The number that made the CFO wake up

If you're an enterprise CFO looking at your Bengaluru cost centre budget in 2026, one number probably ruined a recent morning: total spend on the India GCC has grown by roughly 40-60% in the last three years, and it's harder to hire and retain than it's ever been.

The macro story is spectacular. India's Global Capability Centre ecosystem crossed 2,100 centres in 2026, employs approximately 1.9 million people, and generates roughly $100 billion in annual revenue for the global parents that run them. Bengaluru alone accounts for about 40% of that. Karnataka's state government has publicly targeted 500 more GCCs by 2029.

The reality behind the number is more complicated. Wage inflation for senior engineers is running at double-digit rates. Attrition remains stubbornly high. Hyderabad quietly captured 41% of new GCC decisions in 2026, up from about 25% two years ago. And the specific talent needed for AI-native GCCs is genuinely scarce.

If you run a GCC, are considering one, or work in one, here's what actually matters for 2026 and 2027 planning.

The current shape of the ecosystem

The numbers, curated from analyst reports and government data through mid-2026:

• Total GCCs in India: ~2,100, with roughly two-thirds Fortune 500-owned.
• Employees: ~1.9M direct, plus roughly 3M indirect.
• Annual revenue impact for parents: ~$100B, growing 15-20% YoY.
• Bengaluru share: ~40% of centres, ~45% of headcount.
• Hyderabad share: rising fast, ~20% of centres now but 41% of new-GCC decisions in 2026.
• NCR (Delhi/Noida/Gurugram): ~15%.
• Pune: ~10%.
• Chennai, Mumbai, Ahmedabad, tier-2 cities: growing shares of specialized work.

Nearly every major enterprise now runs a GCC. What used to be a cost centre is now, for many companies, the primary engineering hub — where core product decisions, architecture, and delivery happen. The organizational status of Indian GCCs has genuinely shifted.

The skills shortage that changes the math

The "cheap engineers" era is over. The specific tension:

• Junior engineer supply is enormous. Roughly 1.5M engineering graduates per year in India, and Bengaluru remains the destination.
• Senior engineer supply is scarce. Ten-plus years of experience combined with AI/agent/systems depth is a small pool. Wages have doubled or tripled for this segment in three years.
• The "AI-native engineer" archetype — someone who thinks in agents, LLMs, and modern MLOps — is genuinely in short supply. Fortune 500 GCCs are competing with Anthropic, OpenAI, and well-funded Indian startups for the same handful of people.

The result: the traditional GCC economic model (leverage senior talent's cost differential to run a large team of juniors) is under pressure. Senior talent isn't cheap. And AI is compressing the value of junior talent — if a senior engineer plus Cursor plus Claude Code writes as much software as a team of five juniors used to, why staff the five?

The Hyderabad question

Hyderabad's rise is the single most-discussed shift in the industry. The drivers:

• Lower salary levels compared to Bengaluru — roughly 15-20% for equivalent talent.
• Better infrastructure in newer areas (Gachibowli, Hitec City).
• Aggressive state incentives from the Telangana government.
• Reduced congestion compared to Bengaluru's genuinely difficult commute.
• Perceived better work-life quality — a real recruiting advantage.

New GCC decisions in 2026 broke Hyderabad's way for the first time. Companies like Microsoft, Amazon, and Google have significantly expanded there; several late-stage Indian startups moved engineering functions.

Does this mean Bengaluru is declining? No. But it does mean Bengaluru has real competition for the first time in twenty years, and that changes the calculus for new investments.

What GCCs are actually working on in 2026

The mix has shifted meaningfully.

Legacy focus (declining): IT-services replication, back-office operations, application maintenance. Still a real part of many GCCs but proportionally shrinking.

Current strength: Product engineering, platform engineering, cloud infrastructure, data engineering, security operations. These are the modern GCC's bread and butter.

Growth areas: AI engineering (agent development, model deployment, evals, safety), applied research, developer experience, internal tools that increase parent-company productivity. This is where the bidding wars are.

Sovereign work returning to source: Certain functions that once made sense to offshore — highest-sensitivity data engineering, some IP-critical R&D — are being repatriated as data-residency rules and geopolitics tighten. Modest trend, worth watching.

What's actually working for GCC leaders

The best-performing GCCs we see (and yes, we work with several) share a few habits.

1. Positioning as centres of excellence, not delivery centres

The GCC that produces AI capability the parent needs — and gets sourced by other business units globally — has vastly more strategic weight than the one running managed services. Positioning matters. Some of this is real, some is narrative, both count.

2. Aggressive investment in senior talent

Paying up for senior engineers is now table stakes. The GCCs winning are the ones budgeting for it. Trying to hold senior salaries flat at 2022 levels produces attrition to competitors that will eat you.

3. Genuine ownership of end-to-end products

The best GCCs run entire products or product areas as owner, not just as executor. This means P&L visibility, product roadmap authority, and the ability to say no to bad ideas from HQ. Where this exists, retention is dramatically better.

