The numbers Brief Episode map Takeaways Charts Quotes Clips Glossary Go next Transcript

Episode 113 · Enterprise · 43 min

The centre that became the headquarters

Zinnov's case is that the arbitrage story is finished. India's roughly 1,580 global capability centres now employ 1.7 million people, and the interesting ones have stopped waiting for instructions — a Pune-based centre head is now a global CTO, and a Japanese company's Bangalore office conceived a product that went on to raise close to $100 million. The binding constraint has moved from cost to competence: about 1.5 million STEM graduates a year, and only a fifth of them employable by a global firm.

GCCs & services Hiring & talent Leadership & org

Nilesh

President, Globalization Excellence, Zinnov · with Vishal Krishna

42:57

Said in this episode

▶ 17:17

1.7 million

People employed by GCCs in India

The headcount across India's global capability centres at the time of recording, on a base of roughly 1,580 centres growing about 6% a year.

▶ 34:15

1,580

GCCs operating in India

Nilesh corrects the host's '1,500 plus' to about 1,600, then 1,580 — with 10 new centres set up in one quarter and 19 in the previous one.

▶ 16:19

1.5M → 250k

STEM graduates vs globally employable engineers

About 1.5 million STEM graduates, of whom roughly 20% are ready for global companies; the working figure on air is 250,000, and the ask is to double it to 500,000.

▶ 2:11

~1/3

India cost of a $150,000 US engineer

Same capabilities, same skill set, roughly a third of the cost — the arbitrage that started it, which Nilesh argues should now be spent on capacity rather than banked.

▶ 31:16

10–12%

Attrition in centres Zinnov runs

Against a Bangalore average of about 15%; Zinnov claims two to three points below that, plus time-to-productivity roughly halved.

▶ 9:45

30 → 12,000

One global tech centre's headcount in a decade

Vishal's own observation of centres that landed with 30 people and now run past 12,000, which Nilesh confirms.

The brief

The argument in sixty seconds

Nilesh's claim is that cost arbitrage was only the entry ticket, and that treating it as the point is now the expensive mistake. A US engineer costs upward of $150,000 and the same skills in India cost about a third — but the companies getting real value from that gap spend it on capacity for top-line growth rather than booking it as a saving. What separates a services mindset from a product one, he argues, is a single question: who is your customer? An engineer at a services firm names the client company; an engineer in a product company's Bangalore office should name the same small business the parent serves. India now has roughly 1,580 global capability centres employing 1.7 million people and growing about 6% a year, and the leadership map has inverted — a Pune-based centre head runs a global CTO role, an India site lead has become a company's CHRO, and a Japanese firm's Bangalore office conceived a product that raised close to $100 million. The constraint has moved to competence: India produces about 1.5 million STEM graduates and only a fifth are employable by a global company, against a world short of more than half the AI engineers it needs and US packages touching a million dollars. His prescription is unglamorous — boot camps, college-hire pyramids, small-town centres like Zinnov's 300-person site at Nemili — and his warning is blunt: the India entity only succeeds if the parent in the US actually wants it to.

Worth your time if you are

CTOs sizing up a first engineering centre in India

GCC heads hiring against Google's brand

Services engineers who still wait to be told what to build

State IT officials courting capability centres

Campus hires choosing between a GCC and a startup

Episode map

Where the conversation travels

Every block is a chapter, coloured by what it's about. Click any of it to jump straight to that minute on YouTube.

