10decoders
The approved budget and the real budget are different numbers
The approved GCC budget is usually built from three line items: incorporation, office, and a headcount model taken from an average-salary database. The real budget has twelve line items. The difference between them is not bad math. It is that setup cost and operating cost are treated as separate phases, so the items that belong to both — the IT integration with the parent company, the payroll stack, the audit and transfer-pricing retainer, the attrition buffer — fall through the gap and surface later as unbudgeted surprises.
In 2026, a typical mid-market GCC of 20 to 40 seats in Bengaluru, Chennai, Hyderabad or Pune costs between $500,000 and $1.2 million in total year-one spend, of which the setup capital is usually $120,000 to $250,000 and the rest is salaries, benefits, facilities and compliance. The setup capital is the number most business cases quote. It is also the least useful number, because it excludes the single largest cost driver: what it actually takes to hire and retain the team.
The most expensive version of this mistake is the center that launches under-budgeted and is then forced to make decisions in month six that it should have made before signing the lease — freezing hiring to protect the budget, or pushing salaries below the band that keeps attrition under control. Attrition is the one cost that compounds: every point of attrition above 15% annually adds roughly 8 to 12% to year-one effective payroll once recruiting, onboarding and lost productivity are priced in.
"The number that wins board approval is setup capital. The number that determines whether the GCC survives is total year-one cost. They are not the same number."
The 2026 GCC setup cost breakdown, line by line
The twelve line items below are what a complete year-one cost model contains. The ranges are for a 20–40 seat center in a tier-1 city with a standard engineering/delivery mix. They assume a mix of senior and mid-level hires, a 12-month lease with a 3-month deposit, and no unusual visa, security or regulatory complications. Every line is negotiable; none of them are optional.
| Cost Line Item | What's Included | Typical Range (2026) | Risk if Skipped |
|---|---|---|---|
| Entity incorporation & regulatory setup | Private limited company formation, RBI and FDI compliance, GST and professional tax registration, DIN/DSC, PAN/TAN, opening the bank account, and the statutory registrations a foreign parent needs before hiring the first employee | $4K–$12K | Critical |
| Office lease & fit-out | 12-month rent plus security deposit, fit-out or license fees in a managed space, workstations, meeting rooms, pantry and common area costs. A managed/serviced space trades a higher monthly rate for near-zero capital fit-out | $35K–$110K | High |
| IT infrastructure & security baseline | Laptops and monitors, network and internet with redundancy, device management, SSO and MFA, endpoint protection, backups, and the security policies a parent company's infosec team will require before granting access | $25K–$60K | High |
| Hiring & onboarding | Recruiter fees or internal recruiting cost, background verification, offer-to-joining logistics, and the first-quarter onboarding and training ramp for a 20–40 seat team | $10K–$30K | High |
| Salaries, benefits & statutory contributions | Year-one payroll for the engineering, delivery, and operations team, plus PF, ESIC, gratuity, medical insurance, and the 10–12% annual increment most offers assume. This is 50–65% of total year-one cost and the one number that must come from offer letters, not salary databases | $300K–$800K | Critical |
| HRMS, payroll & time systems | HRMS/payroll platform, time and leave tracking, expense management, and the ITSM tool the parent requires. Many parents double this cost by licensing a parallel corporate stack for the Indian entity instead of a local equivalent | $15K–$60K | Medium |
| IT integration with the parent | VPN or VDI connectivity, identity federation with the parent's directory, data-residency review, and the security review and audit that gates access to parent systems. Almost always 2–4 months longer than the business case assumes | $15K–$70K | High |
| Legal, tax & audit | Annual statutory audit, transfer pricing study and documentation, tax filings, legal retainer, IP registration, and the ongoing compliance calendar the center must run to stay audit-clean | $10K–$35K | High |
| Insurance | Directors' and officers' insurance, cyber insurance, professional indemnity, and group health cover for employees and dependents | $8K–$25K | Medium |
| Attrition & replacement buffer | Recruiting pipeline for natural attrition, notice-period overlap cost, and the retention spend (increments, training, skip-level reviews) that keeps a 20–40 seat center from turning over its entire team in two years | $10K–$40K | Medium |
| Contingency | A 10–15% buffer over the total, drawn on when hiring velocity is slower than modeled, fit-out runs over, or the parent adds a compliance requirement mid-year | $50K–$150K | High |
| Exit & wind-down provision | Notice-period obligations, lease exit penalties, and the transition cost if the center is later consolidated or closed. Priced as an option, exercised rarely, and never modeled — until it is needed | $0–$50K | Low |
Pricing your GCC from a template that missed 60% of the real cost?
