India crossed 2,117 Global Capability Centres in FY26. Of those, 583 belong to mid-market companies not Fortune 500 giants, but enterprises in the $100M–$1B revenue band. That number is the single most useful data point in the entire EOR vs GCC india debate, because it quietly demolishes the assumption most founders operate on: that owning your India team is something you graduate into at 500 heads.
India’s GCC ecosystem now spans 2,117 centres, 2.36 million professionals and roughly $98.4 billion in revenue as of FY26 with 583 of those centres run by mid-market enterprises, per the Zinnov–nasscom India GCC Landscape Report 2026.
Every employer of record provider sells the same true story: hire in India in days, no entity, no compliance exposure, cancel anytime. It’s an excellent product. It is also a rental agreement, and rental agreements have a point past which they are the most expensive way to hold an asset.
The vendors who sell you the rental will never tell you where that point is. They have no commercial reason to. So the useful question is not “which EOR is best” , that market is commoditised and the answer changes quarterly. The useful question is: at what headcount, at what seniority mix, and at what strategic intent does the arithmetic flip?
This guide models that flip with real numbers. It covers what an India entity actually costs to stand up and run, where the crossover lands for most teams, the five hidden costs that don’t appear on an EOR invoice, and the transition sequence that avoids blowing up your team on the way across.
TL;DR
This guide is for founders, CTOs and COOs who already have people in India through an employer of record and are wondering whether it's time to own the entity. It models the EOR vs GCC india decision in cost terms rather than marketing terms.
The headline number: for most companies, the crossover sits between 15 and 30 India heads. Below that, EOR fees are cheaper than the fixed cost of running a compliant Indian subsidiary. Above it, you are paying a rental premium of roughly $4,800–$7,200 per person per year for infrastructure you could own outright.
By the end, you'll be able to calculate your own crossover point, pressure-test whether a global capability center is genuinely the right next step, and sequence the migration without losing people. If you're deciding when to switch from EOR to entity, this is the arithmetic that decision rests on.
What the EOR vs GCC India Decision Actually Is
EOR vs GCC India is the decision between employing your Indian team through a third-party legal employer for a per-head monthly fee, versus incorporating your own Indian subsidiary and employing them directly. EOR is a variable-cost rental model with near-zero setup time. A GCC is a fixed-cost ownership model with higher upfront investment and materially better unit economics at scale.
That definition matters because the two options are usually compared on the wrong axis. Buyers compare setup effort where EOR wins every time. The axis that decides the outcome is total cost per head per year, plus the things you cannot do inside someone else’s entity.
The Real Problem: You Are Paying a Fee That Never Stops Scaling
An EOR service fee is charged per employee, per month, forever. It does not amortise. It does not decline meaningfully with tenure. Your tenth engineer costs the same to rent as your first, and your fiftieth costs the same as your tenth.
Public pricing from early-to-mid 2026 puts global EOR platforms at roughly $400–$700 per employee per month, with India-specialist providers closer to $99–$399. Deal lists from $599, Oyster from $699, Multiplier around $400, RemoFirst from $199. Negotiated enterprise rates for a 15–30 person India team typically land in the $350–$550 band.
Take the midpoint. At $500 per head per month, twenty engineers cost $120,000 a year in service fees alone before salary, before statutory contributions, before laptops. That is pure intermediation cost. Nothing in it touches compensation.
Now the part most dedicated teams miss: the fee is only the visible layer. Sitting underneath it are security deposits (often one to two months of full employment cost, held for the life of the contract), FX spreads of 1–3% applied to every payroll cycle, off-cycle payroll charges, per-termination fees, and equipment procurement markups. In most engagements we’ve reviewed, these push the effective cost 15–25% above the quoted headline fee.
The structural issue is worse than the cost. You are building a team you do not legally employ, cannot issue equity to directly, and cannot hold IP through cleanly without leaning on an assignment chain that runs through a vendor’s contract. For a services business that’s tolerable. For a product company, it becomes a diligence problem the moment someone looks at your cap table.
Modelling the Crossover: When to Switch from EOR to Entity
This is the section the EOR vendors won’t write. The crossover is a simple ratio: annual fixed cost of running your own entity ÷ annual EOR fee per head. Everything else is detailed.
What an Indian entity actually costs to stand up
Setting up a wholly owned subsidiary for a foreign parent involves incorporation, FDI reporting under FEMA, PAN/TAN, GST registration, EPF and ESI registration, professional tax registration in each state you employ in, and a bank account that will take longer than everything else combined.
In engagements we’ve run, realistic one-time cost for a foreign parent lands in the ₹8–20 lakh range including legal, CS and tax advisory and realistic time to a compliant first payroll is 10–16 weeks, not the 2–4 weeks that incorporation timelines alone suggest. Anyone quoting you six weeks end-to-end has not accounted for banking.
