There is a popular idea that when a product is offered for free, the users themselves become the product. MicroAGI, a startup based in Munich, illustrates this concept well with its innovative Shift app, which launched on May 28. This app delivers complimentary professional apartment cleanings across New York City, but there is a significant catch: cleaners wear head-mounted cameras that record their cleaning activities, including scrubbing countertops, mopping floors, and folding laundry.
The primary goal of MicroAGI is to gather vast quantities of real-world visual data necessary for training the next generation of household robots. Robots require substantial data from real living environments to learn effectively, as simulated settings fail to provide the intricacies of actual human tasks performed in everyday scenarios.
How does MicroAGI utilize the recorded data?
Users in New York can sign up for free cleaning services via the app. During these sessions, the cleaner’s actions are recorded, and the footage undergoes anonymization to remove any identifiable details, such as faces or personal items. This footage is utilized as critical training data for artificial intelligence systems and can be sold to AI and robotics firms seeking high-quality physical task data.
What is MicroAGI's global reach?
While the free cleanings in New York garnered attention, the company's operations extend far beyond just one city. MicroAGI boasts a network of over 10,000 contributors across 15 countries, who perform typical tasks while wearing camera rigs for about $20 per hour. This extensive network yielded over $5 million in revenue in the first quarter of 2026, proving that the business model has significant potential. Established in 2025 and headquartered in Munich, MicroAGI also offers research fellowships, providing extensive computing resources to select industry researchers.
Why is this innovative model important for AI and robotics?
In the wider context, MicroAGI addresses a significant issue in the robotics industry. While AI models like GPT benefit from enormous datasets sourced from the internet, physical AI lacks similar access to diverse, real-world data required for practical tasks, such as handling objects or navigating environments. By transforming ordinary tasks into a scalable data generation system, MicroAGI positions itself uniquely in a market that greatly needs this kind of information.
However, there are notable risks involved. Privacy concerns arise due to the sensitive nature of recording people’s homes. Even with safeguards in place to anonymize data, the scrutiny from regulators is likely, particularly in regions with strict privacy laws like the European Union. Furthermore, the sustainability of providing free cleanings raises questions about the long-term viability of this model. The costs involved in compensating thousands of operators worldwide can mount quickly, making it essential to ensure that the resulting datasets can command high prices in the market.