Decoding the 18-Fold Figure
The dramatic '18-fold' increase, highlighted by research firm Gartner, refers to the rising cost required to generate positive returns from AI investments. This finding, discussed at the Gartner IT Symposium/Xpo in Kochi, India, points to a growing gap
between the initial excitement for AI and the harsh economic realities of its implementation. According to Gartner analyst Neha Kumar, while companies are beginning to see some positive outcomes, they must be far more selective about which projects they pursue. The statistic doesn't mean a single AI tool is 18 times more expensive, but that the entire ecosystem of technology, infrastructure, implementation, and oversight needed to achieve a return on investment has become vastly more demanding.
Beyond the Subscription: The AI Iceberg
Many businesses are discovering that the initial software license or API fee is merely the tip of the iceberg. The true cost of ownership is dominated by hidden expenses that accumulate as AI systems are scaled. These include massive infrastructure demands like powerful GPU servers and high-throughput data storage, which can be expensive to provision and even more costly to leave idle. Another significant, often underestimated, cost is data preparation. AI requires clean, structured data, and many firms find their existing data is messy, siloed, and inconsistent, requiring substantial investment in cleanup and management before any AI work can even begin. One analyst notes that hidden data and infrastructure costs can easily double or triple the apparent cost of an AI project.
The Human Factor: Talent and Oversight
AI doesn't just run on data and servers; it runs on people, and the right people are expensive. The demand for specialised talent, from machine learning engineers to data scientists and AI ethics officers, far outstrips supply, driving up salaries. But the human cost doesn't stop there. As AI takes over tasks, work shifts from doing to checking. This requires human oversight to monitor AI outputs for accuracy, bias, and relevance. One study found that up to 40% of expected productivity gains from AI are lost to the time employees spend correcting its output. Furthermore, as AI becomes more integrated, the need for robust governance, compliance, and security frameworks adds another layer of operational cost and complexity.
A Strategic Approach for Indian Businesses
For Indian companies, this news is not a signal to abandon AI, but to approach it with a clear-eyed strategy. The focus must shift from rapid, broad adoption to selective, high-impact projects with a measurable return on investment. Experts advise starting small and avoiding the trap of using the most powerful—and expensive—models for every task when a smaller, cheaper one will suffice. Instead of viewing AI as a simple software purchase, it should be treated as a major business transformation initiative. This involves redesigning workflows and operating models to truly leverage the technology. For many, this will mean moving away from trying to build everything from scratch and instead using pre-trained models or partnering with vendors on outcome-based commercial models, where payment is tied to performance.
















