BROTECH qualified for USD 10,000 in AWS credits in October 2026. We aim to use the credits to grow the infrastructure that BroBot runs on in AWS.
BroBot itself is an edge AI system: it performs face recognition on its own GPU, so no face image leaves the robot for recognition. The robot's internal network never leaves the robot; the cloud talks to it only through its web API and AWS IoT Core. On AWS we run:
- BroBot Cloud, our fleet platform: AWS IoT Core, AWS Lambda, Amazon DynamoDB, Amazon Cognito, Amazon CloudFront, Amazon S3 and Amazon SES.
- The storage for software releases delivered over the air.
- The off-site copy of the encrypted backups each robot takes every night.
- The cloud simulation farm: AWS Batch on Graviton m8g machines, with AWS Step Functions running the nightly workflow.
The farm went live on 30 October 2026 and runs about 106 scenario jobs every night: people crossing in front of the robot, narrow doors, thin barrier tapes, elevators, docking at the charging station and sensor failures. Its results only count once they agree with a reference desktop machine.
The most important job of this infrastructure is to test every software change at scale before it reaches a robot. A service robot shares corridors, doors and elevators with people, and we would rather catch a fault in the cloud overnight than in the field. As more robots go into service, BroBot Cloud will carry more work. We plan to use the credits to run more scenarios on the farm and to extend the fleet cloud as we prepare for that growth.