A Global Capability Centre is a captive unit a foreign company owns in India to run real work, engineering, product, finance, analytics, R&D, cybersecurity and increasingly AI, rather than outsourcing it. India is where most of the world's GCCs already are, and the centre of gravity has moved from cost-saving back offices to strategic capability that the global business depends on. For a foreign company, setting one up well turns on four decisions that this guide maps, and on running it as a capability rather than a cost line. The economics are real, but so is the execution: the entity, the regime, the transfer pricing and the people all have to be right together.
At a glance
- India hosts more than 1,700 GCCs employing over 1.9 million people, with the market projected to roughly USD 99–105 billion and around 2,200 centres by 2030;
- the usual entity is a wholly-owned private limited company, the cleanest base for talent, IP and governance;
- the location decision pairs a tax regime (SEZ, STPI, non-SEZ or GIFT City) with a city, increasingly a tier-2 one;
- the economics hinge on transfer pricing, where a single 15.5% safe-harbour margin for IT services now applies from 2026;
- the centre runs on people, so employment, payroll and parent-company ESOPs are part of the structure, not an afterthought.
What a GCC is, and why India now
A GCC, sometimes called a global in-house centre or captive, is an offshore unit that a multinational owns and operates itself, as distinct from contracting the work to a third-party vendor. The model has matured in a way that changes the decision. Where the first wave was about labour arbitrage and transactional support, the current wave is about capability: GCCs now host product engineering, data and AI, finance and risk, design, and cybersecurity, and a growing number own a global function outright rather than supporting one. That shift is why a GCC is increasingly a board-level decision rather than an operations one.
India holds the largest share of this activity by a wide margin. As of 2025 the country hosts more than 1,700 GCCs employing over 1.9 million people and generating around USD 64.6 billion in export revenue, having added more than 400 centres between 2019 and 2024 and expanded the talent pool by roughly a quarter. Industry projections put the market at around USD 99–105 billion, with some 2,100 to 2,200 centres and 2.5 to 2.8 million professionals, by 2030. The underlying draw is a deep, English-speaking engineering and professional talent pool at a cost and scale few markets can match, alongside an ecosystem of real estate, vendors and advisers built specifically around the model.
The policy tailwind, and the move to tier-2
The economics increasingly point beyond the established hubs. Tier-2 and tier-3 cities can offer talent costs roughly 20 to 35% lower than the metros and meaningfully lower attrition, and projections suggest close to two-fifths of the GCC workforce could sit in those cities by 2030, with over 220 centres already there. Recognising the sector's weight in the economy, policymakers have moved to attract centres to non-metro states, and several states now run their own GCC incentive policies covering capital subsidies, payroll support and stamp-duty relief.
For a new entrant the practical consequence is that the location decision has two parts that used to be one: which city tier to build in, and which tax regime to register under. A talent-and-cost view favours a tier-2 city; a tax view turns on the regime, and the two are chosen together rather than in sequence. The location piece works through both.
The four decisions you actually make
1. The operating model
Whether to build and own a captive from day one, use a build-operate-transfer arrangement under which a partner runs the centre before handing it over, or stay with managed services. The choice sets how much control you keep, how fast you start, and crucially how intellectual property is owned. For a capability you intend to keep, an owned captive is usually the destination, even where build-operate-transfer is the on-ramp. The model-and-entity piece works through the trade-offs.
2. The entity
Most GCCs are set up as a wholly-owned private limited company under the Companies Act, the cleanest base for employing talent, owning intellectual property and presenting clear governance. The alternatives, a limited liability partnership or a branch, carry real trade-offs for an IP-heavy, equity-incentivised centre, set out in the same piece.
3. The location and regime
The centre's tax and operational position depends on pairing a regime, a Special Economic Zone unit, an STPI registration, a plain domestic company, or a GIFT City IFSC unit, with a city. Each regime carries its own tax benefits and operational conditions: an SEZ unit, for instance, carries strong export-income relief but restricts domestic sales. The location piece sets out the regime choice and the city decision together.
4. Tax, transfer pricing and people
A captive almost always charges its parent on a cost-plus basis, which makes transfer pricing the single most important tax question, and the most audited area in India. From 2026 a revised safe-harbour regime offers an optional single 15.5% margin for IT services. Alongside it sit corporate tax and the employment, payroll and parent-company ESOP arrangements that any people-heavy centre needs. The tax-and-transfer-pricing piece covers all three.
