That’s according to a report Monday (June 1) by Bloomberg News, which calls the move an example of how companies are rethinking the way they use AI.
Sources familiar with the matter told Bloomberg that Walmart has begun offering a set number of tokens — a unit of AI computing data — each employee can use with its in-house AI agent Code Puppy. Before this, workers had an unlimited number of tokens, the report added.
Code Puppy was introduced to employees as part of the retailer’s push to incorporate AI across its operations to improve efficiency. Staffers have access to platforms like Anthropic’s Claude and OpenAI’s ChatGPT, the sources said.
A Walmart spokesperson told Bloomberg the company wants employees to apply AI to create value, and that it is guiding workers to use the right AI for the right task.
The report noted that companies are seeing the cost of AI as they integrate it into their workflows, particularly when employees use it extensively, or for jobs where large swaths of data are involved.
Advertisement: Scroll to Continue
Uber has reportedly exhausted its annual AI budget already, while Microsoft is apparently scaling back some offerings, the report added.
“The financial friction both companies hit traces back to a structural mismatch between how AI tools are priced and how enterprise finance teams are built,” PYMNTS wrote last week.
As covered here, annual licenses and seat-based pricing offered CFOs a stable cost structure they could forecast. Token-based consumption, where charges accrue based on volume of text processed and generated, cracked that model open.
“A surge in internal experimentation, a new product feature or a poorly optimized prompt can cause costs to spike in ways that are difficult to anticipate,” PYMNTS added.
“Engineering decisions now carry direct balance-sheet consequences that finance teams weren’t structured to track.”
The problem goes beyond billing volatility, as tokens track volume, not outcome. PYMNTS has reported that companies increasingly use token consumption as their main metric for AI adoption and workflow intensity, but that signal has limits.
“A poorly structured prompt that forces a model to iterate and regenerate consumes more tokens than a concise, targeted query, yet neither necessarily produces useful output,” PYMNTS added.
The Bloomberg report noted that Walmart has been ahead of many fellow retailers in adopting AI, implementing it in everything from supply-chain management to customer experience.
