Gartner: AI coding costs to exceed average developer salaries by 2028

The cost of using AI coding tools is on track to surpass the average software developer's salary by 2028 as organizations scale adoption of AI-powered coding agents and consume increasing volumes of large language model (LLM) tokens, according to new research from Gartner.

The analyst firm warned that many organizations are underestimating the financial implications of moving from pilot projects to enterprise-wide deployments of AI coding assistants, particularly as vendors increasingly adopt consumption-based pricing models tied to token usage.

"Organizations are rapidly moving from experimentation to scaled deployment of AI coding agents, but many are underestimating the financial impact of rising token consumption," said Nitish Tyagi.

AI tokens are the units of data processed by generative AI models and form the basis of pricing for many AI coding platforms. As developers increasingly rely on AI tools for code generation, debugging, testing and software development tasks, token consumption — and associated costs — can rise sharply.

According to Gartner, the industry's transition away from traditional seat-based licensing toward consumption-based pricing is creating a new challenge for software engineering leaders: cost predictability.

Many AI coding vendors provide limited visibility into how token consumption is measured and billed, making it difficult for enterprises to accurately forecast spending or assess return on investment. Gartner said the lack of transparency increases the risk of budget overruns as organizations expand AI-assisted software development programs.

"Most organizations still lack the maturity and frameworks to effectively measure cost versus business impact," Tyagi said. "Software engineering leaders are increasingly concerned as token-driven AI spend becomes harder to justify, with budgets often being depleted earlier than expected."

The report identifies several factors contributing to rising costs, including unrestricted use of autonomous coding agents, excessive context windows that increase token consumption, and inadequate feedback mechanisms to optimize AI usage. Gartner also noted that most AI coding platforms have yet to introduce mature cost-optimization capabilities, leaving enterprises responsible for managing consumption.

The analyst firm expects costs to continue climbing as AI model providers seek to recover growing infrastructure investments and improve profitability. At the same time, developers who currently use AI tools occasionally are expected to become heavier users as AI becomes embedded into everyday software engineering workflows.

To avoid runaway spending, Gartner recommends that organizations establish formal governance frameworks around AI-assisted development.

Among its recommendations are defining when AI coding agents should be used, matching model size and capability to task complexity, implementing intelligent model-routing strategies, and introducing token monitoring and spending thresholds.

The firm also advises organizations to adopt "context engineering" practices that train developers to provide only relevant information to AI systems, reducing unnecessary token consumption without compromising output quality.

Regular reviews of high-token-consuming workflows should become part of software development cycles, Gartner said, enabling teams to identify inefficiencies and share best practices across engineering organizations.

The forecast highlights a growing tension in enterprise AI adoption: while AI coding assistants promise productivity gains and faster software delivery, organizations may soon face a new challenge in ensuring those benefits outweigh rapidly escalating operational costs.