GCCs & servicesHiring & talentLeadership & orgAI & machine learningIndia macroProduct strategy

01Cold open: globalisation minus the arbitrage 0:00 Nilesh defines globalisation excellence against the cost-cutting reflex — a $150,000 US engineer costs a third of that here, but the point is the extra capacity for top-line growth, and the mindset shift from an engineer whose customer is a contract to one whose customer is the parent's customer. 02What Zinnov actually sells 4:56 A four-to-six-week design and discovery pass over the client's products, roadmaps, skills, culture and time zones, then entity setup, hiring, onboarding and L&D — plus the harder half, change management with the US counterparts who have never run a team this far away. 03The two questions every CTO asks 7:52 Operational worries are gone; what CTOs ask now is whether their brand can hire against Microsoft, Google and Amazon, and how fast productivity arrives — to which the answer is that co-located functions often make the global centre more productive, not less. 04Talent first, market second, and the reverse 10:07 Once Amazon, Google, Microsoft and Intuit have people and customers here, India stops being only a talent pool and becomes a market — and Nilesh increasingly sees companies arrive for the market and discover the talent. 05The VCs who had money and nobody to fund 12:10 Setting up Intuit's Bangalore centre in 2004, Nilesh met funds begging for founders when an Infosys job was the gold standard; 10,000-plus startups, Freshworks, Swiggy and Zomato later, exposure and role models — not any single programme — explain the product turn. 06Half the AI engineers, a fifth of the graduates 15:04 Zinnov's landscape study puts India's AI and data talent at about half the US pool while the world is short more than half the AI engineers it needs — and India's own bottleneck is that of 1.5 million STEM graduates only about 20% are employable by a global company. 071.7 million people and a 6% run rate 16:40 Ten new GCCs one quarter and 19 the previous one, 1.7 million people employed and an employable-engineer pool Nilesh describes as steadily climbing, as Zinnov shifts from consultant to operator that hands built centres back to their owners. 08Global roles run out of India 18:48 Most GCC leaders already carry US or European experience, and the traffic has reversed — a customer's Pune-based centre head is now global CTO, an India site lead became the company's CHRO, and Tesco runs several global leaders from here. 09Nearshore, Japan, and a $100M Bangalore product 20:50 US firms still dominate demand at roughly 65%, but Japanese companies are arriving fast — one launched a product conceived entirely in its Bangalore office that raised close to $100 million — while time-zone needs push work to Mexico, Poland and Vietnam without matching India's depth. 10The pyramid, the boot camp, the son test 23:20 Companies will keep having to train their own graduates, which is why Zinnov pushes clients into strong college-hire programmes — the pyramid holds costs down and brings in people who think differently from the people building the product. 11GCCs and IT services: an and, not an or 25:00 Intuit still works with Persistent and Cognizant almost 20 years on at nearly 10x the volume, because centres need 20-30% flex and niche skills — but with LLMs displacing body-shopped work, services firms have to move from people-driven to value-driven consulting. 12LLMs eat the boilerplate; Nemili answers back 27:24 Boilerplate code is going the way of commodity work while people who can build on top of the models stay in demand — and Zinnov's own 300-person centre in Nemili, a small town near Chennai, shows the work no longer needs one of India's dozen big cities. 13Attrition, brand and the preference ladder 29:10 Nothing fundamental changes in the next five years, Nilesh argues — brand, training and culture still decide, and the evidence is in attrition two to three points under Bangalore's 15% average and an engineer's preference ladder that puts Fortune 500s and unicorns above service providers. 14Media lags; 1,580 centres and counting 32:34 Media and entertainment has been a slow mover despite India's own strength in it, while healthcare has raced ahead — and the base is about 1,580 GCCs growing 6% a year, with the US-versus-rest split expected to even out over time. 15States compete, Britain arrives, India's century 35:26 Lloyds Banking is already 700 people in and Barclays moved to Chennai after Lithuania would not scale, while Tamil Nadu, Karnataka and Hyderabad now actively court GCCs — clearing roadblocks rather than selling land. 16IIT Bombay, QuickBooks, and influence 38:10 From electrical engineering at IIT Bombay in 1985 and a masters in image processing to a career rule — the most important skill is the ability to influence others — illustrated by the iPhone and by QuickBooks taking 80% share in three months by deleting debit and credit.

