According to Gartner's market researchers, increasing token consumption and the shift to consumption-based billing models will make coding with generative AI increasingly expensive. By 2028, costs are expected to exceed the global average developer salary; the forecast is based on a worldwide average of around US$2,000 per month.
"Organizations are rapidly moving from experimentation to scaled deployment of AI coding agents, but many are underestimating the financial impact of rising token consumptionn," explained Gartner analyst Nitish Tyagi. However, more discipline in token consumption will not solely arise from developers' decisions, Tyagi elaborated. In his view, they tend to favor convenience and speed over cost-efficiency. Without a regulated operating model for development, costs in companies could rise faster than the productivity gains that AI tools are intended to achieve.
No ROI, nowhere?
"Software engineering leaders are increasingly concerned as token-driven AI spend becomes harder to justify, with budgets often being depleted earlier than expected," said Tyagi. Among others, Uber's CTO Praveen Neppalli Naga caused a stir in April with the statement that the company's annual token budget had already been depleted. Following this, Uber President Andrew Macdonald added in May on a podcast that the benefit of using AI was also unclear. An increase in useful functions for consumers had not materialized. A similar picture is emerging in Germany, where, according to a Bitkom survey, around a third of the companies surveyed have been surprised by the costs of their AI use.