What it costs to run, every year, regardless of headcount
This is the fixed leg of the equation, and it’s where most models go wrong by underestimating 3–4x. The recurring stack:
- Statutory audit and ROC filings non-negotiable, annual.
- Transfer pricing study and Form 3CEB mandatory for a captive serving its parent, and the single most commonly forgotten line item. Cost-plus markup on a GCC typically sits in the 12–18% band and must be documented and defended.
- Payroll and compliance vendor monthly processing, TDS, EPF/ESI remittance, professional tax.
- Company secretary retains board resolutions, statutory registers, annual returns.
- Registered office and India finance/HR coverage even fractional, this is real money.
Fully loaded, that stack runs ₹55 lakh to ₹1 crore per year for a lean centre. Call it $60,000–$110,000 at prevailing rates.
The crossover arithmetic
Divide fixed by variable:
- Lean entity (~$65,000 fixed) against a premium EOR ($7,200/head/year) → crossover at ~9 heads
- Mid-case entity (~$85,000 fixed) against a mid-market EOR ($6,000/head/year) → crossover at ~14 heads
- Fully staffed entity (~$110,000 fixed) against a discounted EOR ($4,800/head/year) → crossover at ~23 heads
Add a transition buffer for parallel running and one-time setup amortised over 24 months, and the practical EOR headcount threshold for most companies sits at 15–30 people in India. Below 15, EOR almost always wins. Above 30, staying on EOR is a deliberate choice to pay a premium for optionality which is sometimes correct, but should be a decision rather than a default.
The second crossover nobody models
There is a threshold that has nothing to do with headcount, and it usually arrives first.
The moment you need to hire an India site leader, a principal engineer, or anyone senior enough to expect equity, EOR becomes an active constraint. Candidates at that level ask who employs them, and “a third-party provider in Bengaluru, on a contract my employer can terminate with 30 days’ notice” is a losing answer against a direct offer from a company with an Indian entity and an ESOP pool.
We see this pattern constantly in IT staffing services engagements that later convert to GCC builds: a team of 12 stalls not on cost but on its inability to attract a leader. If your next three hires are senior, run the EOR or own entity india decision immediately, regardless of what the headcount math says.
Compliance changed underneath you in November 2025
Anyone modelling the EOR vs GCC india question on pre-2026 assumptions is working from a stale compliance picture. India’s four Labour Codes Wages, Industrial Relations, Social Security, and Occupational Safety came into force, consolidating 29 legacy labour laws into a single framework, with central and state rules still being phased in.
This cuts both ways. It temporarily strengthens the EOR case, because a provider absorbs the transition ambiguity. It also means any cost model built on the old wage-definition rules which affect EPF base calculations and therefore statutory load needs rebuilding before you commit to either path.
What This Looks Like in Practice
A Series B fintech running 14 engineers on EOR. The trigger wasn’t cost; it was a VP Engineering search that had stalled for four months because three final-stage candidates declined over employment structure and equity. The team ran entity setup and senior search in parallel; the leadership hire closed against a direct India offer, and the remaining engineers migrated across at renewal. The EOR fees they’d have paid over the following 18 months roughly covered the entire setup cost.
A US healthtech scale from 8 to 35 in India. Here the mistake was sequencing. They incorporated first and hired second, then burned eleven weeks of fixed entity cost with almost nobody on payroll. Run the other way, lock the hiring pipeline so it offers land in the same month the entity goes live and that dead period disappears. Our own recruitment process outsourcing benchmarks put a job description to interview-ready shortlist at 7–10 working days, which means the pipeline can and should start during incorporation, not after.
Decision Framework: Rent, Own, or Run Both
Most teams treat this as binary. It isn’t. The hybrid entity for core and senior roles, EOR for short-horizon or single-country edge cases, is the most common steady state above 30 heads.
| Signal | Stay on EOR | Move to entity / GCC |
| India headcount | Under 15 | 20+, or 15+ with senior hires pending |
| Hiring horizon | Under 18 months, or pilot | Multi-year, core product ownership |
| Seniority mix | IC-heavy, mid-level | Leadership, principal, ESOP-eligible |
| IP ownership | Peripheral work | Core product and roadmap |
| Cost tolerance | Predictable opex preferred | Willing to trade fixed cost for unit economics |
A useful gut check: if you couldn’t shut the India team down inside a quarter without material damage to your roadmap, you are already operating a capability centre. You’re just renting it.
What Most Teams Get Wrong
They model the india team ownership model as a cost decision and discover eighteen months later that it was a talent decision.
