Overview
- Meta is trialing robots capable of cable replacement, server restarts, and equipment inspections.
- There are concerns among employees that these machines may lead to reduced staffing or relegation of technical tasks to lower-wage positions.
- The robots are currently slow and depend on human support.
Meta is conducting trials with robots designed to service the data centers that support its artificial intelligence systems. These machines can perform tasks such as changing networking cables, rebooting servers, and examining equipment.
As reported by WIRED on Friday, Meta is collaborating with robotics firms including Watney Robotics from San Francisco, Kinova based in Quebec, and ABB from Zurich. This initiative could potentially lower labor costs and staffing levels as the company's AI infrastructure continues to grow.
Myriad: How high will Nvidia go? Click to make your prediction.An anonymous employee mentioned that a successful cable-swapping robot could potentially take over up to 80% of the tasks currently performed in specific roles. However, this estimate is not an official projection from Meta, and the robot has not yet reached human-level speed.
Introducing Robots into Server Operations
Meta's robots are being tested for various tasks, including moving server racks, inspecting machinery, swapping cables, and rebooting servers. Nonetheless, these machines still require human oversight and face challenges with navigating obstacles, battery life, visual assessments, and complex cabling.
Other companies, like Nvidia and Alibaba, are also working on enhancing robot training capabilities. In June, Alibaba revealed its Qwen-Robot Suite, which includes models for navigation, object manipulation, and physical simulation. Meanwhile, Nvidia researchers have developed AI agents that train robot fleets.
However, the shortage of real-world data continues to hinder the implementation of robots in warehouse settings. ACE Robotics Chairman Wang Xiaogang noted that embodied AI, which allows machines to navigate and interact with physical environments, could experience a “ChatGPT moment” by the end of 2027.
Additionally, companies are exploring robotic applications beyond data centers. For instance, Bedrock Robotics has rolled out autonomous equipment for construction, while Mercedes-Benz has experimented with humanoid robots for repetitive tasks in factories. Proponents highlight labor shortages and safety concerns, while critics warn that such automation may lead to job losses.
The Intersection of Jobs and AI Expansion
Meta’s initiatives in robotics have sparked worries that the data centers supporting AI could ultimately employ fewer individuals. Employees expressed to WIRED that automation might decrease the need for skilled technicians and push remaining responsibilities onto lower-paid workers following AI-generated directives.
A spokesperson for Meta stated that the firm is actively hiring and training employees amid a shortage of skilled labor.
“America is experiencing its largest infrastructure boom since World War II, and there’s a significant shortage of skilled workers to fill these roles; we need more workers, not fewer,” the spokesperson reportedly said.
This concern contrasts with the broader narrative in the industry advocating for automation. Nvidia CEO Jensen Huang envisions data centers as “AI factories” operated by a combination of people, software agents, and robots. Huang and other robotics advocates argue that machines can help alleviate labor shortages and take on repetitive or hazardous tasks.
On Wednesday, Microsoft co-founder Bill Gates suggested taxing robots and AI tokens, claiming that existing tax laws make machines more financially attractive than human employees. While Meta’s robots currently require human supervision and are not yet capable of complex repairs, their evolution could influence whether the company’s growing data center operations create new jobs or displace existing ones.
Meta has not yet provided a response to a request for comments from Decrypt.