Indian services exports in the AI challenge
Business Standard, 2 March 2026
The question
The export of information technology and IT-enabled services stands as a defining triumph of modern Indian economic history. It is the principal domain where India has achieved global competitiveness at scale. These export revenues exert a powerful benign influence upon the Indian economy. The rapid advancement of AI has generated a narrative of decline. There is fear that AI will hollow out the Indian IT/ITES sector. Nobody knows what will happen in the next decade. It's important to try to peer into the future.
Looking back into the data
What has happened thus far in the data? In the December 2025 quarter, the gross inflow for 'other business services' was $32.1 billion, a strong increase from $29.6 billion in the corresponding quarter a year ago (an 8.5% rise in nominal USD over a year). For IT services, the gross inflow in the December 2025 quarter reached $54.3 billion, a strong increase from $48.3 billion a year ago (a 12.2% rise in nominal USD over a year). These growth rates of the recent past seem decent. Let's look into the future and wonder how things might change.
How much will AI tools harm software development in India?
Much of the current anxiety stems from demonstrations of engineers rapidly building complex systems using AI tools. There is a tendency to extrapolate these localized successes into a macroeconomic upheaval. A distinction must be drawn between technology demonstrations and functional enterprise systems. Large organizations operate with immense friction. Enterprise users do not like to disrupt working systems and established contracts. When J P Morgan deploys AI, perhaps they will send more work to TCS.
At a deeper level, the construction of a software system is a process of knowledge creation. We learn how a system works through the toil of building it. When an AI system generates the code, the requisite knowledge creation in the mind of the human author is bypassed. This creates technical debt and cognitive debt [link, link]. Such systems become brittle. When they fail, the human operator lacks the mental model required to repair them. High volumes of effort are required to debug and modify these kinds of systems. The proliferation of low quality AI-generated code might actually generate a lot of maintenance and remediation work for Indian software engineers [Will Wilson on Core Memory].
Everyone thinking about this field has to take a view on how far modern LLMs will go in achieving true human level capabilities, and when. My view is that present systems are not at the level of trustworthiness to match diligent, competent, sustained human work, and it is not easy to overcome the barriers that they face. The economists are still debating whether the computer revolution made a big difference to productivity growth. The transformation of the world economy using these tools will take a while, and it will be filled with labour-intensive tasks that will come to India.
Indian IT firms are not uniform entities. Specialised companies, such as Persistent or KPIT, have constructed moats based on deep domain knowledge and long-standing customer relationships. This embedded knowledge allows them to adapt AI technologies into specific verticals. A general-purpose AI cannot easily replicate the specialised workflows of the global automotive industry that a firm like KPIT understands.
Consider the large incumbents like Infosys. These corporations are embedded deep inside the operations of global enterprises. They manage the complex, unglamorous plumbing of the Fortune 500. A frontier AI company in California cannot easily dislodge this infrastructure. If anything, the market dynamics run in the opposite direction. Frontier AI firms require distribution. To achieve daily revenues from token sales, they need adoption at global enterprises, and the path to this could run through Indian IT giants, the Indian ITES operations and the Indian GCCs who actually run the enterprise IT of the global giants.
The Indian IT/ITES industry has navigated numerous technological regime changes since the late 1980s. Through client-server architectures, the internet boom, and cloud computing, the underlying constant has remained unchanged: the global economy has a thirst for skilled, inexpensive intellectual capability. The AI revolution does not alter this fundamental arbitrage. Yes, of course, a lot will have to change in India, but this is a clever and resilient community that has ridden many a transformation before.
It is important to maintain perspective on the situation. Indian IT currently accounts for roughly two percent of global IT market capitalisation. We must not overstate the size of the incumbent. In a turbulent global environment, there is in fact ample headroom for expansion.
How should boards, investors and leaders of IT firms respond?
There is a genuine technological revolution afoot. We should not get crushed into gloom thinking there is no future for Indian IT/ITES/GCC. We do need to think in rather new ways insofar as business strategy, the board, the top management and the approach to finance is concerned.
For boards, the historical business model was straightforward. It relied on linear headcount expansion and long-term, annuity-style cash flows. That equilibrium is broken. The Indian IT industry that emerges over the next decade will look structurally different. It is the job of the board to engage in strategy thinking that induces a successful transition out of the present comfortable ways.
We do not know how the future will work out. Firm leadership must vigorously explore a wide frontier of possibilities. Enterprise customers may demand small, private language models trained and operated by Indian vendors. Indian software firms may form joint ventures with AI laboratories to implement technologies for enterprise clients. Specialised firms might partner deeply within their domains to create proprietary products. We may see a division of labour where Indian firms build the AI generators and also operate the human-in-the-loop checking mechanisms. Perhaps this technological shock may catalyze a new breed of global product companies emerging from India.
Coping with this uncertainty by developing a portfolio of risky bets requires new kinds of corporate governance. The organisational DNA of traditional IT firms prioritises stability, process adherence, and reliability. The new environment requires a greater slice of innovation, experimentation, and an appetite for risk. Navigating this shift will require a cultural change in the board and in the senior management.
From an investment and corporate finance point of view, big changes are required. Historically, Indian IT firms have exhibited the financial characteristics of stable utilities. There was a bias toward distributing cash to shareholders rather than reinvesting it in R&D. This capital allocation strategy must change. The financial system must transition from pricing these firms as low-risk utilities to valuing them as complex, adaptive, technology integrators.
Management teams need to communicate with capital markets in new ways. Boards must show their portfolio of technological bets to the financial system, outlining the risk and potential return of their AI investments.
Conclusion
The world will always want intellectual capability. The constraint on growth is not the capability of algorithms, but the availability of human ingenuity. The task for India is to double down on producing, owning, and growing intellectual capital, in the emergence of a firm culture of knowledge and innovation.
Back up to the media page for the year
Back up to Ajay Shah's home page