Key facts
- Meta is testing robots to perform physical tasks within its data centers, including cable management and server restarts.
- The company is using hardware from vendors such as Watney Robotics, Kinova, and ABB.
- Employees express concern that robots could significantly reduce the need for human technicians.
- Meta acknowledges long-term goals for robotics in data centers but also emphasizes hiring and training initiatives.
- Advancements in AI and robotics hardware are making these applications more feasible and cost-effective.
Meta is actively testing robots to automate tasks within its data centers, a move that could significantly reduce its reliance on human labor and control costs as its investment in AI infrastructure grows. Current and former employees familiar with these projects report that the company is evaluating robots from vendors like Watney Robotics, Kinova, and ABB for duties such as plugging in cables, restarting servers, and performing failure inspections.
One data center worker estimates that a successful robotic implementation could automate up to 80% of certain tasks, leading to widespread concern about job security. Meta has also deployed simpler robots for basic functions like pressing power buttons. While a company spokesperson emphasized Meta's commitment to hiring and training skilled workers, citing a national shortage, a senior manager for robotics previously outlined long-term goals for robot deployment.
Advancements in AI and more affordable hardware have made robots increasingly viable for the precise work required in data centers, overcoming previous challenges like equipment damage. Proponents suggest robots can offer consistent, lower-cost output, particularly in remote areas, and handle repetitive or dangerous tasks, potentially freeing human workers for more complex challenges. This trend could also enable data center operations in extreme environments.
The growing use of automation in data centers has received limited attention from activists concerned about the industry's impact, while tech companies often highlight job creation to secure tax breaks. Employees at Meta's Iowa facility have expressed demoralization, fearing job displacement. Some workers also believe Meta is developing AI software to enable the hiring of lower-skilled, lower-paid staff for tasks requiring only basic adherence to instructions.
Despite these concerns, Meta maintains a strong need for skilled labor and has launched programs to train thousands of individuals annually at no cost, guaranteeing employment in several states and partnering with trade unions. However, as communities re-evaluate data center tax incentives, the potential for job replacement through robotics is becoming a more prominent consideration.
Meta has previously showcased its use of self-driving tugger robots for moving server racks and in-house developed inventory robots. Competitors like Microsoft and Google are also investing in data center robotics, and Amazon is using robots for parts recycling. However, current robotic systems still have limitations, such as the inability of some inventory bots to distinguish between different indicator lights, necessitating continued human oversight for certain diagnostic tasks.