Takeaways

Ideas to carry out of this hour

Arbitrage now buys capacity, not savings

A US engineer costs anywhere from $150,000, and the same person with the same skills costs about a third of that in India — that is all cost arbitrage ever meant. Nilesh's argument is that the companies still treating the gap as a bottom-line saving are the ones underusing it: the same money buys more engineers, which buys more problems solved and more top-line growth. Most of the businesses he sees now justify their centres on capacity for growth rather than on a lower cost line.

▶ 2:11 ▶ 2:26 ▶ 2:41

One question separates services minds from product minds

Ask an engineer at a services company who their customer is and they name the company that gave them the contract. Ask an engineer at Intuit in Bangalore and the old answer was Intuit US — the right answer, Nilesh says, is the same one an Intuit engineer in the US would give: a small business filing its taxes. He still meets people hired out of services into product organisations who wait to be told what to build, which is why the mindset transfer, not the hiring, is the hard part.

▶ 3:58 ▶ 4:12 ▶ 4:41

The India centre fails if the parent wants it to

Half of Zinnov's work happens on the other side of the ocean: change management with US counterparts who have never worked with a team that far away, in a different culture and a different communication style. The failure mode is treating your own centre like an outsourcing vendor — withholding customer knowledge and domain context, and never granting empowerment. When he set up the Bangalore centre himself, Nilesh says, the harder job was making the US side understand what it takes to make India work.

▶ 6:40 ▶ 7:09 ▶ 7:37

Aim low on productivity and you will get low

The persistent assumption is that people hired remotely, or in India, are less productive, so leaders quietly plan for less. Nilesh's counter is that you should expect equal or better output from a global centre — and that the centre often beats headquarters because product, support, operations, finance and HR are sitting in one building instead of scattered across time zones. That co-location, not effort, is the mechanism.

▶ 8:52 ▶ 9:07 ▶ 9:22

The talent constraint is employability, not headcount

India graduates about 1.5 million people with a STEM degree of some kind, but only around 20% are ready to be employed by a global company — the figure discussed on air is roughly 250,000 usable engineers, a number Nilesh says keeps rising, and the stated ambition is to double it to 500,000. Meanwhile the world is short more than half the AI engineers it needs, US packages have reached a million dollars, and Zinnov's own landscape study puts India's AI and data pool at about half the American one. Nilesh's fix is public-private skilling rather than waiting for colleges to change.

▶ 15:18 ▶ 15:32 ▶ 16:19

Global roles are being run out of India

Nearly every GCC leader Nilesh meets already carries US, UK or European experience, so the old worry about leadership depth has gone quiet. What is new is the reverse flow: a customer's centre head, still based in Pune, now holds the global CTO role; an India site lead who was also CHRO here took the CHRO job for the whole company; Tesco runs several global leaders from India. The centre has started supplying the org chart above it.

▶ 19:50 ▶ 20:05 ▶ 20:36

GCCs and IT services is an and, not an or

Intuit signed with Persistent and Cognizant when its Bangalore centre was built almost 20 years ago and still works with both — at roughly 10x the volume. Companies need 20-30% flex up and down plus niche skills they will not hire permanently, and once a company has its own people here, the vendor management office moves to India too. The pressure is elsewhere: with LLMs replacing large-headcount work, services firms have to become value-driven rather than people-driven consultancies.

▶ 25:20 ▶ 25:52 ▶ 27:08

The next pool is small-town, and LLMs make it viable

Zinnov runs a centre in Nemili, a small town near Chennai, with more than 300 highly motivated people doing data processing and research for global clients — a place Nilesh had never heard of before he visited. His claim is that remote and hybrid norms plus LLMs, which absorb the commodity and boilerplate layer, make distance cheap while raising the premium on people who can build on top of the models. The cricket analogy does the rest: the talent has always come from everywhere.

▶ 27:54 ▶ 28:24 ▶ 28:39

The numbers, drawn

What the episode measures

Every figure below was said on air — timestamps included, caveats kept.