The cost gap between EOR and entity at 20 heads is real but recoverable six figures a year, meaningful, not existential. The gap that compounds is who you can hire. Engineers evaluating a role treat the employing entity as a signal of permanence, and an EOR arrangement reads as a pilot no matter how committed you actually are.
The second error is the reverse: incorporating far too early because it feels like the serious thing to do. We’ve seen teams carry ₹60 lakh of annual fixed cost against four engineers, purely on the strength of a growth plan that hadn’t happened yet. If you’re under ten heads with no senior hires in the next two quarters, the EOR vs GCC india answer is EOR, and paying for ownership you can’t yet use is just an expensive way to look committed.
The third one that actually costs money during transition is assuming your EOR contract lets you take your people with you. Most don’t, cleanly. Standard agreements carry conversion or buyout clauses, typically pegged to a multiple of monthly fees or a percentage of annual salary, and some carry non-solicit language that survives termination.
Read your buyout clause before you model anything, because it can move your crossover point by several heads on its own. Negotiating that clause down at signing costs nothing; negotiating it at exit costs a lot.
The Transition Sequence
A clean EOR to GCC transition runs in this order. Deviating from it is where teams lose people.
- Audit the EOR contract notice period, conversion fees, non-solicit, deposit return terms.
- Incorporate and register in parallel with hiring do not sequence these, or you’ll pay fixed costs against an empty payroll.
- Lock the transfer pricing model before the first invoice; the markup you pick on day one is the one you’ll defend in assessment.
- Migrate seniors first, ICs second leadership on the new entity makes the change read as investment rather than cost-cutting.
- Running both structures for one to two payroll cycles overlaps is cheaper than a missed salary date.
The one that breaks teams is step four. Migrating the junior half first, while leadership stays on the old arrangement, signals exactly the wrong thing.
Before You Sign the Next Renewal
If you’re at the point where the EOR vs GCC india question has come up in a board meeting, the cheapest thing you can do is model it properly before your next renewal date rather than after it. Renewal is the only moment you have real leverage on conversion terms.
Supersourcing has spent over a decade building India teams on both sides of that line staffing and GCC setup services across 527+ delivered projects, with a 98% candidate joining rate and under 1% drop-off on contract roles. That includes the unglamorous part: reading the buyout clause, running the transfer pricing model, and sequencing hiring so the entity isn’t sitting empty. If your next three India hires include someone senior, that’s usually the signal to start.
Bring us your current EOR quote and your 18-month headcount plan and we’ll model the crossover against your actual numbers with no obligation to switch, and no pitch if the answer is that you should stay on EOR for another year.
Talk to the GCC team → · mayank@engineerbabu.com
FAQ
At what headcount does an EOR stop making sense in India?
For most companies, between 15 and 30 heads. Below 15, EOR fees are almost always lower than the fixed annual cost of a compliant entity. Between 15 and 30 it depends on your negotiated per-head rate and how lean your entity stack is. Above 30, you’re paying a rental premium defensible only if you genuinely need the optionality.
What does it actually cost to set up a subsidiary in India for a foreign company?
Budget ₹8–20 lakh one-time for incorporation, FEMA/FDI filings, registrations and advisory, and ₹55 lakh–₹1 crore annually for the fixed run-rate covering audit, transfer pricing, payroll, CS retainer and India finance or HR coverage. The timeline to first compliant payroll is realistically 10–16 weeks, with banking as the usual bottleneck.
Can EOR employees be transferred to your own entity?
Usually yes, but rarely for free. Most contracts include conversion or buyout provisions, and some include non-solicit clauses that survive termination. Pull your agreement before modelling anything the buyout cost can shift the EOR vs GCC india crossover by several heads. Negotiate the clause at signing, not at exit.
Is a GCC only for large enterprises?
No, and the data is unambiguous. India hosts 583 mid-market GCCs run by companies in the $100M–$1B revenue band, and 1 in 3 Fortune 500 companies operates a GCC in India alongside them. Modern centres routinely start at 20–40 people.
What are the hidden costs of an EOR in India?
Security deposits held for the contract’s life, FX spreads of 1–3% per payroll cycle, off-cycle payroll charges, per-termination fees, equipment markups, and conversion fees at exit. Together these typically add 15–25% to the quoted headline rate. Ask for a fully itemised quote including exit costs before comparing providers.
How do we pressure-test our own crossover before committing?
Build it from your own numbers: fixed entity run-rate divided by your negotiated per-head annual EOR fee, then add your buyout cost amortised over 24 months. If the answer is within five heads of your current team size, get a second opinion before you sign anything. That’s the band where the decision is genuinely close and the downside of getting it wrong is highest.
