A Global Capability Centre is a captive: your entity, your employees, your P&L. That single fact separates it from every vendor arrangement and explains both its advantages and its costs.
One disambiguation first, because it causes real confusion in this market: GCC here means Global Capability Centre. In a regional context the same letters mean Gulf Cooperation Council.
What you are actually taking on
Not a team. A company.
- A legal entity — incorporation, directors, registered office, statutory filings that continue whether or not the centre is busy.
- Employment obligations under Indian law: provident fund, gratuity, statutory leave, notice periods, and the compliance calendar that goes with them.
- Facilities, and increasingly a hybrid policy that has to be decided rather than inherited.
- A hiring function that works in a market where the competitive dynamics are unlike the parent's.
- Local leadership with real authority. A centre run entirely by remote decision becomes an order-taking unit, which is the most common way GCCs underdeliver.
None of this is difficult. All of it is work that has nothing to do with the engineering you set the centre up for, and it is the part organisations consistently underestimate.
The economics
A GCC has the lowest steady-state cost per engineer of any model and the highest fixed cost. That is the whole trade.
| Fixed cost | Cost per engineer | Breaks even | |
|---|---|---|---|
| Staff augmentation | None | Highest | Immediately |
| Dedicated pod | Low | High | Quickly |
| ODC | Moderate | Moderate | Around fifteen engineers |
| GCC | High | Lowest | Commonly cited near fifty, over years |
The number matters less than the horizon. A fifty-person team needed for eighteen months does not justify an entity; a twenty-person team needed for a decade might.
Where to put it
The choice is a trade between depth and stability rather than a ranking.
Bengaluru and Hyderabad have the deepest pools — you can hire almost any specialism — along with the highest salary inflation and attrition, because every other GCC is hiring there too.
Pune and Chennai trade some depth for noticeably better retention, and both have strong engineering bases tied to manufacturing and services respectively.
Tier-two cities offer materially lower cost and the best retention. The constraint is the senior market: hiring a principal engineer or an experienced engineering manager is genuinely harder, and that constraint usually decides the question for a centre that needs its own technical leadership.
The honest framing is that you are choosing which problem to have.
The two ways in
Direct setup
Incorporate, hire a country lead, build from there. Full control from day one, and appropriate if the parent already has India experience or an existing entity.
The risk is sequencing: you are hiring your first engineers before you have a manager who knows the market, or hiring a manager before there is anything to manage. That gap is where the first year goes.
Build-Operate-Transfer
A vendor builds and runs the unit under its own entity, then transfers it on a fixed date at a fixed price. You get delivery from month two rather than month twelve, and the entity is incorporated when the capability is proven rather than before.
Usually cheaper in the first two years, and it defers the irreversible decision until you have evidence. The clause that matters is the transfer terms being fixed at signature — where they are left open, it is outsourcing with an implication.
The failure mode nobody plans for
Not cost, and not quality. Mandate drift.
A centre set up to own a product ends up doing maintenance for the parent's teams, because that is what gets asked for in the first year while capability is being proven. Two years later it is a support function with a hiring problem, because nobody senior wants to join one.
The defences are structural: give the centre end-to-end ownership of something from early on, put a decision-maker in it rather than a coordinator, and measure it on outcomes rather than on tickets closed.
The question that comes first
Not which city or which model. Is this capability permanent and core?
If yes, a GCC is worth its overhead and BOT is usually the lower-risk route to it. If you are not certain, an ODC gives you most of the delivery benefit and none of the statutory commitment — and an entity built for a capability that turns out to be temporary is expensive in a way that takes years to unwind.