One-third of companies are skipping commercial software purchases to build custom tools using AI coding agents. While employees report higher productivity, only 37% of firms have seen an improvement in operating profits. This shift in IT spending highlights rising costs and uncertain financial returns, which investors should monitor.
A growing number of businesses are moving away from traditional software providers, choosing to build their own tools using AI coding agents instead. Data indicates that about 33% of firms that use artificial intelligence have canceled plans to buy commercial software, opting for in-house development. This trend is particularly strong in the technology sector, where 41% of companies are bypassing external vendors to create custom solutions tailored to their specific needs.
This shift signals a change in how companies allocate their information technology budgets. For years, enterprises relied on buying licenses for packaged software. Now, the ability to generate code quickly via AI is making it easier for them to build features internally. However, this strategy comes with its own financial challenges. While 80% of employees report that AI coding agents have made them significantly more productive, this efficiency has not yet translated into better company-wide financial performance. Only 37% of surveyed organizations reported a positive impact on their earnings before interest and taxes.
The disconnect between employee productivity and bottom-line profit is largely due to high operating expenses. About 20% of companies identified the rising costs of AI processing—often referred to as token costs—as a major barrier to scaling these projects. These fees for computing power can quickly add up, sometimes offsetting the savings gained from not purchasing commercial software licenses. Consequently, many businesses are finding that the cost of maintaining their own AI-built tools can be higher than expected.
This trend creates a mixed outlook for the IT industry. Traditional software companies that rely on license fees may face pressure if large enterprises continue to pivot toward internal development. Conversely, IT services firms may find new opportunities if companies need help setting up and managing these complex, custom-built AI environments. Despite these challenges, companies are not slowing down their AI investments. Roughly 60% of firms plan to increase their capital allocation toward AI throughout the next fiscal cycle, suggesting they believe in the long-term potential of the technology despite current profitability struggles.
Investors should pay close attention to how this trend affects the margins of both software providers and their enterprise clients. In the coming quarters, the key monitorable will be whether companies can effectively manage the high operational costs of AI development and turn that increased employee productivity into actual profit growth. If these costs remain elevated, firms may eventually reconsider their strategy, potentially shifting back to established vendors if the return on investment remains weak.