Inside Siemens' Testing of AI-Powered Humanoid Robots

AI robot workers are officially here. 

AI-powered humanoid robots have been deployed at Siemens' electronics factory in Erlangen, Germany, marking a significant milestone in the practical application of AI-driven robotics in industrial settings.

The trial, conducted in partnership with AI and robotics company Humanoid, successfully demonstrated how machine learning and autonomous systems can be integrated into industrial logistics workflows. 

Siemens says it represents the first step in a broader collaboration between the two companies to test and validate AI-powered humanoid robots in real-world operational environments.

Stephan Schlauß, Global Head of Manufacturing Motion Control at Siemens, says: "As Siemens' customer zero, the Electronics Factory Erlangen is excited to partner with the Humanoid team.

"We are tackling production automation, discovering new opportunities for Siemens and are eager to advance this promising technology across our factory network to deliver customer value."

AI-driven automation in practice

The proof of concept (POC) was structured in two phases, with the first focused on in-house development and demonstration. 

During this stage, the Humanoid team created a physical twin for testing, optimisation and rapid iteration throughout the development process. 

This approach allowed the AI systems to be trained and refined before deployment. 

The second phase saw a two-week deployment at the Siemens Electronics Factory where partners assessed the robots' autonomous capabilities in a live production environment. 

The robots were tasked with destacking within Siemens' logistics process, autonomously picking up totes from storage stacks, transporting them to conveyors and placing them at designated pickup points for human operators. 

This sequence was repeated until the stack was fully empty, with AI systems managing navigation, object recognition and task sequencing.

Measuring AI performance metrics

According to Siemens, the POC measured both the performance and reliability of the robots under autonomous operation.

The target metrics were met in full and included a throughput of 60 tote moves each hour, operation with two different tote sizes and continuous autonomous task execution for more than 30 minutes.

Success was also measured through overall pick and place success rate and autonomous pick and place success rate, which were both above 90%.

Artem Sokolov, Founder and CEO of Humanoid, says: "At Humanoid, we are a commercially driven company. Our focus is on creating robots that deliver measurable value in real-world settings.

"Working closely with industrial and technology partners allows us to validate our systems against real operational requirements and understand which use cases matter outside the lab.

"This joint POC with Siemens showed clear potential for practical deployment of humanoid robots. We see them move steadily toward the real world, and partnerships like this one help accelerate that transition."

Neural networks and sensor integration

Based in the UK, Humanoid aims to create commercially scalable and safe humanoid robots powered by advanced AI systems.

Artem says: "Our team is developing robots designed to seamlessly integrate into various industries, tackling tasks deemed hazardous or monotonous. This is not about replacing human workers but rather about augmenting our capabilities and freeing individuals to pursue more fulfilling endeavours."

The HMND 01 ALPHA BIPEDAL features 29 degrees of freedom and is equipped with AI-enabling technology including RGB cameras, depth sensors and 6D F/T sensors. The system uses end-to-end reasoning and skills powered by NVIDIA processing, allowing the robot to process visual data and make autonomous decisions in real time.

Humanoid also makes the HMND 01 ALPHA WHEELED, which combines a humanoid upper body with a wheeled platform. This model has been used in a POC with Schaeffler to pick metallic bearing rings in a near-production environment, demonstrating the adaptability of AI systems across different form factors.

Artem explains: "At Humanoid, early POCs are one of our key priorities because they allow us to iterate faster – to go into the real world as early as possible and learn what our future customers truly need."

"This project proved that even at the pre-alpha stage, our platform can deliver tangible value in operational settings. We see high potential for more operational applications with humanoids at Schaeffler's factories."

The company's approach reflects a broader trend in AI development towards rapid deployment and real-world learning rather than extended laboratory testing.