Rerun, a Stockholm-based startup focused on developing data infrastructure for Physical AI, has secured $17 million in seed funding. The round was led by Point Nine, with participation from Sunflower Capital, Costanoa Ventures, and Seedcamp, along with notable angel investors such as Guillermo Rauch (CEO of Vercel), Oliver Cameron (CEO of Odyssey), Wes McKinney (creator of pandas, Apache Arrow and Ibis), Eric Jang (VP of AI at 1X) and Nicolas Dessaigne (General Partner at Y Combinator). The company is working on building a multimodal data stack specifically designed for robotics, autonomous vehicles, drones, and spatial computing.
Traditional data tools have struggled to meet the needs of Physical AI, and Rerun aims to address this gap with a new database and cloud platform that can manage complex data like video streams and 3D scenes. This funding will help expand Rerun’s team and support the development of their commercial offering, which is set to launch later this year. The company’s open-source visualization tools have already attracted attention from major players in the field, including Meta and Google.
Founded in 2022 by Nikolaus West, Emil Ernerfeldt, and Moritz Schiebold, Rerun focuses on developing a multimodal data stack designed for Physical AI applications, such as robotics, autonomous vehicles, drones, and spatial computing. Rerun combines open-source visualization tools with a high-performance database to manage complex, evolving data types, such as video streams and 3D scenes. This approach aims to address the limitations of traditional data tools, which are not well-suited for the unique demands of Physical AI. Rerun’s solutions aim to improve data management, experimentation, and training for AI systems interacting with the physical world. The company has gained attention in the field, with its open-source project being used by companies like Meta and Google.
“Most of the world’s GDP exists in the physical world, creating an enormous opportunity for Physical AI to transform our economy,” said Nikolaus West, Co-Founder and CEO of Rerun. “But the current data stack wasn’t built for this new reality.”
AI systems increasingly interact with the physical world through robotics, autonomous vehicles, drones, and spatial computing devices. These systems use cameras and sensors to perceive their environment and perform actions based on that perception. The datasets these systems produce don’t fit nicely into tools built for classic software development. The difficulty of understanding and working with this data is significantly slowing down industry progress.
This is at a time when Physical AI is seeing skyrocketing investment. As more humanoid robots, autonomous vehicles, and spatial computing devices move toward commercial deployment, data infrastructure is becoming a key bottleneck.
“Say a robot drives into a wall. Why did it do that? You need tools that allow you to see the world through the eyes of the robot to know why it acted the way it did,” says West. “That’s what we’ve built.”
Rerun is launching a new database and cloud data platform purpose-built for Physical AI data and workflows. The database is designed to manage complex multimodal data such as video streams, 3D scenes, and tensors that evolve at varying rates over time. Combined with Rerun’s popular open source visualization toolkit, this first-of-its-kind, new multimodal data stack supports everything from observability to experimentation and training. The new commercial offering is in development with select design partners and scheduled for general availability later in the year.
Rerun’s open source project–which made it simple to create high performance visualizations of Physical AI data–hit a nerve with the community. Today it has a vibrant Discord community, and is used in open source projects by companies like Meta, Google, Hugging Face’s LeRobot, and Unitree.
“Previously, everyone—from robotics to autonomous vehicles to AR— built much of their visualization and data infrastructure in-house. But despite a lot of investment, even the biggest companies in the world weren’t happy with the results,” says West. “We realized they were building on the wrong foundation. The old stack lacked performance and flexibility. Inspired by the flexible data models in gaming engines, we rewrote ours from the ground up in Rust.”
The Rerun team of robotics, autonomy, drone and data systems veterans come from Apple, AWS, Meta, Unity, Zenly and Zipline. It includes the creator of the rosbag format, key contributors to Apache Datafusion, and their CTO, Emil Ernerfeldt, is the creator of egui, the biggest open source GUI framework in Rust.
Traditional data tools for robotics are file oriented and lack the flexible data management needed for rapid experimentation and dataset curation. On the other hand, classic data warehouses and lakehouse architectures lack deep support for Physical AI data, which forces developers to maintain complex translation layers and duplicate code between systems. Rerun is solving these problems by bringing its flexible data model for Physical AI into the database. That gives teams built-in visualization, high performance, and simpler, more flexible systems.
“All companies developing Physical AI are bottlenecked on speed of product development,” said Moritz Schiebold, Co-Founder and COO. “There is a data flywheel in Physical AI. The smarter your system is, the more units you get to deploy, which allows you to collect more data to train on, which makes your system even smarter. Rerun helps teams run more experiments faster, which is the biggest lever on development speed.”
“Rerun’s mission to build the multimodal data stack for Physical AI is wildly ambitious and we’re excited to support Emil, Moritz, Niko and team on their journey,” said Ricardo Sequerra Amram, a partner at Point Nine. “Their work on open source has allowed them to gain the trust of some of the most ambitious companies in the world of Physical AI. As AI goes into the world of atoms, we believe Rerun can be the foundational platform enabling companies to build, iterate and ship better products continuously.”
The company plans to double its team size by the end of the year, focusing on building out its commercial database offering and making the open source project more deeply integrated into the Robotics and Physical AI ecosystems.
Click to read more funding news.
Read the orginal article: https://arcticstartup.com/rerun-raises-17m-seed/
