Indian IT companies are successfully using AI to reduce legacy software issues, but clients in finance and manufacturing now fear that AI adoption itself creates hidden costs and technical risks. For investors, this highlights the need to monitor how IT firms manage AI project profitability amid rising compute and token expenses.
Indian IT services companies have started the current financial year with a dual mission: aggressively reducing clients' old, outdated software issues—often called legacy technical debt—while simultaneously building new artificial intelligence (AI) capabilities. While this shift has led to improved productivity and deal wins, a new concern is emerging among clients in sectors like banking (BFSI) and manufacturing.
The Rise of AI-Spawned Technical Debt
While AI tools are helping companies fix old software faster, leaders in the manufacturing and financial sectors are expressing caution. The concern is that the rapid, automated nature of AI coding could lead to a new form of technical debt. Unlike traditional coding, which is human-readable and easier to track, AI-generated code often lacks full transparency. If developers do not have complete visibility into every line of code produced by AI, it can lead to systems that are difficult to maintain or upgrade later.
Furthermore, clients are wary of the 'probabilistic' nature of AI. In mission-critical environments, such as banking transactions or large-scale manufacturing processes, an AI model that is only 'mostly correct' is not sufficient. These uncertainties, combined with a lack of clear regulatory guidelines, are causing some companies to pause before moving AI projects from testing phases into full-scale production.
Impact on IT Company Financials
For IT services firms, this is a delicate balancing act. While companies like Tech Mahindra have reported strong financial performance—highlighted by a 53% year-on-year increase in EBIT for the first quarter of FY27 and $1.078 billion in new deal wins—the broader sector is navigating a shift in client spending. Management teams at major players like TCS and Wipro have noted that clients are becoming more selective with their technology budgets, prioritizing projects that offer clear, measurable returns.
A major hurdle for IT firms is managing the financial burden of AI deployment. The compute power and 'token' costs required to run advanced AI models can be significant. If these expenses are not managed efficiently, they can erode profit margins for both the IT firm and the client. Companies like Aurionpro Solutions have already begun focusing on AI-native platforms specifically designed to help BFSI clients optimize these costs, reflecting a broader trend where IT firms must now offer 'AI FinOps'—services that track business outcomes against the actual cost of running AI models.
Future Monitorables for Investors
As the industry matures, the focus is shifting from simply adopting AI to proving its long-term financial viability. For investors, the key monitorable is not just the total value of AI deals signed, but how IT firms maintain their profit margins while absorbing the high costs of AI compute and token usage.
Successful firms will likely be those that can guide clients through the 'AI debt' cycle, ensuring that new digital implementations do not create expensive maintenance problems down the road. Moving forward, shareholders may watch for management commentary on how effectively these firms are balancing discretionary spending pressures with the need to build sustainable, high-efficiency AI solutions for their largest corporate clients.