between the hardware and application-specific features" [Jacobi company page]. This includes the recently announced Jacobi Vision library, which integrates depth data and point clouds directly into the motion planning process for robots working in messy, unstructured environments. The most concrete application of this full-stack approach is the OmniPalletizer, a turnkey system that takes mixed-case flow from conveyor belts and builds store-ready pallets, enforcing stacking rules and learning from experience. It's this solution that won the company the DHL Fast Forward Challenge Americas Edition in 2025 and will be demonstrated live at the MODEX trade show in 2026 [Jacobi Robotics wins DHL innovation award].
A team built in the lab
The company's technical credibility is its most tangible asset. The founding team reads like a who's who from UC Berkeley's robotics labs, including Yahav Avigal, Max Cao, Jeff Ichnowski, and Lars Berscheid [Forbes profile]. Sitting above them as Chief Scientist is Ken Goldberg, the William S. Floyd Distinguished Chair of Engineering at UC Berkeley, a foundational figure in robotic manipulation. This isn't a team that pivoted from web apps; they are academic roboticists commercializing research that, in some cases, they authored. Their investor syndicate reflects this, heavy with venture firms like The House Fund and Berkeley SkyDeck Fund that have deep ties to the university ecosystem.
The following table outlines the key founding team members and their roles:
| Name | Role / Association | Background |
|---|---|---|
| Ken Goldberg | Chief Scientist | William S. Floyd Distinguished Chair of Engineering, UC Berkeley [NYU Computer Science Department, 2026] |
| Lars Berscheid | Co-Founder | Founder of Jacobi Robotics [Crunchbase Person Profile, 2026] |
| Yahav Avigal | Co-Founder | Researcher at UC Berkeley AI Research Lab [Forbes profile] |
| Max Cao | Co-Founder | Researcher at UC Berkeley AI Research Lab [Forbes profile] |
| Jeff Ichnowski | Co-Founder | Led the FogROS project for cloud robotics at Berkeley [TechCrunch, 2022] |
Where the proof needs to land
The risks for Jacobi are not about technical novelty, but commercial traction. The impressive deployment metrics cited are from unnamed "partners," and the public record lacks the kind of detailed, logo-bearing case studies that convince cautious operations VPs at large logistics firms. The motion planning space also has focused competitors, like Realtime Robotics, which offer their own real-time collision avoidance systems. Jacobi's answer appears to be a broader platform play and a focus on specific, high-value applications like palletizing.
- The credibility gap. Strong academic pedigrees open doors, but Fortune 500 procurement offices close deals on proven ROI with referenceable customers. The DHL award is a start, but the sales cycle in heavy industry is long.
- The platform paradox. Selling a developer-friendly library (
pip install) and a turnkey palletizing solution are different motions with different buyers. One requires building a community of engineers, the other requires winning competitive bids against established integrators. - The integration burden. Promising to slot into existing ROS workflows is smart, but every warehouse environment is a unique snowflake of legacy equipment and custom software. The last 10% of integration often consumes 90% of the time.
The company's next twelve months will be about moving from promising technology to proven deployments. A successful live demonstration at MODEX 2026, coupled with announcing a first major named customer for the OmniPalletizer, would provide the concrete evidence the story currently lacks.
The unit economics of waiting
The math for Jacobi's target customer is simple, even if the software is complex. Take a mid-sized warehouse facing labor shortages and peak-season chaos. Deploying a traditional robotic palletizing cell might cost $250,000 in hardware and, critically, another $100,000 and four weeks of engineering time for programming and integration. If the robot sits idle for a month during setup, that's a month of lost productivity and a month longer until the investment pays back.
Jacobi's claim is that its software can cut that engineering time from four weeks to less than a day, shaving perhaps $80,000 off the project cost and putting the robot to work immediately. The savings aren't just in cash; they're in the acceleration of the automation project itself. For a warehouse operator, that means the ROI timeline compresses from maybe 18 months down to 12. That's the kind of calculation that gets a project approved.
Ultimately, Jacobi Robotics isn't trying to beat other robotics software startups in a features checklist. It's aiming at the incumbent that every warehouse manager already knows: the daunting, expensive, and slow process of custom integration. If the team can turn their academic breakthroughs into a product that truly makes industrial robots plug-and-play, they won't just sell software. They'll sell time.
Sources
- [Jacobi Robotics] AI Startup Jacobi Robotics Launches with $5 Million Seed Round | https://jacobirobotics.com/blog/jacobi-launches-with-5m-financing
- [Jacobi Robotics] Company Page | https://jacobirobotics.com/company
- [Forbes] Jacobi Robotics Profile | https://www.forbes.com/profile/jacobi-robotics/
- [Manufacturing AUTOMATION] Jacobi Robotics Signs Collaboration With ABB Robotics | https://www.automationmag.com/jacobi-robotics-abb-collaboration-2026
- [Jacobi Robotics] Jacobi Robotics wins DHL innovation award | https://jacobirobotics.com/blog/dhl-award-2025
- [NYU Computer Science Department] Ken Goldberg Biography | https://cs.nyu.edu/~goldberg/
- [Crunchbase] Lars Berscheid Profile | https://www.crunchbase.com/person/lars-berscheid
- [TechCrunch] FogROS brings robotic cloud computing to ROS | https://techcrunch.com/2022/05/23/fogros-brings-robotic-cloud-computing-to-the-robot-operating-system/
- [PACK EXPO] Jacobi Robotics Exhibitor Profile | https://packexpo24.mapyourshow.com/8_0/exhibitor/exhibitor-details.cfm?exhid=16001021