microAGI says Shift reached $100M ARR six months after launch
The company reports a $100 million annualised revenue milestone for Shift as it expands human-demonstration data collection and industrial robot deployment.
microAGI says Shift has grown from zero to $100 million in annual recurring revenue within six months of launch, according to a company graphic describing the milestone as an annualised run rate. The claim comes roughly three months after its $55 million seed round led by Hummingbird Ventures, with Northzone, LocalGlobe, Village Global and redalpine participating. Shift is the company’s real-world data-collection network for physical AI, recording people performing manual tasks and converting demonstrations into training data for robot models and industrial deployments.
Shift records first-person human activity rather than relying primarily on teleoperation. microAGI pairs that data layer with Atlas, which captures customer-specific production data, trains or fine-tunes control policies in simulation and deploys them onto third-party robot hardware. The company describes Atlas as hardware- and model-agnostic and says it supports industrial customers across manufacturing and logistics.
The $100 million figure is company-reported annualised revenue, rather than $100 million of revenue earned over the six-month period. On September 22, microAGI hosted its first European Physical AI Summit at its Munich workshop, bringing together industrial companies, robot manufacturers, frontier AI researchers and public-sector participants around factory deployment. It has also opened London as its first priority market outside Germany and is building a local deployment team around Atlas.
Animesh Garg joined as Chief Research Officer on September 29, bringing the team from Usable Robots. Garg previously spent six years at NVIDIA Research, served as Chief Scientific Officer at Apptronik and leads the People, AI and Robotics group at Georgia Tech. At microAGI, he leads post-training work adapting general-purpose robot models to specific production tasks and sites.
A July partnership with Google Cloud and NVIDIA expands microAGI’s compute stack through GB300 NVL72 systems and RTX PRO 6000 Blackwell GPUs for training and inference. The company says deployments feed operational data into subsequent training, connecting Shift’s human demonstrations with Atlas deployments and task-specific post-training.
Founded in 2025, microAGI is headquartered in Munich, with robotics research in Zürich and operations including London and New York. The company described its $55 million seed round as the largest seed financing raised by a German startup to date.