Viksit Bharat@2047: Through the Lens of AI and Global Capability Centers
Ask most people in major cities in India what “Viksit Bharat” means, and you’ll get a version of the same answer: a developed India by 2047, the hundredth year of independence. It’s a big, almost audacious target — a $30–40 trillion economy, built on inclusive growth, technological self-reliance, and a much louder voice on the world stage.
The government has organised the vision around four groups it wants to lift: youth (Yuva), the poor (Garib), women (Mahila), and farmers (Kisan). Underneath all of it sits Atmanirbhar Bharat, the push for self-reliance, paired with an equally strong appetite for global partnerships and leadership in innovation and governance.
Two things keep coming up whenever this vision gets discussed in policy circles: artificial intelligence and the explosive growth of Global Capability Centers, or GCCs. Together they’re doing a lot of the heavy lifting — creating high-value jobs, seeding indigenous innovation, and pulling India deeper into global supply and value chains. Prime Minister Shri Narendra Modi has said more than once that he wants India among the world’s top three AI powers — and not merely as a consumer of AI built elsewhere, but as a creator of sovereign, inclusive AI built on Indian terms.
What follows is a look at how this vision came to be, where AI and GCCs fit into it (with examples), and what stands in the way between now and 2047.
Where This Ambition Comes From
It is worth reflecting on how far the journey has come from its humble beginnings. India in 1947 was a low-income economy just beginning to find its footing; today it’s the world’s fifth largest. That arc runs through the 1991 liberalization reforms, the Digital India push that took off around 2015, and more recently the production-linked incentive (PLI) schemes that tried to pull manufacturing back onshore.
Getting to Viksit Bharat means sustaining something close to 8% annual GDP growth for two decades — a shift away from an economy driven mostly by domestic consumption toward one driven by manufacturing and innovation. That requires infrastructure most people take for granted in richer countries: better roads and ports, yes, but also the quieter digital plumbing — UPI, Aadhaar — that already underpins daily transactions for hundreds of millions of Indians. It requires skilling at a scale the National Education Policy 2020 is only beginning to attempt, and a genuine push toward net-zero, including a renewed bet on nuclear power.
AI and GCCs matter here because they act as multipliers. They stretch the value of human capital further than it would otherwise go, and they pull in foreign direct investment that infrastructure alone can’t. India’s tech sector, GCCs included, is already a meaningful slice of GDP, and under an aggressive-adoption scenario, AI alone could add close to $1.7 trillion to the economy by 2035.
Technology as the Connective Tissue
Beyond the four social pillars, there are strategic ones too: economic competitiveness, national security, global partnerships, strong legal and regulatory frameworks. Technology threads through all of them — semiconductors, quantum computing, supercomputing, and AI aren’t separate initiatives so much as the backbone that makes self-reliance possible at all.
AI and GCCs matter here because they act as multipliers. They stretch the value of human capital further than it would otherwise go, and they pull in foreign direct investment that infrastructure alone can’t.
The youth cohort — sometimes called Amrit Peedhi — is where a lot of this energy is concentrated. India already has the world’s third-largest startup ecosystem, and GCCs plus AI are turning what used to be described somewhat abstractly as a “demographic dividend” into actual jobs. Women-led enterprises are growing. Farmers are getting access, however unevenly, to precision agriculture tools built on AI. None of this is evenly distributed yet, but the direction is consistent.
AI as the Engine
India’s official framing is clear, concise, and impactful: “Make AI in India, Make AI Work for India.” The IndiaAI Mission, launched in March 2024 with an outlay of ₹10,372 crore, organizes the work around seven areas — compute infrastructure, foundational models, datasets, applications, entrepreneurship, skilling, and safe AI.
On the compute side, capacity has grown fast — from around 10,000 GPUs to more than 38,000, made available to researchers and startups at subsidized rates of roughly ₹65 an hour. The stated goal is 100,000-plus publicly accessible GPUs, with private capacity pushing the national total well past 200,000.
The more interesting story, though, is on the model side. Sarvam AI, a Bengaluru startup, was selected to build a sovereign large language model — one trained on Indian data, tuned to Indian languages and cultural context rather than adapted after the fact from a Western model. BHASHINI, the government’s multilingual AI initiative, supports similar work for public services. NITI Aayog has projected that AI could help push growth toward 8% annually, potentially lifting GDP to around $21 trillion by 2047, well above baseline projections without it.
The sector-by-sector picture is wide: crop monitoring and yield prediction in agriculture, diagnostics and drug discovery in healthcare, personalized learning tools in education, predictive maintenance in manufacturing, fraud detection in finance. NITI Aayog’s roadmap singles out financial services, pharmaceuticals, manufacturing, and automobiles as priority sectors. The employment numbers being floated are large — up to 4 million new “AI-first” jobs by around 2030, with demand for AI talent expected to climb from roughly 800,000–850,000 today to over 1.25 million.
Sarvam AI and the Case for Sovereign Models
Sarvam is a useful example of what “sovereign AI” actually looks like in practice. It’s building large language models trained on Indian datasets, capable across more than 20 languages, designed for voice-first use — which matters enormously in a country where a large share of the population is more comfortable speaking than typing. Access to IndiaAI Mission compute lets Sarvam train these models domestically rather than renting capacity or IP from abroad.
The output feeds into “BharatGen,” aimed at public-service applications: a government chatbot that responds in a local dialect, for instance, isn’t a novelty here — it’s a genuine attempt to narrow the digital divide, and potentially something India could eventually export to other countries in the Global South facing similar language diversity.
