India has 2.36 million engineers in GCCs. Founders still can't hire. Here's why.

The number everyone cites, and the number nobody does

India now hosts 2,117 Global Capability Centres (GCCs) employing 2.36 million tech professionals. The Zinnov-NASSCOM figures are everywhere: 32% growth since FY2021, $98.4 billion in annual revenue, 40% of the global GCC workforce spread across Bengaluru, Hyderabad, Pune, and Chennai, with another 425,000–450,000 jobs expected to be added through FY2026.

Every conversation about Indian tech hiring invokes these numbers. Usually to make one of two points: that GCCs are good for the Indian economy (true), or that they have made it harder for startups to hire (partially true, though not for the reasons usually given).

The number that doesn't make the conference slide: at senior engineering levels, GCC attrition runs at 22 to 30 percent. For AI and ML roles, it sits at the top of that range.

2.36 million engineers in GCCs, nearly a quarter of the senior cohort turning over every year, and founders report that hiring has never felt harder. The paradox resolves once you disaggregate who is leaving GCCs and why.

Why 70% of GCC leavers aren't motivated by your salary budget

A 2026 Zinnov analysis of senior engineers leaving Indian GCCs found that 42% cited limited career progression as their primary reason, 28% cited lack of ownership over product outcomes, and only 19% cited salary gaps. The remaining 11% cited geography, work model, or personal circumstances.

Career progression and ownership account for 70% of GCC attrition. The salary gap receives most of the attention in founder hiring conversations and drives less than a fifth of actual exits.

Most startup hiring playbooks focus on closing the compensation gap. That is optimising for the 19%.

The engineers leaving GCCs in volume aren't doing it because a startup outbid their employer. They're leaving because a senior engineering role at a GCC, regardless of the parent company's prestige, typically reduces to: deliver to spec, operate within a defined boundary, report upward into a global org that makes architectural decisions somewhere else. Bengaluru builds. Singapore reviews. San Francisco decides.

For engineers who are comfortable with that structure, the GCC model is a reasonable arrangement: stable employment, solid pay, clear career ladder to a certain rung. For engineers who have built enough skill that they want to own something (the architecture, the product roadmap, the customer relationship), that model hits its ceiling around year five.

The experience-band problem

GCCs and startups don't compete uniformly across all experience levels. The market has stratified by band, and most founders are recruiting in the wrong one.

Engineers in the 2–6 year band are the hardest to recruit from GCCs. They're building their financial cushion, haven't hit the career ceiling yet, and don't have the risk appetite for a 30-person startup where equity is speculative and the base salary may be modestly lower. The math doesn't work at that career stage.

The 7–12 year band is a different story. This is where GCC attrition concentrates. Engineers here have seen what the senior IC path looks like inside a captive centre: there's a ceiling, it's visible, and clearing it usually means moving into management or lateraling to a new GCC. Many have worked at three or four organisations and can clearly distinguish between 'you'll work on hard problems' (true in most GCCs) and 'you'll own the outcome' (usually not true in a captive model).

Founders who recruit as if GCCs are a uniform competitor are solving the wrong problem. The right question is simpler: are we sourcing in the experience band where our offer is competitive, and are we leading with what that band actually wants?

What founders who are winning are doing differently

Three patterns appear consistently among founders who report strong engineering hiring outcomes in the current market.

Lead with the problem, not the company. The opening frame of a GCC engineer's day is often 'implement this spec.' At an early-stage startup it's 'figure out what to build.' Founders who open recruiting conversations with the specific technical or product problem they're working on, not the funding history or perks, report higher response rates from senior GCC candidates. The problem is the pitch.

Make ownership specific. 'Significant ownership' and 'a seat at the table' mean nothing to someone who has heard them fifteen times in the past year. Founders who are winning make the claim concrete: 'You'll own the API layer, including deciding what we don't build. You'll be in every customer conversation about it until we hire someone whose full-time job that is.' Vague ownership claims are indistinguishable from marketing. Specific ones are checkable, and that is the point.

Size equity to compute. A candidate who can't construct a plausible upside scenario from an equity grant treats it as noise, not motivation. Pre-Series B companies winning the most GCC-background offers are typically in the 0.2–0.8% range for founding-adjacent engineering hires, not co-founder grants, but large enough that the math is worth doing under conservative exit assumptions.

The GenAI wildcard

There is a structural shift happening inside GCCs that hasn't fully propagated into the hiring market yet.

GCCs were built on a specific model: high-quality engineering talent performing well-defined tasks at a cost advantage relative to the parent company's home market. A significant portion of that work (requirements translation, test generation, code review, boilerplate service delivery) is being automated by AI tooling faster inside large GCCs than inside smaller startups. Enterprise procurement budgets mean GCCs can access the tooling. But they are also the entities whose existing workflows are most directly in the path of that automation.

The result is a cohort of engineers in the 4–8 year experience band who will be released from their current functions faster than new GCC roles can absorb them. GCC revenue CAGR is 9.9%. If a meaningful portion of that revenue is associated with tasks that AI tooling now handles at a fraction of the cost, employment demand for that work will contract.

This is directional, not certain. But the direction is legible from the data, and the timeline is probably shorter than most people in the current hiring conversation are assuming.

Founders who build recruiting relationships with senior GCC engineers now, before this acceleration becomes visible in the market, are positioning to hire from the next wave of leavers, whether those engineers depart by choice or by circumstance.

What to change this month

If GCC candidates are unresponsive to your outreach, convert poorly on interviews, or accept offers and then renege, the problem is almost always one of three things.

You're recruiting in the 2–5 year band and expecting startup-risk appetite that isn't there yet. Move your sourcing toward engineers with 7+ years of experience, where the career ceiling is visible and the motivation for ownership is genuine.

Your ownership pitch is vague. Specific ownership claims are checkable; vague ones are discountable. Get specific about what the engineer will own, what decisions they will make, what will change in the product because of their presence.

Your equity is too small to compute. If the math doesn't produce a number worth reasoning about, it produces nothing. Size it so a candidate can run the upside scenario, even the conservative one, and get to a figure that motivates rather than discourages.

India's GCC sector hasn't thinned out the startup engineering pool. It reorganised it. The engineers who leave GCCs are the ones who want something GCCs structurally cannot offer. Founders who recruit against that motivation are fishing where the fish actually are.