TechCrunch News 03月21日 01:11
Rerun’s open-source AI platform for robots, drones and cars revs up with $17M seed
index_new5.html
../../../zaker_core/zaker_tpl_static/wap/tpl_guoji1.html

 

Rerun是一家位于斯德哥尔摩的开发工具初创公司,专注于为“物理AI”构建数据堆栈。该公司近期获得了1700万美元的种子轮融资,总融资额达到2020万美元。Rerun旨在帮助AI开发者解决代码到实际应用的难题,其数据平台专为处理物理AI数据和工作流程设计,支持多模态数据,如视频流、3D场景和张量。该平台提供可视化调试功能,帮助开发者追踪机器人运动,理解问题所在,并支持从真实世界传感器导入数据以进行空间可视化和分析。Rerun的开源平台已经被Meta、Google等公司用于其他开源项目,团队成员来自苹果、AWS、Meta等知名公司。

💡Rerun是一家为“物理AI”构建数据堆栈的初创公司,旨在解决AI在物理世界应用中的数据匹配问题。

💰Rerun近期获得了1700万美元的种子轮融资,总融资额达到2020万美元,由Point Nine Capital领投。

⚙️Rerun的数据平台专为物理AI数据和工作流程设计,支持多模态数据,提供可视化调试功能,帮助开发者追踪机器人运动。

🌐Rerun的开源平台已被Meta、Google、Hugging Face和Unitree等公司用于其他开源项目,团队成员来自多家知名科技公司。

📈Rerun认为物理世界AI存在“数据飞轮”效应,更智能的系统可以部署更多机器人,产生更多数据,从而训练更好的AI模型。

As Generative AI moves increasingly into the physical world through robotics, vehicles and drones, the need to map cloud-based AIs to real-world environments can lead to a huge mismatch. Rerun, a Stockholm-based dev tools startup, which builds a data stack for this emergent kind of “Physical AI”, has secured $17 million in seed funding to help AI developers smooth out the bumps from code to road, as it were.

The seed round was led by Point Nine capital out of Germany, bringing the 2022-founded startup’s total funding to date to $20.2 million.

Nikolaus West, Co-Founder and CEO of Rerun, told TechCrunch that AI which operates in the physical world is already booming. Per West, Rerun’s database and cloud data platform is built from the ground-up for Physical AI data and workflows, managing multimodal data such as video streams, 3D scenes, and tensors.

The startup’s multimodal data stack supports visual debugging, helping developers to track back through their bots’ movements to better understand where things might be going wrong — importing data from real world sensors to power spatial visualization and analytics.

“The data infrastructure that exists for AI does not really understand physical AI data. And old-school robotics tools are not built for the machine learning world, so the data breaks, and trying to use both these things together creates a huge amount of friction for the teams,” West explained.

He suggests there is a potential “data flywheel” effect in creating AI geared to the physical world. The smarter the system is, the more robotics units can be deployed, which in turn means more data to train the AI model on, ensuring better outcomes, per West.

But to get there the startup’s conviction is that AI developers need dedicated tools to help their models move smoothly through the real world.

Visualizations of Physical AI data created by Rerun’s open source platform have been incorporated into other open source projects by companies such as Meta, Google, Hugging Face, and Unitree. 

The Rerun team is drawn from veterans from the likes of Apple, AWS, Meta, Unity, Zenly and Zipline. CTO, Emil Ernerfeldt, is also the creator of egui, the biggest open source GUI framework in Rust.

Commenting in a statement, Ricardo Sequerra Amram, a partner at Point Nine, said Rerun’s strength in open source led to the investment — which is a notably large seed round for a European startup.

“Rerun’s work on open source has allowed them to gain the trust of some of the most ambitious companies in the world of Physical AI,” he suggested.

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

Fish AI Reader

Fish AI Reader

AI辅助创作,多种专业模板,深度分析,高质量内容生成。从观点提取到深度思考,FishAI为您提供全方位的创作支持。新版本引入自定义参数,让您的创作更加个性化和精准。

FishAI

FishAI

鱼阅,AI 时代的下一个智能信息助手,助你摆脱信息焦虑

联系邮箱 441953276@qq.com

相关标签

Rerun 物理AI 数据堆栈 融资 开源
相关文章