AI Is Cutting Indian IT in Two. The Halves Are Not Equal

THE RISE OF FOREIGN-OWNED GCCS AND THE QUIET DECLINE OF INDIA'S IT GIANTS

The story told about Indian IT is a story of disruption. Artificial intelligence, the argument goes, is automating the work that built a $224 billion export industry, the code-writing, the testing, the legacy migration, and the national champions that built it are scrambling to adapt. That story is true. It is also incomplete in a way that matters.

Illustration by Geopolitics Next

The more significant shift has been happening for a decade, accelerated by AI, and has received almost no analytical attention commensurate with its scale. As of FY2025, Global Capability Centres, the wholly owned Indian subsidiaries of multinational corporations, and Indian IT firms now contribute equally to India’s technology export revenue. Fifty percent each. That milestone, confirmed by NASSCOM’s FY2025 Strategic Review, marks the end of Indian IT’s structural dominance of India’s technology export economy. It is not a forecast. It has already happened.

The instinctive reaction is concern. The more interesting response is to examine precisely what has changed, and whether the change is unambiguously bad.

TWO BLADES, ONE SCISSORS

The compression facing Indian IT is typically framed as an AI story. Kotak Institutional Equities quantifies over $10 billion in annualised Indian IT revenue at risk from AI-driven pricing deflation alone. HCLTech’s CEO publicly documented contracts repricing for equivalent scope at renewal, a compression he described as structural rather than cyclical. This is real, it is verified, and it is not the most consequential part of the structural shift.

The second blade is GCC insourcing, and it operates through an entirely different mechanism. When a multinational builds a GCC, it does not reprice its outsourcing contract. It eliminates the vendor relationship entirely. The budget line moves from external IT spend to internal operating cost. The revenue does not flow to a cheaper competitor.

It permanently exits Indian IT’s addressable market. HCLTech’s CEO described insourcing as having “deflated the total addressable market for IT services,” the only public acknowledgement from any of the five major Indian IT firms that this mechanism is categorically distinct from pricing pressure. The distinction matters enormously for how the threat is modelled and managed.

What makes this a pair of scissors rather than two parallel threats is the interaction between the blades. AI is not only compressing Indian IT’s existing revenue through pricing deflation. It is simultaneously lowering the minimum viable headcount required to operate a GCC, meaning companies that previously could not justify the fixed investment in building a captive centre are crossing the economic threshold for the first time.

The addressable market for GCC formation is expanding precisely as Indian IT’s core revenue base is contracting. The hinge between these two movements is AI itself. One dynamic produces the other.

The State Street and HCLTech case is the clearest verified instance of the second blade landing. When State Street dissolved its joint venture with HCLTech in late 2023, consolidating the operation as a wholly owned subsidiary, HCLTech recorded an impact of approximately 0.8 percentage points on its financial services vertical revenue across consecutive quarters. State Street described the move as enabling the transformation of its operating model and the unlocking of efficiency savings. It was not a contract cancellation. It was a structural reclassification.

WHAT A GCC ACTUALLY IS?

The compression story only makes sense if the two models, the Indian IT outsourcing firm and GCC, are competing on equivalent terms. They are not. A GCC is not a cheaper TCS. It is a categorically different entity. As a wholly owned subsidiary, it sits inside the parent’s organisational chart, attends global product and strategy meetings, and operates with full access to proprietary data.

When AI tools reduce engineering hours by up to 55 percent in controlled studies, consistent with experimental evidence from Brookings Institution research, the productivity gain flows directly and entirely to the parent’s P&L. The GCC model gets structurally richer from AI with every cycle.

An Indian IT vendor faces the same efficiency gains but a fundamentally different incentive structure. Surfacing AI productivity to the client invites repricing at the next contract renewal. The commercial logic of the outsourcing model creates a structural disincentive to foreground efficiency. This asymmetry is not a management failure. It is embedded in the architecture of the relationship.

And it compounds: over time, the entity with full data access, aligned AI incentives, and institutional knowledge built within a single legal entity will pull progressively further ahead of the entity that sits behind a procurement layer, works on anonymised data subsets, and rotates talent across accounts at 13 to 22 percent annual attrition.

The GCC model also generates a form of strategic lock-in that outsourcing contracts never could. Fixed infrastructure, multi-year talent commitments, and institutional knowledge built inside a parent’s legal entity are considerably harder to unwind than an outsourcing contract cancellable at 90 days’ notice.

When 174 of the Fortune 500 have physical infrastructure and talent relationships in India, they have built structural dependencies that run in both directions. The switching cost is not trivial, and it runs against the narrative that India is in a position of pure dependency.

THE GENUINE UPSIDE, AND ITS LIMITS

The GCC boom is not bad for India. It is important to state this without qualification before examining what it is not. GCC employment in India now stands at 2.36 million professionals, generating gross value added per employee of approximately $32,500, eleven times India’s national GVA per capita, comparable to South Korean productivity levels.

