Microagi Taps Google Cloud to Power Embodied AI Robotics
Munich-based startup Microagi is betting on Google Cloud and NVIDIA’s Blackwell platform to bridge the gap between impressive robot prototypes and reliable industrial labor. By leveraging high-performance compute, the company aims to scale its "Atlas" platform, which trains task-specific models to navigate complex, real-world physical environments for enterprise clients.
The collaboration grants Microagi access to NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs and GB300 NVL72 rack-scale systems. This technical foundation allows the startup to process massive multimodal datasets, including video, to refine robots for specific commercial roles. The goal is to move beyond generic automation by offering customizable software packages for sectors like hospitality and industrial manufacturing.
Microagi, founded in 2025, positions its Atlas platform as a hardware-agnostic layer that integrates with existing frontier AI models to prevent vendor lock-in. Founder Bercan Kilic noted that the partnership provides the necessary infrastructure to support the company’s intensive model inference pipeline. By utilizing Google Cloud's AI stack, Microagi intends to streamline workflows and accelerate the deployment of intelligent machines that perform consistent, practical work on the assembly line.
Comments (0)
No comments yet. Be the first!