What happened
Walmart Implements Usage Limits Assistant is at the center of this update. Walmart has started restricting employee access to its internal AI tool, Code Puppy, due to unexpectedly high demands on the underlying large language model, prompting new cost-control measures amid evolving AI pricing models.
Walmart Restricts AI Assistant Usage Amid Cost Concerns
Walmart has reportedly begun imposing limitations on the use of its internal AI assistant, Code Puppy, after the demands on the supporting large language model (LLM) exceeded initial expectations. While employees were initially encouraged to use Code Puppy freely for tasks such as spreadsheet analysis and presentation creation, the company is now allotting a fixed number of AI tokens per employee, effectively capping usage.
Transition to Pay-Per-Use AI Models Drives Change
This policy shift reflects a broader industry trend where AI service providers are moving away from fixed-price subscriptions toward pay-per-use models. Such changes increase operational costs for large enterprises, particularly Walmart, which employs approximately 2.1 million workers. Even modest AI usage per employee can accumulate substantial expenses.
Walmart’s guidance encourages employees to leverage AI tools strategically, selecting the most appropriate platform for each task. The company also provides access to additional AI platforms beyond Code Puppy, all funded by Walmart.
Balancing AI Productivity Gains with Financial Realities
Walmart has expanded its AI initiatives, including training 1.5 million associates on AI applications to boost productivity without resorting to layoffs. However, as AI interactions become directly billable, the company faces challenges in balancing productivity improvements against escalating costs.
One contributing factor is the prior practice of measuring productivity by AI usage intensity, which led some employees to maximize token consumption—known as ‘token maxxing’—to enhance performance metrics. While this approach was once encouraged, it now generates higher AI service bills due to the complexity and volume of requests.
More advanced LLMs, particularly those capable of recursive or introspective processing, consume greater token counts per task, raising usage costs. Walmart’s emphasis on careful model selection aims to prevent expensive AI models from being used for routine activities.
Complex AI Workflows and Cost Implications
The use of multi-agent AI workflows, where iterative loops involve several AI agents working toward a goal, can also increase costs unexpectedly. Suboptimal outputs requiring repeated refinements translate directly into higher expenses.
Several AI providers, including Anthropic and OpenAI, have already transitioned to token-based billing for their premium enterprise plans. Microsoft’s GitHub Copilot adopted a paid usage model starting June 1, 2025, reflecting the emerging standard in AI service economics. Similarly, Uber disclosed that its 2026 AI budget was exhausted within the first four months, underscoring the financial impact of these billing changes.
Walmart’s Strategy for Sustainable AI Adoption
By enforcing per-employee token limits, Walmart aims to manage ongoing AI costs, encourage more deliberate use of AI technologies, and establish measurable return-on-investment metrics. This approach represents a pragmatic response to the evolving financial landscape of enterprise AI deployment.
Image source: Pixabay, under license.
Related coverage: AI Chronicle analysis and updates.
Why it matters
This update influences the AI race across model providers, infrastructure leaders, and enterprise adoption decisions.
Meta Secures 1 GW of Solar Power to Fuel Data Centers and Reduce Carbon Emissions
Robots Set New Records at Beijing Half-Marathon, Surpassing Human Performance
IBM Unveils Bob, an AI Platform to Optimize Software Development Costs and Governance