More than 70 percent of GCCs operating in India are multifunctional hubs where teams exercise genuine product ownership rather than executing instructions from abroad. The engineer at a Goldman Sachs or Microsoft GCC in Bengaluru is not performing the same function as the engineer at a mid-tier Indian IT firm rotating through commodity application maintenance.

They are inside the global strategic conversation, trained not just in technical skills but in how world-class institutions solve problems that matter to them competitively.

That is a generational upgrade in the quality of human capital development happening on Indian soil.

The talent premium GCCs pay, 20 to 40 percent above Indian IT market rates, pulls up the compensation floor across the entire technology labour market, benefiting engineers who never set foot in a GCC. The geopolitical embeddedness created by 2,117 foreign-owned technology operations across India’s major technology cities generates private-sector constituencies for stable bilateral relationships in Washington, London, and Brussels that no government initiative could manufacture.

Bengaluru, home to over 880 GCCs, produced 8 of India’s 10 new unicorns in 2024 and 2025. The geographic co-location is not coincidental.

The density of world-class institutional knowledge in one city, built across thousands of engineers working inside global strategic decision-making processes, is the raw material from which the next generation of Indian technology founders will most plausibly emerge.

The precedent is documented elsewhere: Israel’s 434 multinational R&D centres employ one third of its technology workforce, contribute 40 percent of its R&D expenditure, and have produced a cyclical talent flow, from MNC centre to startup to senior role and back, that Startup Nation Central describes as foundational to Israel’s standing as the world’s most startup-dense economy per capita.

Unit 8200, the Israeli military’s signals intelligence unit, produced Check Point, Waze, Wix, and Wiz among dozens of others: an institution that trained people in high-stakes technical problem-solving at world-class standards before releasing them into the commercial ecosystem. GCCs are not the same institution, but the structural logic is the same.

Concentrated exposure to global-standard technical and strategic work, inside institutions with genuine stakes in the outcome, builds the kind of human capital that eventually re-enters an economy as founders rather than employees. India has not yet produced the documented alumni-to-founder pipeline at the scale Israel has. But Israel took decades to compound that flywheel. India is a decade into building the conditions for it.

The honest complication is distributional. GCC employment is concentrated in four metropolitan areas, skewed toward mid-to-senior professionals, and structurally oriented toward the top quartile of India’s technology workforce. The Indian IT outsourcing model, for all its operational limitations, performed a function the GCC model cannot replicate: it absorbed mass quantities of engineering graduates from tier-2 and tier-3 cities, trained them in standardised processes, and provided a first career rung at scale.

TCS hiring tens of thousands of freshers annually was not merely a business model. It was a social mobility pipeline. As that pipeline compresses, TCS made its largest layoff in company history across FY2026, while the broader sector’s net hiring approached zero; no replacement mechanism has been constructed for the cohorts whose first rung is disappearing.

India’s government is responding rationally on the metrics it is measuring. The Karnataka GCC Policy 2024-2029, MeitY guidelines, and SEZ incentives offering full income tax exemption on export profits for five years are all designed to attract more GCCs. That is the correct policy response if the objective is FDI, GDP, and productivity. It is an incomplete policy response if the objective also includes managing the distributional consequences of a structural transformation in the sector that has been India’s primary engine of middle-class formation for thirty years. No published government framework distinguishes the macroeconomic character of GCC export revenue from Indian IT export revenue. The headline number grows. The questions beneath it remain unasked.

THE TRADE THAT WAS NOT CONSCIOUSLY MADE

India stumbled into the outsourcing miracle. The 1991 liberalisation, the Y2K demand surge, the English-language engineering graduate surplus: these were structural accidents that became deliberate policy only after the value they created became visible. India built world-class institutions around what had emerged organically. It is repeating the pattern with GCCs, attracting them rationally, benefiting from them genuinely, and constructing no framework for what is being exchanged in the process.

What is being exchanged is specific. The value created by Indian technology talent is increasingly world-class. The ownership of that value, the intellectual property, the client relationships, the strategic decision-making, the capital returns, remains abroad. India captures the labour income from its technology economy.

The capital income compounds elsewhere. That is not a sovereignty crisis. It is a structural choice about what kind of technology economy India is building. And it is a choice that has not been examined as one.

The question this poses is not whether GCCs are good or bad for India. The evidence on that is genuinely mixed, and anyone who resolves it cleanly in either direction is simplifying a complex distributional reality. The question is whether India is consciously managing the most significant structural transformation of its most important export industry in a generation, or whether it is, once again, ratifying an outcome after the architecture has already been built.

India has the talent to be the world’s best technology location. It also has the foundations, if it chooses to build them, to produce the next generation of technology principals rather than the next generation of premium technology employees. The 50/50 split is not a verdict. It is a moment at which a choice, previously unmade, becomes visible. What India does with that visibility is the question the next decade will answer.

---

BY GEOPOLITICS NEXT

CURATED BY TEAM GEOSTRATA

info@thegeostrata.com