AI Coding Tools Could Soon Cost More Than Developers’ Salaries

Companies could soon spend as much on a developer’s AI token usage as they do on the developer’s salary, according to Gartner.

The research firm expects AI coding costs to match or exceed the average software developer’s monthly pay within the next two years.

The increase is being driven by wider adoption of generative AI and coding agents, along with a shift from fixed subscriptions to consumption-based pricing.

Based On Global Pay

Gartner’s forecast uses a global average developer salary of about $2,000 per month.

It does not suggest that AI usage will cost more than every developer’s salary, particularly in countries such as the United States, where annual pay can reach six figures.

However, Gartner Senior Principal Analyst Nitish Tyagi said some organisations are already reporting much larger bills.

He has heard cases where a developer consumed $20,000 worth of AI resources in one month, while a business user generated a bill of $32,000.

Costs Are Hard To Control

Gartner issued the forecast to warn companies about what could happen if they fail to monitor and control token usage.

Many AI coding providers do not yet offer mature tools for reducing or managing consumption costs.

Prices may also continue rising as vendors invest in infrastructure and try to make their AI services profitable.

This makes it difficult for companies to forecast spending and calculate whether the productivity gains justify the cost.

AI Agents Increase Usage

Agent-based workflows are particularly difficult to control because they can continue completing tasks and consuming tokens with limited human involvement.

Large context windows can also increase costs when models repeatedly process unnecessary information.

As a result, companies may use their allocated budgets earlier than expected and struggle to explain the resulting expenses.

Many organisations also lack the processes required to measure the return on investment from AI coding tools.

Governance Is Needed

Tyagi warned that AI expenses could rise faster than the productivity benefits without a controlled engineering model.

Companies need to monitor token usage, limit unnecessary context, and determine which tasks require advanced AI models.

Without these controls, rising consumption-based charges could turn AI coding tools into an expense comparable to developer payroll.