Skip to main content

MicroAGI

Category: Robotics / Embodied AI

German embodied AI research lab and data infrastructure company that collects real-world first-person video of human tasks through its consumer brand Shift to train AI robots. MicroAGI was founded in 2025. The company is led by Bercan Kilic. Based in Munich, Germany. Team size: 11-50. Total funding raised: $5M. Latest round: Pre-Seed. Key investors include Earthling VC, Redalpine, Village Global.

Founded
2025
Headquarters
Munich, Germany
Team size
11-50
Total funding
$5M

Value proposition

Solving the embodied AI "100,000-year data gap" by building the largest, most diverse dataset of real-world first-person video of human task execution, collected via a gig-economy marketplace and promotional free services.

Products and solutions

Shift (consumer-facing app — free NYC home cleaning service using camera-capped cleaners for embodied AI training data), Shift contributor platform (14,000+ operators across 15 countries recording everyday tasks for pay, $5M+ paid out in Q1 2026), Atlas (autonomous robotics for industrial companies), Internal research lab working on VLA (Vision-Language-Action) models and end-to-end physical AGI

Unique value

Offers free home cleaning services to generate high-quality, diverse, real-world embodied AI training data at scale — with proprietary data processing that commands premium pricing from AI labs and robotics companies.

Target customer

AI labs, robotics companies (selling training data); and consumers (free cleaning service in NYC, soon expanding to SF, London, Zurich, Munich)

Industries served

Embodied AI / Robotics training data, Household services, Gig economy data collection

Technology advantage

Proprietary data processing technology that produces higher-quality training data commanding premium prices; automated in-device face/PII blurring using ML models on smart glasses before cloud upload; operates its own research lab for end-to-end physical AGI; founded by former F1 aerodynamic engineers and Alan Turing Institute researcher

How they differentiate

Vertically integrated data pipeline from capture (Shift marketplace/cleaning) to processing (proprietary anonymization tech) to model training (internal VLA research lab); geographic diversity with operations in 15+ countries including underserved markets (Bulgaria, Georgia, South Africa, Turkey); diverse task coverage across messy real-world environments

Main competitors

Scale AI (data labeling for AI, moving into physical world data), Micro1 (human data platform for AI training, $41.6M raised), Encord (data annotation for computer vision/AI)

Key partnerships

Phoenix Court (VC backer/investor), Earthling VC, Redalpine, Village Global

Notable customers

Works with world's leading robotics and AI companies (specific names not publicly disclosed per Webrazzi source)

Major milestones

Founded August 2025 in Munich hacker house by ex-F1 engineers and Alan Turing Institute researcher, grew to 35+ employees by March 2026, launched Shift app with free NYC home cleaning promo going viral with 8M+ views and thousands of bookings (May 2026), 14,000+ operators across 15 countries, $5M+ paid out to contributors in Q1 2026, launched Turkey operations March 2026, plans to expand to SF, London, Zurich, Munich

Market positioning

Early-stage entrant in the emerging market for real-world embodied AI training data, positioned as a full-stack data research and deployment lab covering capture, processing, and model development — operating at the intersection of gig-economy data collection and frontier AI research.

Geographic focus

Global — 15+ countries (Germany HQ, US operations in NYC, with presence in Bulgaria, Georgia, South Africa, Turkey, UK, Switzerland)

About Bercan Kilic

Ex-Formula One aerodynamic engineer at Red Bull Racing & Red Bull Technology; Ex-professional esports player; RWTH Aachen University graduate.

Official website: