The Industry That Refuses to Die, Agentic AI, the End of Labour Arbitrage, and India’s Next Great Reinvention

The Industry That Refuses to Die, Agentic AI, the End of Labour Arbitrage, and India’s Next Great Reinvention

The technology-services industry has a peculiar distinction: every few years, we decide it is finished. We have been told that automation would hollow it out, that the cloud would commoditise it, and that platforms would replace it. Each time, the obituary was written early, and each time, the industry’s capacity for reinvention proved to be its ultimate product. The accompanying analysis by B.V.R. Subrahmanyam and Debjani Ghosh, writing in their capacities at NITI Aayog, captures this recurring pattern and argues that the current moment is no different—except that the stakes are higher, the change is faster, and the opportunity for those who adapt is greater than ever.

The technology that now threatens to disrupt the industry is agentic AI—autonomous systems capable of executing multi-step workflows without continuous human supervision. This is not merely an incremental improvement on previous generations of automation. It is a fundamental shift from a model of labour-intensive ‘effort’ to one of high-impact ‘outcomes’. It compresses the core of the industry—software development, testing, and operations—and forces a reconsideration of everything from business models to workforce composition.

Markets, when they wobbled after recent autonomous agent announcements, fell into a familiar reflex: extrapolating new technology to extinction-level outcomes. They are excellent at sensing disruption but far less good at modelling the transition. The authors argue that this is not cyclical turbulence; it is a structural rewrite. The old playbook rewarded scale; the new one rewards judgement, orchestration, and domain depth—things machines cannot do alone.

The shift is from labour arbitrage to intelligence arbitrage. For decades, the Indian IT industry’s competitive advantage rested on the ability to deploy large teams of skilled engineers at costs significantly below those in developed countries. That model, built on linear headcount growth, is being outpaced by AI-native delivery. The winners in the new era will not be firms that blindly chase automation but those that rebuild themselves around it, creating a ‘human + agent + platform’ model that combines the best of both worlds.

The authors are not alarmists; they are strategists. They lay out a vision for how the industry can navigate this transition, identifying six strategic frontier plays that, if executed effectively, could put the industry on track to achieve $750-850 billion in annual revenue by 2035, sustaining a 7-8 per cent share of GDP and expanding its global market share beyond 25 per cent. This is not a defensive posture; it is an aggressive, forward-looking agenda for reinvention.

The Transition: From Labour Arbitrage to Intelligence Arbitrage

To understand the magnitude of the shift, one must appreciate what labour arbitrage meant for the Indian IT industry. For three decades, Indian firms built a global business on the simple proposition that they could deliver the same quality of work as Western firms at a fraction of the cost. They hired thousands of engineers, trained them in the latest technologies, and deployed them in teams that grew linearly with project scope. Scale was the measure of success; headcount was the driver of revenue.

Agentic AI fundamentally alters this equation. When software can write code, test it, and deploy it autonomously, the need for large teams of engineers performing routine tasks diminishes. The value shifts from the number of people deployed to the quality of the outcomes delivered. Firms that continue to bill by the hour will find their revenues compressing as hours shrink. Firms that shift to pricing by outcomes will capture the value of AI-driven productivity gains.

This is not a prediction of mass unemployment. It is a prediction of mass redeployment. Engineers will need to move from routine coding to higher-value activities: designing systems, orchestrating AI agents, ensuring quality and security, and solving complex domain-specific problems. The workforce will be smaller, but it will be more skilled, more productive, and better compensated.

The Six Strategic Frontier Plays

The authors outline six strategic frontiers that Indian IT firms must pursue to navigate this transition successfully. Each represents a significant opportunity, but each also requires substantial investment and a willingness to move beyond familiar business models.

1. Agentic AI Scaling. The concept of ‘services as software’ (SaS) captures the idea that AI agents can deliver outcomes that previously required human effort. The opportunity is to scale this model to serve both mid-market businesses and to modernise legacy systems. This is not about replacing humans; it is about augmenting them, creating non-linear efficiency gains that benefit both providers and clients.

2. Human-tech Hybrid. The workforce of the future will be a partnership between humans and AI agents. Firms must design operating models that optimise this partnership, deploying agents for routine tasks and humans for judgment, creativity, and client relationships. This requires new approaches to training, career development, and performance measurement.

