As IT companies hold off on mass hiring and shave off headcount numbers from the middle, industry leaders like Tata Consultancy Services’ (TCS) Chairman N. Chandrasekaran talk about re-evaluating “sub-optimal centres,” that have become “very difficult” to manage.

Small IT Centres to Morph into AI Governance Hubs

For twenty years, if you wanted to build an IT centre in India, you picked one of six cities and one of two job descriptions: “support” or “delivery.” That playbook is dead. The next wave of centres opening in Coimbatore, Jaipur, Lucknow, and Bhubaneswar aren’t chasing cheaper coders — they’re being built to own AI governance for their global parents. If you’re hiring, building, or job-hunting in Indian tech right now, this is the shift you need to understand before it’s old news.

The numbers tell you this isn’t a trend piece

India already hosts over 1,800 Global Capability Centres, more than half the world’s total, employing close to two million professionals and generating $64 billion a year. That’s the base. What’s changing is what these centres are being asked to do.

Over 40% of Indian GCCs are now actively leading their parent organisations’ AI strategy — not executing someone else’s roadmap, setting it. That’s a genuinely new mandate, and it’s arriving at the same moment governments are pouring policy weight behind it: India’s AI Governance Guidelines, unveiled at the AI Impact Summit 2026, stood up new institutions — an AI Governance Group, a Technology & Policy Expert Committee, an AI Safety Institute — built around a “whole-of-government” model that treats AI oversight as infrastructure, not an afterthought.

Put those two facts together and you get the real story: the centre of gravity for AI governance work is moving to exactly the kind of mid-sized Indian IT hub that used to be synonymous with ticket queues and back-office finance.

Why smaller cities, and why now

Three forces are converging, and none of them are temporary.

Metro saturation is real. Bengaluru, Hyderabad, Pune, Chennai, Mumbai, and Delhi-NCR are hitting infrastructure ceilings — congestion, real estate costs, and talent markets so tight that attrition in metros runs 20–25% versus 12–15% in Tier-2 cities. When a GCC role in Coimbatore pays 20–30% less to run and keeps people for years instead of quarters, the math does itself.

Governance work doesn’t need a metro. Unlike frontier model training, which wants dense compute and deep research clusters, AI governance — auditing model outputs, running compliance monitoring, managing agent oversight, building the guardrails a global enterprise needs before it lets an AI system touch customer data — is knowledge work. It needs trained people and clean processes more than it needs a Bengaluru zip code. That’s exactly the profile Tier-2 and Tier-3 cities can now offer: these cities produce roughly 60% of India’s fresh engineering, arts, and science graduates.

State governments are building the plumbing for it. Uttar Pradesh’s 2026 Data Centre Policy is explicit about this — it’s not just chasing data centre megawatts, it’s building AI research clusters and HPC infrastructure into the same corridors, with projects like PRAGATI in Noida signalling that “AI-ready” and “governance-ready” are being treated as the same infrastructure problem. This isn’t one state’s experiment; it’s the template other states are already copying.

What “AI governance hub” actually means on the ground

Strip the buzzword and here’s the job: these centres are moving off routine finance and basic tech support and picking up AI infrastructure oversight, agent governance, and higher-value monitoring work that used to sit at headquarters. Concretely, that’s teams running end-to-end AI risk audits, monitoring model drift and output quality at scale, managing the compliance layer for AI systems operating across multiple regulatory jurisdictions, and increasingly, cybersecurity operations that have shifted from a compliance checkbox to full detection-and-response ownership.

This is the same trajectory GCCs already ran with cybersecurity and semiconductor design — India now has over fifty semiconductor-focused GCCs and full-stack security operations that used to sit only at HQ. AI governance is following the identical arc, just faster.

What this means if you’re hiring

If you’re a recruiter or a founder reading this thinking “great, more centres, more headcount” — that’s the wrong takeaway. The real shift is in what you’re hiring for. A Tier-2 AI governance hub doesn’t want a hundred people who can execute a script; it wants a much smaller bench of people who can read a model’s failure modes, translate a regulatory requirement into a technical control, and make a judgment call on risk without escalating every decision back to a headquarters that’s ten time zones away.

That’s a different hiring funnel than the one built for offshore delivery. It rewards domain fluency — compliance, risk, applied ML — over raw seat-filling, and it’s going to reward the mid-sized cities that invest in that talent pipeline now, before every multinational is fishing in the same pond.

The catch nobody’s advertising

Every glossy report on Tier-2 GCC expansion glosses over one line: infrastructure “readiness” isn’t uniform. A city can have fibre and a shiny IT park and still lack the specialised compliance vendors, legal expertise, and audit ecosystem that AI governance actually depends on. Building physical capacity is the easy 80%. Building the governance ecosystem — the auditors, the legal-tech vendors, the academic partnerships that keep the talent pipeline fresh — is the harder 20%, and it’s the part that will separate the cities that genuinely become governance hubs from the ones that just get a fancier IT park with an “AI-ready” sign on it.

The centres that get this right won’t look like today’s GCCs at all. The ones that don’t will still call themselves “AI hubs” — they just won’t be doing much governing.

The bottom line

None of this is a five-year forecast — it’s already underway, and the window to position ahead of it is closing fast. If you’re building talent strategy for a Tier-2 centre, stop benchmarking yourself against yesterday’s delivery-hub metrics and start benchmarking against governance capability: how fast can your people read risk, not just execute tickets. If you’re a professional in one of these cities, the AI governance skill set — risk auditing, compliance translation, model oversight — is about to be worth more than another certification in a tool that’ll be obsolete in eighteen months. The cities and the people who make that pivot early won’t just catch the AI wave. They’ll be the ones writing its rules.

Read what our Co-Founder and CBO [amongd others] thought on The Hindu Business Line