Amazon invests $100M to double its robot manufacturing
Amazon is investing over $100 million in a new Indiana facility to double its robot-making capacity, accelerating the automation of its global logistics network.

Amazon is committing over $100 million to construct a 585,000-square-foot robotics manufacturing plant in Greenwood, Indiana. Scheduled to open by 2028, the facility is expected to create 300 engineering and manufacturing jobs with average annual salaries of nearly $100,000. Working in tandem with another new facility in Texas, this site will effectively double the e-commerce giant's internal capacity for producing robots.
The expansion builds on an already massive automated infrastructure. Amazon has manufactured more than one million robots to date, which currently assist in processing 75 percent of customer orders across more than 300 facilities globally. For logistics and robotics practitioners, this massive scaling effort signals a shift from experimental automation to highly standardized, high-volume industrial deployment, raising the baseline for warehouse efficiency worldwide.
This industrial push contrasts sharply with the nascent state of the broader humanoid robotics market. Recent data compiled by Reuters indicates that global humanoid robot sales reached only 7,000 units last year, highlighting the gap between experimental humanoids and Amazon's practical, task-oriented mobile robots. While startups like AGIBOT have deployed humanoid systems across 100 interaction points at China's Chimelong Spaceship Park and delivered 20,000 production units, most humanoid applications remain focused on novelty and public demonstrations, such as Unitree's G1 models performing on television.
For robotics developers and software engineers, Amazon's aggressive expansion underscores the immediate commercial value of specialized, non-humanoid automation. While companies like PitPro Automation are successfully deploying specialized systems—such as a tire-changing setup that swaps four tires in under 15 minutes—the industry's primary bottleneck remains physical-world data. Startups like HANDO Labs are attempting to solve this by crowdsourcing chore data to train models, but Amazon's massive, closed-loop infrastructure gives it a distinct data advantage that independent developers must work to replicate.
This is our own summary of reporting by Superhuman AI



