Decoding the 18-Fold Figure
The explosive '18-fold' increase refers to how much more companies are spending on AI now compared to two years ago. This startling statistic was highlighted at Gartner's IT Symposium in Kochi in November 2026. It reflects a market shifting from small-scale
experiments to broad, enterprise-wide AI deployment. While more than half of organisations are seeing some positive results from AI, the cost to achieve those gains is rising dramatically. According to Gartner analyst Neha Kumar, this signals a major change in the economics of AI adoption. The initial excitement is now meeting the hard reality of what it actually costs to make AI work not just technically, but commercially.
The Hidden Iceberg of AI Costs
The initial license or subscription fee for an AI model is just the tip of the iceberg. True AI costs, or the Total Cost of Ownership (TCO), can be three to five times higher than initial quotes. The bulk of expenses lies in hidden or underestimated areas. These include the immense cost of data preparation, cleaning messy data, and integrating AI into existing, often clunky, legacy systems. There are also significant 'people costs'. The demand for AI talent has driven salaries for these roles up to three or four times higher than average, yet the skills themselves have a short shelf life. Furthermore, ongoing operational costs like model monitoring, maintenance, retraining to prevent 'drift', and governance are substantial and often unbudgeted. Even variable costs like token usage for generative AI can lead to surprise bills as adoption scales.
Why AI Returns Are So Elusive
Even with massive spending, a clear financial return remains frustratingly difficult to pin down. According to one Gartner survey, while 77% of board members expect cost savings from AI, 97% of organisations have failed to translate productivity gains into actual savings. The problem isn't always the technology; it's often organisational. Many of AI's benefits are indirect and long-term, such as better decision-making or improved customer experience, which don't show up immediately on a balance sheet. There's also the 'productivity J-curve' to consider: productivity often dips initially as employees learn new tools and workflows before rising over the long term. This makes it hard for leaders to justify the investment when short-term financial impact is what stakeholders demand. As a result, many companies struggle to prove value, with only about 29% of executives reporting they can measure AI ROI confidently.
A Smarter Path to AI Value
The spiraling costs and elusive ROI don't mean AI is a bad investment. Instead, it signals a need for a more strategic approach. Experts advise moving away from a 'fear of missing out' mindset and toward projects with clear, measurable goals. This starts with being more selective about which AI initiatives to pursue. Business leaders must look beyond simple productivity metrics like 'time saved' and connect them to concrete financial outcomes like reduced operational expenses or new revenue streams. This involves redesigning workflows and operating models to truly leverage the technology, a priority for 86% of CEOs in one Gartner survey. Implementing stronger governance and cost controls, such as usage limits and cost-per-run tracking, is also critical to prevent budgets from spiraling out of control. Ultimately, success requires treating AI not as a simple IT upgrade, but as a fundamental business transformation that requires careful planning, realistic budgeting, and a long-term vision.
















