Amazon announced a $100 million plant in Indiana on Sept. 24 to build the gear that runs its shipping network [1]. Four days later, Walmart committed $300 million to an Ohio building made just for televisions, furniture, and large items that picking robots cannot lift [2].

The big picture:

Both stores want to move orders faster, but their money went to opposite ends of the warehouse.

I keep coming back to how close these two dates landed. Read together, they show a hard truth: you can no longer spread one automation payback number across a network. It splits by what you move.

Amazon has built more than one million robots in the U.S., working across over 300 sites to touch 75% of its shipped orders [1]. Making this gear in-house lowers the cost of each unit. Amazon spends money today to make the next batch of machines cheaper.

The plant itself covers 585,000 square feet and brings 300 jobs that pay close to $100,000 a year on average [1]. It is Amazon’s second robot-making move in two months, after an August build in Texas [1].

Walmart sent its $300 million the other way, building an Ohio site for large, bulky items [2]. An autonomous mobile robot (AMR) moves a small bin, not a sofa, and automated storage towers hold small goods, not mattresses. Walmart is buying concrete and human labor, not robot cells.

It also bought nearly 100 acres for the land, and the new center will sit beside its Monroe, Ohio site and the 200-plus workers there [2].

By the numbers

  • $100M+ — Amazon robot plant: Amazon will put more than $100 million into a large Indiana factory to make gear for its shipping network [1].
  • 1 million+ — Robots built: Amazon has made over one million robots in the U.S. that touch 75% of its shipped orders [1].
  • $300M — Walmart bulky-goods bet: Walmart will spend more than $300 million on an Ohio center for large goods, adding over 300 jobs [2].
  • 3.75% to 4% — Fed policy rate: The Federal Reserve raised rates a quarter point on Sept. 16, raising the bar for any long-term project [3].

What I’d watch:

My read is that stock-keeping unit (SKU) shape drives this choice, not the wage-inflation story that sold the last wave of robots. Uniform goods fit a robot cell, while bulky items still need a building and people.

  • Flexible fleets over fixed towers: Operators renting AMRs under Robotics-as-a-Service (RaaS) deals keep the option to walk away, while a fixed automated storage tower ties the payback curve to one building.
  • Segmented return models: The teams I watch closely are cutting their return models per SKU class. They split robot-ready aisles from oversized aisles instead of using one blended number across the whole network.
  • The rate sets the bar: With the Fed at 3.75% to 4% [3], the shortest payback wins the cash, favoring rented fleets and tight processes over multi-year fixed builds.

The catch

I could be wrong, but this split is not perfectly clean. Walmart already runs robots in parts of its network, and Amazon’s new plant makes parts rather than a finished warehouse.

The honest limit is that neither press release gives a payback figure. This split is my reading of two spending choices, not a published metric from either company.

At a glance

  • The Big Shift: In the same week, Amazon put over $100 million into building its own robots, while Walmart put $300 million into a building for the large goods those robots cannot pick.
  • Why It Matters: Automation payback has split by SKU geometry, meaning one blended return metric across a network now misprices the real costs.
  • What I’d Watch: How leaders change their spending models for different warehouse zones.
  • Flexible fleets: Rented AMR fleets keep the exit open, making a bad bet a simple lease cancellation rather than a stranded tower.
  • Per-SKU models: Splitting the return model by robot-ready and oversized aisles isolates the real payback.
  • The rate bar: A 3.75% to 4% policy rate rewards the shortest payback, favoring rented robots over fixed builds.
  • The Catch: Neither company shared a payback figure, and Walmart already automates elsewhere—this is a read of two spending choices, not a stated number.

Related reading

Sources

[1] Amazon, “Amazon to Create 300 High-Paying Jobs at New Advanced Manufacturing Facility in Greenwood, Indiana” (Sept. 24, 2026) — https://press.aboutamazon.com/job-creation-and-investment/2026/9/amazon-to-create-300-high-paying-jobs-at-new-advanced-manufacturing-facility-in-greenwood-indiana [2] JobsOhio / REDI Cincinnati, “Walmart Plans $300 Million Fulfillment Center in Greater Cincinnati, Creating More Than 300 Jobs” (Sept. 28, 2026) — https://www.jobsohio.com/newsroom/news-press/walmart-plans-300-million-fulfillment-center-in-greater-cincinnati-creating-more-than-300-jobs [3] Board of Governors of the Federal Reserve System, “Implementation Note issued September 16, 2026” — https://www.federalreserve.gov/newsevents/pressreleases/monetary20260916a1.htm