Sometime in the last eighteen months, a decision quietly moved. At a growing number of Fortune 500 and Forbes Global 2000 companies, the call on which product ships next, how a new AI feature gets built, or which market gets an engineering surge, no longer originates at headquarters and travels outward as an instruction. It originates in Bengaluru, Hyderabad or Pune, and travels the other way. Nobody announced this. There was no press release, no reorganisation memo with a new org chart attached. The center that was supposed to execute the strategy has, in a growing number of cases, become the place where the strategy gets made.
This is the real story inside India’s Global Capability Centre boom, and it is a more interesting one than the headcount numbers suggest, though the headcount numbers are themselves striking. According to the Zinnov-Nasscom India GCC Landscape Report published in March 2026, the country now hosts 2,117 GCCs operating across 3,728 units, employing 2.36 million people and generating $98.4 billion in revenue, up 32 percent since FY2021. Some 506 of the world’s Forbes Global 2000 companies now run a capability center on Indian soil. Those are the numbers a CFO reads in a board pack. They are not, on their own, the reason a CEO should be paying closer attention right now.
The mandate changed faster than the org chart
What should worry, or at least occupy, a chief executive is a different figure buried deeper in the same report: 96 percent of GCCs established since FY2021 launched with product or portfolio ownership from day one, skipping the decade-long “crawl, walk, run” progression that defined the first two waves of Indian offshoring. Centers that once graduated slowly from ticket resolution to application maintenance to, eventually, limited product work now open with a mandate to own outcomes. Nasscom’s president, Rajesh Nambiar, put the shift plainly at the organisation’s 2026 GCC Summit: the operative question for a multinational is no longer where work can be done cheapest, but where the enterprise can build resilient, senior-owned capability that can actually run something end to end.
That is a genuine structural change, not a rebrand. Zinnov’s own maturity framework now classifies 39 percent of Indian GCCs as “Portfolio Hubs” with end-to-end ownership from India, and a further 5 percent as “Transformation Hubs” carrying AI-led operations and CXO-level mandates housed entirely in the Indian center. Put differently: a meaningful slice of the largest companies on earth now has senior executive decision-making, not just delivery capacity, sitting eight and a half hours ahead of New York and four and a half ahead of London.
Here is the part that should trouble a board, though: authority has a habit of lagging capability, and nobody has fully reconciled the two. HSBC’s India-based global services leadership speaks of the center being valued for “transformation, resilience, enterprise leadership and strategic influence” rather than scale. That is the rhetoric. The practice, according to executives tracking the sector closely, has not caught up. The gap identified repeatedly at this year’s summit was blunt: companies now talk about Indian centers owning products and making AI-driven decisions, but relatively few have actually rewritten the governance structures, the escalation paths, the equity participation or the P&L accountability that would make that ownership real rather than delegated in name only.
That mismatch is not a paperwork problem. A center that is functionally running a global product line, but whose leadership still reports through a cost-center lens, still gets measured on SLA and unit economics, and still cannot make a call without a sign-off routed through a headquarters time zone, is a center primed for exactly the kind of attrition and disengagement that shows up two years later as an unexplained talent crisis. The people capable of running Transformation Hub-grade work are, not coincidentally, the people with the most external options.
The pipeline that built this system is being switched off underneath it
The second, less discussed tension is generational, and it is being created by the same technology the GCCs are now mandated to own. India’s capability centers did not become deep enough to run global products by hiring senior architects from abroad. They became deep enough by running one of the more effective apprenticeship systems in the modern services economy: enormous entry-level cohorts, trained over years into mid-level and then senior engineers, product owners and eventually center heads. That pyramid, not any single policy decision, is what produced the 2.36 million professionals now on the books.
AI is compressing exactly the base of that pyramid. Work that trained a first-year analyst, code review, basic quality assurance, documentation, first-pass reconciliation, is precisely the work generative tools now do fastest and cheapest. The Observer Research Foundation’s recent analysis of the sector frames this correctly as a task-reallocation problem rather than a wholesale jobs problem: demand is rising sharply for AI product engineering, model governance and domain-led transformation roles, even as the traditional entry-level, FTE-based delivery model gets squeezed. But a task-reallocation problem still leaves an open question nobody in the GCC boom is answering out loud: if you stop training juniors at the rate you once did, where does the senior specialist layer come from in 2031, when today’s cohort would have matured into it?
A labour market splitting in two
That question is sharpened by what is happening to compensation, because the GCC boom and the slow-motion crisis at India’s traditional IT services firms are the same story told from opposite sides of the same balance sheet. Salary data for 2026 shows the split clearly: at the major listed IT services companies, average hikes for the broad workforce have settled around 7 to 8 percent, below the national average, and after inflation amount to a real increase of roughly 2.5 to 3 percent. The stated reasons are structural, not cyclical: top-five IT services revenue growth slowed to 4 to 6 percent in FY26 from 9 to 14 percent two years earlier, and AI tooling has absorbed an estimated 8 to 14 percent of internal engineering productivity without a matching increase in what clients are willing to pay for it.
GCCs are living a different reality. Broad-base hikes there run 9 to 13 percent, AI and machine learning specialists are seeing 16 to 28 percent, and at India’s homegrown product companies, genuinely scarce GenAI talent is commanding 22 to 40 percent increases. Compensation analysts tracking the sector put the overall gap at 1.7 to 2.4 times the hike an equivalent engineer receives at a services firm. A senior AI engineer at a GCC in Bengaluru can now expect total compensation in the ₹78 to ₹120 lakh range at the principal level, a scale that was, until recently, reserved for a small cadre of returning diaspora talent.
This is where the two tensions collide. The IT services sector has historically been the farm system, the place where a large share of India’s engineers cut their teeth before some of them moved into product companies or GCCs. If services firms cannot compete on compensation for AI-relevant skills, and simultaneously reduce entry-level intake because AI has eaten the work that once justified hiring juniors, the ecosystem that has fed the GCC boom for two decades starts to thin from the bottom at precisely the moment GCCs need more of it from the top. The GCC sector is, in effect, drawing down a training system it does not own and has done little to replenish.
What this actually means for a CEO’s desk
None of this is an argument against the India GCC model; the economics remain compelling and the strategic logic, done properly, is sound. But it reframes the decision from an operational one, handled by a chief people officer and a real-estate committee, into a capital-allocation and governance question that belongs at board level. Three things follow. First, if a center has been handed Portfolio Hub or Transformation Hub-level ownership, its governance, equity participation and decision rights need to be redesigned to match, not left running on a cost-center operating model with a strategic label pasted over it; the gap between rhetoric and authority is where the best people leave first. Second, entry-level hiring at the GCC itself should be treated as a long-duration investment in the specialist bench a company will need in five to seven years, not a line item to be trimmed for the same productivity arithmetic that is hollowing out the services sector. Third, and most uncomfortably, executives should stop asking whether their Indian center is saving money and start asking who, in practice, is making the decisions that used to be made at headquarters, and whether the organisation’s formal authority structure has caught up with that fact. In a growing number of the world’s largest companies, it quietly hasn’t, and the distance between where decisions are actually being made and where the org chart says they are made is not a detail. It is the risk.
