Walmart posted Q2 FY27 on the morning of August 20 and raised its full-year outlook. Revenue of $187.9B, up 5.9%. Adjusted EPS of $0.81, up 19.1%. Full-year sales growth guidance moved to 4%-5% from 3.5%-4.5%, adjusted EPS to $2.80-$2.87.
The market spent the day on $2.9B of tariff refunds and price investment. The load-bearing numbers came later, in the supply chain block of the earnings call.
Three ratios
More than 3,100 U.S. stores are now served by some level of automated freight.
Automated facilities process more than half of the company’s eCommerce fulfillment volume.
Fee-based fast deliveries hit a record 37% of store-fulfilled deliveries. Sub-30-minute delivery covers 38 U.S. markets, and fast-delivery volume rose 48%.
Those three cut the same object at three points. The path a case of goods takes from a distribution center to a shelf to a doorstep is being handed to machines one segment at a time.
The same call carried the growth side. Walmart U.S. comparable sales excluding fuel rose 2.6%, led by transactions rather than ticket. Sam’s Club U.S. comps rose 4.4% on a 7% increase in transactions. Global eCommerce rose 23% and Walmart U.S. eCommerce rose 24%, the tenth consecutive quarter above 20%. U.S. Marketplace sales rose 52%, global advertising 38%, membership income nearly 17%.
The business is growing. It is growing fast.
The denominator is the number that did not move
Walmart employs about 2.1 million people worldwide. Chief people officer Donna Morris has said the total stays roughly flat for the next three years while the mix of jobs changes, and she has put it more plainly than that: six years into her tenure, the company has about the same number of employees it had when she arrived, on much higher revenue.
Set that against the growth line and the mechanism resolves.
U.S. eCommerce has compounded above 20% for 10 quarters. In the retail model that existed before automation, fulfillment headcount tracked units shipped almost linearly. Picking, packing, loading, stocking, cycle counting. Every one of those is labor-dense. Walmart’s version is that volume climbs, the headcount number is held, and the gap is closed by automated freight in 3,100 stores and by automated facilities absorbing more than half the eCommerce flow.
Nobody is being laid off. Walmart announced no store layoffs this quarter, and the company’s public posture all year has been retraining its associates rather than replacing them. Morris has called it unfortunate that other companies cite AI as their reason for cutting.
So there is no severed headcount to count here. What gets absorbed is the hiring that would otherwise have happened.
Lowe’s, one day earlier, is the other half
On August 19, Lowe’s put its first cumulative number on Mylow, its AI shopping agent: roughly 25 million customer questions fielded, and online shoppers who use it converting at three times the rate of those who don’t. The same report showed comparable transactions down 2.1%. We wrote that day that reading the two together inverts the story. The agent did not bring new customers in. It kept the customer who would have driven to a store to ask one question on his phone instead.
Two American retailers, one day apart, and AI is playing opposite roles that arrive at the same place.
At Lowe’s, demand is shrinking and the agent’s job is to hold a smaller pool more tightly. The cost is that the person who has worked the plumbing aisle for 15 years stops being asked.
At Walmart, demand is growing and automation’s job is to catch the overflow. The cost is that the jobs that overflow would have created do not get created.
One hollows out existing roles. The other prevents new ones. Both end with the same result: retail headcount stops tracking retail volume.
Which hours go first
eCommerce fulfillment back-of-house leads. With automated facilities already handling more than half of eCommerce volume, incremental hiring into picking, packing and sortation has stalled. Nearly half of Marketplace volume moved through Walmart Fulfillment Services this quarter, up roughly 400 basis points year over year, which routes that growth into the same automated stack.
Store receiving and stocking is second. Automated freight across 3,100-plus stores rewrites the hours between the dock door and the shelf. That work is the overnight stocking shift, one of the most standard entry points into retail employment in the country.
Delivery dispatch is third. Sub-30-minute coverage across 38 markets with fast-delivery volume up 48% is not a routing problem a scheduler solves by hand. Those decisions already sit in software.
Customer questions are fourth, and barely started. Usage of Sparky, Walmart’s AI shopping tool, rose 70% year over year, and customers who use it spend 40% more per order. Lowe’s 25 million questions show what tools in this category consume: the conversation in front of the shelf.
Walmart has built new roles against this, including an agent developer function that builds internal AI tooling, plus automation maintenance and data work. Those jobs are real. They are not the same order of magnitude as the hiring being absorbed, and they do not start at the same entry point.
The three-year window
Morris named three years. If Walmart’s business keeps compounding at the rate of the past two years through roughly 2029 and the headcount number holds at 2.1 million, the largest private employer in the United States will have completed a displacement at scale without a single layoff announcement.
That process does not enter the Challenger job cut tally. It does not trigger a WARN notice. No internal memo leaks. It shows up in exactly one place: the number of job postings.
It is also the same event we have been tracking all year in a different shape. Globant’s Q2 on August 13 carried 2,673 fewer people and revenue per head up to $95,800, which is IT services raising productivity by shrinking the denominator. Walmart is running the other operation: hold the denominator, let the numerator grow past it. For a financial statement those are different maneuvers. For someone looking for an entry-level retail job, they are the same one.
We covered how the company framed its Silicon Valley tech reorganization in July. That one moved people at headquarters. This one moves no one. It moves the distance between the dock and the shelf.
Sources
- Walmart Releases Q2 FY27 Earnings — Walmart corporate newsroom, August 20, 2026
- Walmart Q2 Earnings Call Highlights — automation and fulfillment ratios from the call
- Walmart CEO wants everybody to make it to the other side — the 2.1 million headcount policy and Donna Morris’s framing
