Gartner is warning that AI coding tools could cost more to run than hiring a developer by 2028, mainly because token usage is rising and pricing is shifting toward consumption-based models. The firm says companies are underestimating how quickly AI coding spend can grow once these tools move from experimentation into everyday use.
Why Costs Rise
The core issue is token consumption. AI coding agents process more tokens as they handle larger tasks, longer context windows and more autonomous workflows, and that directly drives up cost under consumption-based pricing. Gartner says many vendors still do not give enterprises enough transparency into how those costs are calculated and billed.
That makes forecasting difficult. Without clear visibility into usage, organizations can overspend before they realize the tools are eating into budgets faster than the productivity gains justify.
Governance Gaps
Gartner says the cost problem is not only about pricing, but also about how organizations govern usage. Common mistakes include giving agents too much autonomy, feeding them bloated context, and failing to review how much token-heavy work is actually being done.
The company argues that developers often optimize for speed and convenience rather than cost efficiency, so token discipline will not happen on its own. That is why it recommends a governed operating model instead of letting teams use AI coding agents ad hoc.
What Gartner Recommends
Gartner suggests that engineering leaders define when AI coding agents should be used and how much autonomy they should have. It also recommends matching model size to task complexity so smaller models handle simpler work while frontier models are reserved for higher-value tasks.
It further advises teams to practice context engineering, which means giving the model only the information it needs. Gartner also wants organizations to set token thresholds, escalation rules and automated monitoring so usage stays predictable.
Practical Impact
For companies adopting AI coding agents at scale, the message is simple: productivity gains can be real, but so can runaway costs. Gartner’s warning suggests that success will depend less on whether teams use AI and more on whether they use it with discipline.