4. AI-native operating models

Some GCCs are still running like it's 2019. The best are running like their parent company's most modern engineering org — Cursor, Claude Code, Copilot, internal Slack agents, extensive AI-assisted operations. This is a productivity multiplier and a talent attractor.

5. Real career paths that don't require moving abroad

Historically, the best Indian talent hit a ceiling at the GCC and left for the parent HQ or for local startups. The GCCs solving this have real senior technical tracks locally, real principal engineer roles, and real influence over global technical decisions.

Where founders and agencies fit

If you're not running a GCC but adjacent to the ecosystem — a founder, an agency, a services firm — the boom creates real opportunities.

Consulting to help stand up new GCCs. Every new Fortune 500 decision to open in India creates 6-12 months of vendor-heavy activity. Site selection, hiring, initial delivery capability.

Specialized services that GCCs don't want to build in-house. AI/agent development, security engineering, dev tooling. GCCs would rather buy these than staff for them.

Talent partnerships. Agencies that can lend senior talent on flexible terms serve real GCC needs during ramp-up and specialization phases.

Bridge to global buyers. Indian agencies that can serve GCC parents directly — not through the GCC — capture higher-margin work.

At Xenolve we do a mix of these. Our engineering team works alongside GCC teams for specific engagements, and we build the AI capability many GCCs would rather not stand up internally. If your GCC is evaluating a specific AI or agent build, or you're weighing whether to buy or build a capability, reach out — we've been on both sides of that decision.

What this means for individual engineers

If you work in a Bengaluru or Hyderabad GCC in 2026, the market is favorable but the ground is shifting. Practical observations:

• The premium for AI-native skills is real and probably growing. Investing in agent development, MLOps, and applied AI is the highest-ROI skill investment right now.
• The premium for pure code output is compressing. If your day job is writing standard CRUD, AI is coming for your leverage. Move up-stack.
• The best GCCs are worth working for; the median ones are not what they used to be. Choose employer carefully.
• Startups pay less base but the equity story remains real. If you have runway to bear the risk, the upside is genuine.
• Location choice matters. Hyderabad is a real alternative for certain roles.

The 2027 outlook

Prediction range for what happens next:

High confidence:

• Continued 15-20% revenue growth for the GCC ecosystem in aggregate.
• Continued wage inflation for senior AI talent — probably 15-25% annually.
• Hyderabad's share of new decisions holds around 40%; Bengaluru remains the largest overall.
• More sovereign-sensitive work moving to India as data-residency rules tighten globally.

Medium confidence:

• Two or three of the largest Bengaluru GCCs restructure or downsize as AI productivity gains hit their traditional workforce mix.
• Government policy on GCCs sharpens — likely more incentives for tier-2 cities, possibly stricter data residency for certain sectors.
• The senior-engineer premium starts to compress by late 2027 as more talent moves up and as AI tools narrow the gap.

Lower confidence:

• Whether the ecosystem crosses $150B in revenue by 2028 or stalls closer to $120B.
• Whether Hyderabad overtakes Bengaluru in absolute terms this decade (probably not, but closer than most expect).
• Whether a black-swan geopolitical event (US visa changes, India-China tensions, Middle East disruptions) rewrites the map.

Frequently asked questions

Should our company open a new GCC in Bengaluru or Hyderabad? If you need immediate scale and top-tier senior talent, Bengaluru. If you want lower cost with strong quality, Hyderabad. If your work is genuinely specialized (biotech, semiconductor, aerospace), consider tier-2 alternatives (Coimbatore, Kochi, Ahmedabad, Vizag).

Is the talent quality still competitive globally? For senior talent, yes — and often superior for specific stacks. For junior talent, quantity is high, quality is variable, careful selection matters.

Are Indian salaries really approaching Western levels? For the top 5-10% of senior AI engineers in the largest cities, effective compensation (base + equity + retention grants) can approach mid-tier US levels. Median compensation is nowhere near.

What about GCC alternatives — Philippines, Poland, Egypt? Real alternatives for specific roles, but the depth of the Indian ecosystem — talent pool, tooling maturity, English proficiency, time-zone flexibility — is hard to match. Diversification makes sense; wholesale relocation rarely does.

How do agencies compete with GCCs for senior talent? The answer is often not to compete directly. Offer meaningfully different work — client variety, technology exposure, faster career growth than a large GCC — and pay competitively but not at parent-company parity. Some senior engineers prefer agency variety over GCC stability.

The strategic read

Bengaluru's GCC ecosystem is not slowing down. It's maturing, differentiating, and confronting its first serious internal competition from Hyderabad. The economics have shifted permanently — this is a genuinely expensive market for senior AI talent, and that's not going back.

For enterprises with GCCs, the question is whether your India centre is positioned as strategic infrastructure or as a cost line item. Those are very different futures, and 2027 is when the divergence starts to be irreversible.

For agencies serving the ecosystem — us included — the opportunity is to be the specialized capability that GCCs don't want to build. AI, security, developer tooling, applied research. If you're a founder thinking about the space, or a GCC leader planning a 2027 investment, let's talk.

The next chapter of the Indian tech story is being written in 2026. It's a great time to be part of it, and a strategically important time to understand what's actually happening.