Conversation share

portion of the hour spent on each theme

GCCs & services30%

Hiring & talent20%

Leadership & org12%

AI & machine learning12%

India macro10%

Product strategy6%

Computed from the chapter map of this episode.

What the arbitrage actually costs

US$ per engineer per year

As stated on air: a US engineer costs 'anywhere from $150,000' and the same engineer with the same capabilities in India about a third of that — plotted here as $50,000. Nilesh's point is that the gap should buy extra capacity for top-line growth, not a lower cost line.▶ 2:11

India's STEM employability gap

thousand engineers

Figures as stated in conversation: about 1.5 million STEM graduates, of whom roughly 20% are employable by global companies, with 250,000 used as the working number and 500,000 as the goal. The two are given loosely — 20% of 1.5 million would be 300,000.▶ 16:19

Worth keeping

Lines that stay

Don't think of globalisation as just a means to cut cost — create excellence for the world and for your own businesses.

— Nilesh ▶ 1:41

The entity in India will only succeed if the entity in the US wants it to succeed.

— Nilesh ▶ 7:24

You should expect equal or better productivity in your global organisations. Don't aim for low — then you'll get low.

— Nilesh ▶ 9:07

It's almost the reverse happening now — global roles are being run out of India.

— Nilesh ▶ 20:36

You can't build a product that you would use, because your customers are probably half your age. You have to think about what my son would want.

— Nilesh ▶ 24:34

Clips that travel

Short on time? Start here

CTOs sizing up a first engineering centre in India

The two questions every CTO asks

Brand versus Google in the hiring market, and why co-located functions make the global centre more productive than headquarters expects.

7:52 → 10:07 · 2 min ▶ Watch clip

GCC heads hiring against Google's brand

Half the AI engineers the world needs

The AI talent shortage, million-dollar US packages, and India's real bottleneck: 1.5 million graduates, a fifth of them employable.

14:06 → 16:40 · 3 min ▶ Watch clip

Anyone told the India centre is a support function

Global roles run out of India

A Pune-based centre head made global CTO, a site lead made CHRO, Tesco's global leaders here — the org chart inverting in real time.

18:48 → 20:50 · 2 min ▶ Watch clip

IT services leaders watching GCCs take their work

An and, not an or

Intuit's 20-year vendor relationships at 10x, the 20-30% flex GCCs cannot carry, and why LLMs force value-driven consulting.

25:00 → 27:24 · 2 min ▶ Watch clip

Engineers outside India's dozen big cities

LLMs, the boilerplate and a town called Nemili

Where commodity work goes when models absorb it, and the 300-person small-town centre that already serves global clients.

27:24 → 29:10 · 2 min ▶ Watch clip

Glossary

The jargon, unpacked

GCC (global capability centre)

A multinational's own offshore office — engineering, operations, support or all three — staffed by its own employees rather than a vendor's; India has about 1,580 of them employing 1.7 million people.

Cost arbitrage

Paying less for the same skills because of living-cost differences between countries — a $150,000 US engineer for roughly a third of that in India.

Globalization excellence

Zinnov's framing for its practice: build global teams for innovation and growth rather than for cost, and hold them to headquarters-level standards.

Nearshore

Placing work in a country close to the customer's time zone — Mexico, Costa Rica or Colombia for the US, Poland or Romania for Europe — chosen for overlap hours rather than scale.

Talent pyramid

The ratio of junior to senior engineers in an organisation; strong college-hire programmes keep the base wide, which controls cost and keeps fresh thinking coming in.

Time to maturity

How long a new centre takes to reach full productivity — the metric Zinnov claims to halve, alongside attrition two to three points under the local average.

Vendor management office

The team that runs a company's relationships with its IT services suppliers — increasingly relocated to India once the company has its own centre here.

Connections

If this resonated, go here next

EP 09 · Enterprise

Full transcript

The whole conversation, searchable

Auto-generated captions, lightly cleaned. Click a timestamp to open that moment on YouTube.