How to set up a Global Capability Center, and what changes at mid-market scale

Setting up a Global Capability Center is six pieces of work. Choose a country and an operating model. Register a legal entity. Stand up tax, payroll, and employment compliance. Sign for space and secure the IT. Hire a leader who can run the place. Then ramp the team. Every published setup guide lists roughly that sequence, and the sequence does not change with the size of the team going into it.

That last part is the interesting part, and this article gets to it after the steps. The steps come first, in full, because a company deciding whether to build one deserves the real list rather than a shortened version that flatters the alternative.

The six steps to set up a Global Capability Center

You set up a GCC by building a foreign subsidiary and then staffing it. Here is what those two sentences actually contain. India is the reference market below, because it is where the mechanics are documented publicly and where most of the sector's data lives.

• Location and operating model. Choose the city, then choose how you build. Own and operate it yourself, run a build-operate-transfer arrangement with a partner who hands you the entity later, or start under an employer of record while the entity forms behind it. This choice sets everything downstream, including how long the ramp takes.
• The legal entity. The common vehicle in India is a private limited company, with a branch office or liaison office as alternatives. That means name reservation, incorporation filings with the Ministry of Corporate Affairs, director identification numbers and digital signatures, then PAN, TAN, and GST registrations. Foreign ownership adds FEMA and RBI reporting on top of all of it.
• Tax and transfer pricing. A captive center that serves only its parent is a related party, so the parent pays it on a cost-plus basis at an arm's-length markup. India's safe-harbour rules set that markup for eligible technology and technology-enabled services at 15.5 percent as of 2025. The rate has been revised more than once, so confirm the current notification before modeling anything on it. Transfer-pricing documentation becomes an annual obligation, and tax advisors report that scrutiny has widened from the markup itself to the operating-expense cost base it applies to.
• Employment and payroll infrastructure. Registration under the Shops and Establishments Act, provident fund and state insurance enrolment, gratuity, and a POSH committee. Payroll, benefits, and a working knowledge of Indian employment law all have to exist before the first offer letter goes out.
• Space, security, and IT. A lease in a market you have never operated in, network and endpoint security that satisfies your auditors and your customers' auditors, and a data-protection posture under India's DPDP Act.
• Leadership and the first cohort. Someone senior has to run the site, hold the standard, and represent the parent locally. Most build plans underestimate this one. The center's culture is set by whoever takes that seat, and the search for that person is competitive against every other company building in the same city.

How long a GCC takes to set up, and what it costs

The honest range is wide, and it depends almost entirely on how much you are building at once.

Industry setup guides converge on the same shape. A lean center running one function, with a partner carrying the entity, can reach go-live in roughly three to six months, and build-operate-transfer arrangements are commonly quoted at 90 to 120 days. A full own-and-operate build across several functions runs nine to eighteen months before the center reaches operational maturity. Inside those windows, company formation alone takes up to a month, employment and payroll setup fifteen to sixty days, and a defensible data-protection posture thirty to a hundred and twenty days.

On money, the same guides put an initial build somewhere between $500,000 and $3 million, with legal and company registration running from a few thousand dollars into the tens of thousands, and annual operating cost for a small to mid-sized center between $700,000 and $1.5 million. Those are wide bands on purpose, because they cover very different ambitions. What sits inside them is mostly fixed.

What changes at $50 million to $500 million in revenue

Nothing on the list gets shorter when the company gets smaller. That is the whole difficulty.

A company with three thousand offshore seats and a company with twelve register the same entity, file the same transfer-pricing documentation, sit the same statutory audit, and need the same senior person on the ground. At three thousand people that overhead is noise. At twelve it is the largest line in the budget.

The subtler problem is the one that shows up in year two. A GCC build gets quoted as a project and then behaves like a subsidiary. Once the entity exists, it exists every year: annual filings, a board, an audit, transfer-pricing documentation, and a standing local employment obligation that does not pause when the parent's plans change. A middle-market company that stands one up has not bought a team. It has acquired a foreign legal entity with a team inside it, and an entity gets closed down rather than cancelled.

A GCC build gets quoted as a project. What you sign for is a subsidiary you have to keep alive every year after.

None of this is an argument against the model. The architecture is genuinely good, and forty years of enterprise experience says so. NASSCOM's 2024 landscape report with Zinnov counts more than 1,700 of these centers in India alone, employing over 1.9 million professionals and generating $64.6 billion that year. Companies do not run something at that scale by accident. We covered the era-by-era version of that story in how the model got smaller every decade, and the definitional version in what a Global Capability Center is.

The alternative to a full GCC build

For a company between $50 million and $500 million in revenue, the practical alternative is to keep the architecture and drop the wrapper.

The valuable part of a GCC was never the entity or the lease. It was the design: senior people embedded in one company, holding its context, working its hours and inside its systems, getting better at its specific work every quarter. The entity, the real estate, and the local management layer were the delivery mechanism that design needed in 1997. They were never the thing producing the return.

Separate the two and the decision changes shape. It stops being a question of which country and how many quarters, and becomes a question of which roles are ready to be held by someone senior and who designs the environment around them. That second question is where most offshore attempts are actually decided, and it is the difference between an embedded team and an outsourced one.

That rebuilt version has a name. Kayana creates Mini-GCCs for PE-backed middle-market companies: the same embedded architecture the enterprises spent forty years proving, without the entity, the real estate, or the multi-year build. We design and operate the environment the team works inside, and the design shows up in one number. Voluntary turnover across the teams we build runs 3 to 5 percent in year one. People who stay hold the context, and context is what compounds. The Mini-GCC model is its own deep dive.

Quick answers

The category argument continues in the Mini-GCC model. The full framework is in the book. Operational Alpha →