Gartner predicts AI coding costs will exceed developer pay

A new forecast from the research firm Gartner warns that the cost of using artificial intelligence to write software could surpass the average developer's salary by 2028. The prediction, released Wednesday, points to a fundamental shift in how these tools are priced and a widespread lack of visibility into their escalating costs.

The central issue is the industry-wide move from flat subscription fees to consumption-based pricing, where companies pay for each "token" an AI model processes. Tokens are the units of data that power generative AI. As developers integrate AI coding assistants like Claude Code, GitHub Copilot, and others into their daily workflows, token consumption-and the associated bills-are soaring unpredictably.

According to Gartner, many vendors lack transparency in how they calculate and bill for token use, making it nearly impossible for technology leaders to forecast or control spending. Software engineering budgets are being depleted earlier than expected, and leaders are finding it increasingly difficult to justify the expenses. A related survey from KPMG, also released Wednesday, found that just over a quarter of enterprise leaders have complete, real-time...