Rerun: Open Source AI Platform Raises $17M Seed Funding

The Challenge of Bridging Cloud AI and the Physical World

The increasing integration of generative AI into robotics, autonomous vehicles, and drones presents a significant challenge: effectively mapping cloud-based AI systems to real-world environments. This transition often results in substantial discrepancies.

To address this issue, Rerun, a Stockholm-based startup specializing in developer tools for “physical AI,” has successfully raised $17 million in seed funding. Their focus is on providing a robust data stack to facilitate a smoother transition from code development to real-world application.

Seed Funding and Company Background

This seed round was spearheaded by Point Nine, a German venture capital firm, bringing Rerun’s total funding to $20.2 million since its founding in 2022.

Rerun’s Platform for Physical AI

According to Nikolaus West, co-founder and CEO of Rerun, the demand for AI operating within the physical realm is experiencing rapid growth.

Rerun’s database and cloud data platform are specifically engineered for the unique demands of physical AI data and workflows. It is designed to manage complex, multimodal data types, including video streams, 3D environments, and tensors.

The platform’s multimodal data stack provides powerful visual debugging capabilities. This allows developers to trace the movements of their robots and systems, identifying the root causes of errors and improving performance.

Addressing Data Infrastructure Gaps

West explains that existing data infrastructure is often ill-equipped to handle the complexities of physical AI data.

Traditional robotics tools are similarly not optimized for modern machine learning workflows, leading to data fragmentation and increased friction for development teams.

The Potential for a Data Flywheel

Rerun envisions a “data flywheel” effect, where increasingly intelligent AI systems enable the deployment of more robotics units.

This expanded deployment generates more data, which in turn fuels further AI model training and improved outcomes.

However, realizing this potential requires dedicated tools to ensure seamless model operation in the real world.

Open Source Adoption and Industry Recognition

Visualizations of physical AI data generated by Rerun’s open source platform have been integrated into various other open source projects.

Notable companies utilizing these visualizations include Meta, Google, Hugging Face, and Unitree.

Team Expertise

The Rerun team comprises experienced professionals from leading technology companies such as Apple, AWS, Meta, Unity, Zenly, and Zipline.

Emil Ernerfeldt, the company’s CTO, is also the creator of egui, a prominent open source GUI framework written in Rust.

Investor Perspective

Ricardo Sequerra Amram, a partner at Point Nine, highlighted Rerun’s strength in open source as a key factor in their investment decision.

He noted that this is a particularly substantial seed round for a European startup, reflecting the confidence in Rerun’s potential.

“Rerun’s commitment to open source has fostered trust among some of the most innovative companies in the physical AI space,” Amram stated.

Additional Investors

Sunflower Capital also participated in Rerun’s seed round, alongside existing investors Costanoa Ventures and Seedcamp.