Microsoft and the Classroom
Microsoft’s Bengaluru R&D team built AI tools that help Karnataka’s teachers generate personalized lesson plans, now integrated with the government’s DIKSHA education platform. In healthcare, a similar partnership with Apollo Hospitals produced a clinical AI assistant that reportedly saves doctors about 20% of the time they’d otherwise spend on data entry and record-keeping — time that goes back into seeing patients. Small efficiency gains like these, multiplied across a system serving over a billion people, add up.
None of this is without friction. Data quality remains inconsistent. Talent retention is a real worry given how aggressively global firms compete for the same AI engineers. Compute-hungry training runs carry a real energy cost. The responses so far — responsible AI guidelines, large-scale reskilling programs like FutureSkills PRIME (which has already reskilled over 300,000 people), and deeper public-private collaboration — are reasonable starts, but nobody would call the problem solved.
GCCs — No Longer the Back Office
Global Capability Centers are, in essence, offshore units that multinational companies set up and fully own, rather than outsourcing to a third party, covering IT, R&D, analytics, finance, and increasingly, core product development. India now hosts the largest concentration of these centers anywhere in the world: roughly 2,100+ centers spread across 3,600+ individual units as of FY26, employing 2.2 million+ people and generating close to $98 billion+ in value. The ecosystem has grown 32% in size since FY21, with more than 500 new centers opening in recent years.
What’s changed isn’t just the scale, it’s the nature of the work. Nearly half of these centers — 46% — now function as genuine “portfolio” or “transformation” hubs rather than cost-saving back offices, a marked shift from where things stood even a few years ago. AI and machine learning now run through more than 1,200 GCCs, supported by over 250 dedicated AI centers of excellence and more than 250,000 AI professionals — roughly 28% of the entire global GCC AI workforce sits in India. Hiring reflects this: an estimated 510,000 jobs are expected in 2026 alone, and 64% of them will require AI or data skills. Bengaluru remains the anchor, with around 1,080 units, followed by Hyderabad and the National Capital Region, while Tier-2 cities are now the fastest-growing segment of the map.
The roster of companies setting up shop keeps widening too: Forbes Global 2000 firms, private-equity-backed companies, and newer entrants like Anthropic and Marriott. Increasingly, the innovation flow runs in both directions — products and solutions built in Indian GCCs are shipped out globally, not just adapted from headquarters.
SAP Labs and Joule
SAP’s Bengaluru center built Joule, a generative AI copilot that sits across SAP’s enterprise software suite, letting users automate tasks and pull insights through natural-language queries. It was conceived and built in India, then rolled out globally — a fairly clean example of a GCC moving from support function to genuine product owner, which is exactly the kind of IP-building Atmanirbhar Bharat is meant to encourage.
Walmart and Retail Intelligence
Walmart Global Tech India uses machine learning out of its Bengaluru hub for inventory forecasting, real-time product substitutions, and personalized recommendations — work that ultimately optimizes supply chains at global scale and cuts waste. There’s a sustainability angle too, in more efficient logistics. The savings run into the billions, and there’s an obvious path for these techniques to filter into Indian retail through Walmart’s stake in Flipkart, potentially pulling local small businesses into more sophisticated supply networks.
Pharma and Healthcare GCCs
Novo Nordisk’s India operation leverages AI across drug development support, regulatory documentation, and personalized diabetes care. Amgen’s Hyderabad center focuses on precision oncology analytics, backed by an investment north of $200 million. Siemens Healthineers uses AI in diagnostic imaging and radiology. Collectively, this work shortens R&D timelines and lowers costs, which matters directly for India’s stated goal of $350 billion in pharmaceutical exports by 2047.
Manufacturing and Energy
In steel and energy, GCCs are using computer vision for defect detection and predictive maintenance, and for optimizing things like waste-heat recovery. At least one center reported a meaningful drop in downtime and carbon footprint as a result. Applied more broadly, this kind of work supports both the green transition and the manufacturing self-reliance that PLI schemes are trying to build.
Beyond the direct economic output, GCCs are quietly building India’s digital sovereignty, cybersecurity capability, data localization practices, and a deep bench of trained talent that didn’t exist at this scale a decade ago. Many now work closely with local startups and universities, which is arguably where the longer-term payoff lies.
Where AI and GCCs Reinforce Each Other
The connection between the two isn’t incidental. Indigenous AI models reduce reliance on imported technology; GCC-driven R&D feeds into the broader push around semiconductors and AI hardware under Semicon India. Together they generate millions of jobs, give youth a reason to stay and build rather than emigrate, open doors for women in tech, and modernize agriculture and healthcare from the ground up.
GCCs are turning India into more of an innovation exporter than an outsourcing destination — and combined with digital public infrastructure like UPI and the Open Network for Digital Commerce (ONDC), India is increasingly exporting its governance models, not just its labour.
There’s a global dimension too. That said, geopolitical friction and global competition for talent means policy must stay nimble, especially around data protection law and IP frameworks.
Closing Thoughts
Viksit Bharat@2047 isn’t just an economic target, it’s closer to a national reinvention — and AI and GCCs are two of its clearest working examples. Sarvam’s sovereign models, SAP’s and Walmart’s product innovations, and the healthcare breakthroughs coming out of pharma GCCs all point to something concrete rather than aspirational: this is already happening, unevenly but genuinely.
Whether India actually gets there by 2047 depends less on any single technology and more on follow-through — largely around policy continuity, sustained investment in skills, and enough patience to let institutions mature. If it works, the result won’t just be a bigger economy; it’ll be a different kind of development story, one built as much on indigenous capability as on capital. That’s the harder version of the bet India has placed, and the next two decades will show whether it pays off.