The build, in sequence
A GCC is not a single transaction but a sequence, and most overruns come from treating parallel workstreams as if they were sequential. In outline the build runs: incorporate the wholly-owned company and put in the resident director and the FEMA reporting; choose and register the regime, SEZ, STPI or domestic, and secure the unit approval where relevant; sign real estate; hire the leadership and begin the talent ramp; put the intercompany services agreement and the transfer-pricing policy in place before the first invoice; and then go live and scale. Each of these runs on its own clock, and the entity, the regime approval and the talent market are the three that most often gate the timeline. Sequencing them in parallel, with the transfer-pricing policy set before billing starts rather than after, is much of the practical work.
The economics, in outline
A captive is, in tax terms, a cost-plus entity: it incurs the costs of running the centre and bills the parent those costs plus a markup, and that markup is its taxable profit in India. The headline saving comes from the talent-cost differential, but the net economics depend on the markup the transfer-pricing rules require, the regime's tax treatment of that profit, and the all-in cost per seat once real estate, attrition and compliance are counted. Read the GCC as an owned capability with a defensible cost-plus economic model, not as a place where work simply becomes cheaper; the centres that endure are the ones built and priced on that basis.
Who a GCC suits
It suits a foreign company with enough sustained, core work to justify owning a team rather than buying a service: a technology or product firm building engineering and AI capability, a bank or insurer centralising finance, risk and operations, or an industrial group running design and analytics. Financial-institution GCCs have a particular route through GIFT City, covered in the BFSI piece. A GCC is less suited to small or short-horizon needs, where outsourcing or a managed centre is usually the better answer, and the honest adviser will say so.
Where this goes wrong
- Underpricing the captive. Transfer pricing on a cost-plus captive is the most audited area in India; getting the margin or the documentation wrong invites adjustment and penalty years later.
- Choosing the wrong regime. An SEZ unit carries strong export-income relief but operational restrictions; the right regime depends on the centre's profile, not the headline benefit.
- Treating it as a pure cost centre. The centres that endure are run as capability, with retention, IP and governance planned from the start, not bolted on.
- Underestimating the ramp. Talent, real estate, regime approvals and the entity each run on their own clock; sequencing them in parallel is most of the execution.
- Billing before the policy is set. Invoicing the parent before the intercompany agreement and transfer-pricing policy are in place is a common, avoidable exposure.
How ATB Corporate helps
We set up and structure GCCs for foreign companies end to end: the operating-model decision, the entity and its governance, the regime and location choice, the transfer-pricing policy and documentation, and the employment, payroll and ESOP arrangements. For financial-institution groups, including those based in the UAE and the wider Gulf, we structure the route through GIFT City. The aim is a centre that is defensible on tax and built to retain people, not just one that opens on time.
Talk to ATB about setting up your India GCC →
FAQ
What is a Global Capability Centre?
A GCC, or global in-house centre, is a captive unit a multinational owns and runs in India to deliver work such as engineering, product, finance, analytics, R&D or cybersecurity, as opposed to outsourcing that work to a third-party vendor.
Why do companies set up GCCs in India?
For a deep, English-speaking engineering and professional talent pool at scale and cost few markets match, and increasingly to own a strategic capability rather than only to save cost. India hosts more than 1,700 GCCs, and the market is projected to roughly USD 99–105 billion by 2030.
What entity do you use for a GCC?
Usually a wholly-owned private limited company under the Companies Act, which is the cleanest structure for employing talent, owning intellectual property and governance. The alternatives carry trade-offs covered in the model-and-entity piece.
What is the biggest tax risk in a GCC?
Transfer pricing. Because a captive charges its parent on a cost-plus basis, the margin and its documentation are the most scrutinised area; from 2026 a single 15.5% safe-harbour margin for IT services offers an optional certainty route.
How long does it take to set one up?
It depends on the model, the regime and the talent market, but the entity, the regime approval and the leadership hire are the usual gating items. Running the workstreams in parallel, and setting the transfer-pricing policy before billing, is what keeps the timeline honest.
Captive or build-operate-transfer?
A captive gives the most control and the cleanest IP; build-operate-transfer trades some of that for a faster, partner-run start with a later handover. The model-and-entity piece works through the choice.
Key references
Zinnov–nasscom India GCC Landscape reports; EY GCC outlook; Income-tax Rules 2026 (safe-harbour revisions); Companies Act 2013; SEZ Act and STPI scheme; IFSCA framework for GIFT City.
This article is general information and not tax or legal advice. Laws and IFSCA rules change, and positions should be confirmed for your specific circumstances before being relied upon.