We build and audit global capability centers for a living. Our GCC setup workshop produces a twelve-line-item cost model, a hiring plan priced from actual offer letters, and a year-one budget that survives the first board review — before you sign the lease.
Get a Realistic GCC Cost Model →Where the 2026 cost models most often go wrong
The first mistake is modeling the team at database salaries. Published averages are six to eighteen months old, and the 2026 market has compressed the spread: strong senior engineers and delivery leads in tier-1 cities now clear the top of what most business cases modeled as the midpoint. A cost model built from a salary database will be wrong by the amount of two or three offer-letter rounds. The model must be built from offers issued in the last 90 days for the specific roles and the specific city.
The second mistake is treating IT integration as a setup event instead of a program. Federating identity, standing up VDI or a managed data path, and passing the parent's security review routinely takes two to four months and needs its own owners on both sides. Until it is done, the center cannot run real work, and the clock on the lease and the payroll does not stop. The integration budget belongs in the setup estimate, not in the operations phase where it will compete with payroll for approval.
The third mistake is forgetting that the first year has two payroll ramps. Headcount is rarely hired all at once. Each intake round carries a recruiting cost and a training cost, and until the team reaches critical mass, output does not scale with cost. A model that prices 30 seats at an average salary but forgets the ramp is really pricing 30 seats arriving on day one. The gap between the two is a common source of the 60–120% overshoot.
Under-Budgeted Captive
Entity, office and an average-salary headcount model. No IT integration, no compliance calendar, no attrition buffer, no contingency. Approved on a number that is 40–60% of the real total, then forced into hiring freezes and retention risk in month six when the missing lines arrive as invoices.
Complete Cost Model
All twelve line items priced from actual offer letters, real lease terms and vendor quotes. Integration funded in the setup phase, contingency funded at 10–15%, attrition priced as a line item. The center launches with a number that the board can audit at month twelve and find true.
Compounding GCC
Year-one setup is amortized across year-three cost per FTE, and the center starts to be measured on delivered value rather than unit cost. The savings the business case promised are real because they were modeled completely. The next wave of hiring is funded by the first wave's results, not by a new approval cycle.
The GCC cost checklist: what a year-one budget that survives the audit covers
"Every GCC business case gets the setup capital right and the total year-one cost wrong. The centers that survive are the ones whose sponsors were willing to see the whole number before the ink dried on the lease."
What to do this week
01Rebuild the cost model as twelve line items, not three
Take whatever spreadsheet approved the current plan and expand it to the twelve lines above: entity, office, IT baseline, hiring, payroll, systems, integration, legal and audit, insurance, attrition, contingency and exit. Even before you change a number, the structure will show you which line items are missing from the approved budget. If the current model has no contingency and no integration line, the gap between approved and actual is already priced in.
02Price the payroll line from three recent offer letters per role
Pick the five roles that dominate the headcount plan. For each, get three offer letters from the last 90 days — from recruiters, peers, or the offer your own hiring team has already made. Rebuild the payroll line from those numbers for your city and your seniority mix. This one exercise removes the largest single source of error from the model, and it takes an afternoon.
03Get a written timeline for parent IT integration before you sign the lease
Ask the parent's infosec and platform teams for the actual sequence: identity federation, VDI or data path, security review, data-residency decision. Each step has a real duration and a real owner. Put the dates in the project plan. If the integration timeline runs two to four months — and it usually does — the setup budget and the go-live date must be built around it, or the center will pay rent and payroll before it can do any real work.
04Get a complete model in front of you before the next approval cycle
The GCCs that survive are not the cheapest on paper; they are the ones whose sponsors saw the whole number and funded it. A complete cost model — twelve lines, priced from offers, with integration and contingency included — is what turns a captive from a cost risk into a deliverable that the board can audit at month twelve. We build this model for every GCC we stand up, and we will build yours in a two-week engagement that ends with a budget the first-year audit will confirm.
Let 10decoders price your GCC before you sign the lease
We have stood up and audited global capability centers across India. Our GCC setup engagement produces the twelve-line cost model, the offer-letter-priced hiring plan, the IT integration timeline, and the year-one budget that survives the first board review — before you commit to an office, a payroll, or a parent IT program.