Mosaico’s open-source software solves data plumbing for robotics and physical AI by turning sensor data into a highly organised and training-ready asset.
-
Mosaico introduces an autonomous sensor data management layer that saves roboticist teams countless hours of manual sorting
-
Founded by three trained roboticists who’d previously built every stage of the autonomous vehicle pipeline at Ambarella
-
The technology operates as pluggable software and requires no pre-built robotics data infrastructure
-
Beyond physical AI and robotics, the software will benefit developers working on technology for IoT, smart cities, agritech, and more.
-
Funding will support product development and commercial expansion in the EU and open source distribution worldwide
Mosaico, the open-source platform for organising and managing sensor data for robotics teams, has raised a €7.5M seed round. The round was led by United Ventures with participation from OpenOcean and returning investor Connect Ventures. Founded in 2023, Mosaico has developed a software that automatically stores and sorts petabytes of data, saving roboticists countless labour hours while giving them full control over the data, applications, and results. The investment will support planned geographical expansion into the United States and bolster its position in the nascent AI-integrated IoT and robotics market.
The process of ingesting, transforming, and storing data from robotic sensors is referred to as “data plumbing,” and it is indispensable to the overall reliability, security, and efficiency of every robot. The data plumbing process, however, is one of the messiest and most cumbersome for roboticists, who spend as much as a quarter of their workforce and half of their time manually searching through petabytes of sensor data stored in folders. The Manufacturing Leadership Council estimates that 70% of manufacturers are still entering data manually, even as 44% of manufacturing leaders have at least doubled the amount of data collected in the last two years.
Since robotic sensors indiscriminately collect enormous amounts of data, finding one data point such as one five-second clip of a specific autonomous vehicle stopping at a crosswalk can take hours. Meanwhile, data plumbing has only grown in importance as generative and physical AI enters the market and presents new possibilities for robotics, IoT, agritech, and smart city technologies. In order for AI algorithms to be meaningfully integrated into these sectors, super efficient data plumbing is absolutely necessary, and can’t be left to overworked robotics teams.
Mosaico was founded by three veterans of the autonomous vehicle industry who were well-acquainted with the problems of managing multimodal sensor data manually. To fix it, they created a Databricks-style platform specifically for robotics-integrated industries. The open-source and pluggable data engine operates as a horizontal layer. It manages petabytes of time-series sensor data and replaces manual search with semantic queries. Roboticists can query massive data catalogues in plain text and instantly retrieve the matching data. It enables developers to find specific data clips in seconds, accelerating model training and validation. Its semantic search replaces manual search, so a roboticist can make a query about a scenario such as, “Show me every time Car X stopped at a crosswalk.” SQL-like queries also enable detailed filtering.
Also Read: Bleu Robotics raises €5M to scale humanoid robot deployment in factories
Gabriele Lini, Mosaico co-founder and CEO, said, “My co-founders and I spent 15 years building the complete pipeline for autonomous vehicles at Ambarella, so we understood the manual data plumbing problem and how unsuited the process was to the ascendant AI-integrated IoT and robotics market. Rather than create an end-to-end vertical solution, we decided to focus all our energy on fixing the data layer. Now we are the only fully open-source and on-prem-by-design solution on the market. This funding round not only validates our vision but is going to make it possible to expand to developers in and beyond Europe.”
Mosaico’s open-source community has grown to over 1,200 members since December 2025, many of whom are high-level engineers and researchers at organisations like Google DeepMind, Amazon, Zoox, Hexagon, ETH Zürich, EPFL, Politecnico di Milano, and other major organisations. Given that the physical AI industry is not yet ready to buy at scale, Mosaico is prioritising growing its open-source community and becoming the industry standard. As it matures, Mosaico will shift its focus to sales. The company is already closing its first commercial contracts, with active enterprise deployments already underway at leading industrial players.
Giulia Giovannini, Partner at United Ventures, said, “Every wave of AI has been unlocked by the infrastructure built beneath it, and physical AI is no different: as robotics and IoT scale, messy manual data plumbing becomes the real bottleneck. Mosaico’s founders saw this early and built the open, horizontal layer to fix it. Gabriele and his team have the credibility and ambition to make it the industry standard.”
Patrik Backman, General Partner at OpenOcean, said: “Scaling MySQL for the web era and MariaDB for the cloud era showed us that every platform shift needs its own open data layer. Only open source can become the standard a whole industry can build on, and physical AI is the next platform shift. What’s more, robotics can be an order of magnitude larger than previous software revolutions, because it will reach into every factory, farm, vehicle and city, running on sensor data that no one can yet manage at scale. Mosaico is building the data layer for this scaling market on the right principles: open source to its core and on-prem by design. We’re proud to back Gabriele and the team as they build the standard that physical AI will run on.”
This seed round brings Mosaico’s total funding to €8.5M, following its €1M pre-seed in February led by Connect Ventures. The round will support general product development and Mosaico’s plans to expand its open source distribution globally.


