Gartner predicts that by 2028, the rising costs of AI coding agents will exceed the average developer's salary as consumption-based pricing models lead to unpredictable and ballooning expenses.

Key Points

• Gartner reports that monthly AI coding costs per developer are surging from $20 to as much as $5,000 due to token-based consumption pricing.
• Many organizations lack the visibility and frameworks required to track token usage, making it difficult to forecast budgets or justify business value.
• Factors driving costs include autonomous agent tasks, bloated context windows, and a lack of built-in cost controls from AI software vendors.
• Some companies are already reporting monthly AI expenditures as high as $7,500 per employee, prompting firms like Microsoft to adjust their usage strategies.
• Gartner recommends implementing "model routing" to assign simpler tasks to cheaper models and practicing "context engineering" to reduce unnecessary token consumption.

Why it Matters

The shift from flat-rate subscriptions to consumption-based pricing creates significant financial risk for enterprises that fail to monitor AI usage. Without strict cost governance, companies may find that the productivity gains provided by AI tools are entirely offset by unsustainable operational expenses.