Walmart Is Applying Everyday Low Cost To Its Own AI

Walmart has placed a per-associate limit on Code Puppy, an artificial-intelligence tool it built in-house, after demand for it rose sharply. Each employee now receives a fixed allowance of the “tokens” that meter AI computing, in place of the unlimited access they had before, a change Bloomberg reported this week. Speaking to reporters in Bentonville during Walmart’s annual Associates Week, Walmart Global Chief Technology Officer Suresh Kumar pointed to a cause that has less to do with the tool’s novelty than with how people use it. Associates, he said, keep asking it to solve the same problems.

Code Puppy lets employees with no engineering background build simple software and assemble materials like spreadsheets and presentations through plain-language prompts. Kumar told reporters that roughly as many non-engineers now use it as engineers, reaching down to hourly associates, a spread that accounts for both its popularity and its cost. Associates also have access to outside systems including Claude and ChatGPT, Bloomberg reported.

The surge brought a pattern Kumar described as “large classes of problems that people are doing again and again.” Running an identical request through an AI agent costs far more than a conventional search, and Kumar placed the cap inside Walmart’s everyday-low-cost discipline, the EDLC counterpart to its everyday low prices, now turned toward what the company spends on computing. The limit, in his account, is meant to preserve broad AI access while ending the habit of paying repeatedly for answers the company already holds.

What Walmart does with that repetition is the more revealing part. When many associates request the same thing, Kumar said, the company treats it as evidence the capability is worth building once and offering to everyone, rather than regenerating it prompt by prompt. Heavier development stays on a separate platform built for engineers. The token allowance is therefore doing two things at once, holding down duplicative spend and converting aggregate demand into a list of what Walmart should build at enterprise scale.

The Cap Is Internal. The Mindset Behind It Is Not.

Nothing Walmart said this week connects Code Puppy to the agents suppliers and advertisers use. What carries over is the posture. The partner-facing agents Walmart has built over the past year point in the same direction Kumar described, toward efficiency and use that earns its cost. Marty, the partner agent introduced in July 2025 for suppliers, sellers, and advertisers, handles onboarding, order management, and campaigns. Inside Walmart Connect, an agentic advertising assistant tied to Marty is in beta for Sponsored Search, answering plain-language questions on bidding, keywords, and billing, and Walmart Connect has said it is on track to reach every Sponsored Search advertiser in the Ad Center during the first half of 2026.

For the brands on the other side of those tools, the useful signal is the thinking rather than the cap. Walmart Connect has reported that 97 percent of queries to the advertising assistant are unique, which it presents as evidence that advertisers use it for genuine campaign work rather than repeated lookups, the mirror image of the duplication Kumar is pricing out internally. The read does not divide cleanly between first-party suppliers and third-party sellers, since Marty reaches both and the Connect assistant reaches any brand buying Sponsored Search. Supplier and advertiser teams deciding how much of a planning or campaign workflow to route through these agents should assume Walmart is measuring that consumption inside the partner tools as deliberately as it measures it in its own.

None of this signals a retreat. Furner told reporters Walmart is now doing things with AI it would not have imagined attempting five years ago, and the partner agents are scheduled to widen through the first half of the year, not contract. What Walmart put on the record this week was the discipline underneath that expansion, the everyday-low-cost stance that underwrites its prices, now applied to what it spends on AI.

Winning With Walmart

Winning With Walmart is an independent platform for suppliers, sellers, solution providers, and industry experts.

Articles are developed by staff editors, contributing experts, and trusted partners.

Some are researched and drafted with the assistance of AI tools and are reviewed, fact-checked, and edited by Winning With Walmart editors before publication.

Editorial judgment, sourcing decisions, and final approval rest with the publication's human editors in every case.

We are committed to accuracy and fairness. If you believe this article contains an error, we welcome your feedback.