3. Software Reboot. India’s share of the global Software-as-a-Service (SaaS) market is currently around 1 per cent. The opportunity is to dramatically expand this share by rearchitecting existing software value pools—customer relationship management (CRM), enterprise resource planning (ERP), and others—using AI-native approaches. This is not about competing with established players on their own terms; it is about leapfrogging them with superior technology.

4. Infrastructure Hub. AI requires massive computing power. India’s current data centre capacity is 1.4 gigawatts. The authors call for expanding this to 12 gigawatts, establishing India as a global hub for AI-ready data centres. This requires investment in physical infrastructure, power supply, and connectivity, as well as the development of sovereign cloud stacks that ensure data security and regulatory compliance.

5. Innovation Hub. The global research and development (R&D) operations market is estimated at $1.1 trillion. India can capture a significant share of this by establishing specialised Centres of Excellence (CoEs) in sectors such as medical technology, automotive, and industrial manufacturing. These CoEs would not merely execute tasks defined elsewhere; they would drive innovation, solving problems and creating intellectual property.

6. India for India. The most pressing challenges India faces—in healthcare, agriculture, education, and governance—are also opportunities to build AI models tailored to local contexts. Solutions developed for India can then be exported to other countries facing similar challenges. This is not a diversion from global ambitions; it is a pathway to them.

The Reinvestment Imperative

Realising this vision requires reinvestment at scale. The authors are explicit: firms must shift from billing hours to pricing outcomes. They must invest 1-2 per cent of revenue into defensible intellectual property and research and development. They must rebuild their operating models around AI rather than bolting it onto existing structures.

This is not easy. It requires courage to cannibalise existing revenue streams, patience to wait for new ones to mature, and discipline to maintain focus on long-term goals while managing short-term pressures. But the alternative—defensive reaction, clinging to outdated models, hoping that the disruption will pass—is not viable. Firms that choose that path will be commoditised, their margins compressed, their relevance diminished.

Conclusion: The Industry That Refuses to Die

The Indian IT industry has defied predictions of its demise before. It survived automation, cloud computing, and the rise of platforms. Each time, it reinvented itself, finding new sources of value, new markets to serve, new ways to compete. The authors’ confidence that it can do so again is not naive optimism; it is based on a track record of resilience and adaptation.

But the current moment is different. The speed of change is unprecedented, compressing adoption cycles from years to quarters. The nature of the change is structural, rewriting the fundamental economics of the industry. The opportunity, for those who seize it, is enormous—a path to $750-850 billion in annual revenue by 2035, a sustained share of GDP, and a position as a global leader in AI-native services.

The obituaries are being written again. They are, as always, premature. The industry that refuses to die is preparing for its next great reinvention.

---

Q&A Section

Q1: What is “agentic AI,” and why is it considered a more consequential inflection point for the IT services industry than previous disruptions?
A1: Agentic AI refers to autonomous systems capable of executing multi-step workflows without continuous human supervision. Unlike previous generations of automation, which handled discrete, well-defined tasks, agentic AI can plan, execute, and adjust sequences of actions across multiple systems. This fundamentally alters the economics of software development, testing, and operations—the core of the IT services industry. The shift is from a model of labour-intensive ‘effort’ to one of high-impact ‘outcomes’. Previous disruptions (automation, cloud, platforms) were absorbed through incremental adaptation; agentic AI compresses the core of the industry itself, requiring a structural rewrite of business models, workforce composition, and value propositions. The authors describe it as moving from labour arbitrage to intelligence arbitrage.

Q2: What is the distinction between “labour arbitrage” and “intelligence arbitrage,” and why does this shift matter for Indian IT firms?
A2: Labour arbitrage was the foundation of the Indian IT industry’s success: deploying large teams of skilled engineers at costs significantly below Western competitors, with revenue growing linearly with headcount. Intelligence arbitrage refers to capturing value through superior judgment, orchestration, and domain depth—capabilities that AI cannot replicate alone. In this new model, firms succeed not by scaling headcount but by optimising the partnership between humans and AI agents, delivering outcomes that are greater than the sum of their parts. This matters because the old playbook (linear growth, effort-based billing) is becoming obsolete. Firms that continue to bill by the hour will see revenues compress as hours shrink. Firms that shift to pricing outcomes will capture the value of AI-driven productivity gains. The workforce will be smaller but more skilled, productive, and better compensated.

Q3: What are the six strategic frontier plays that the authors identify for India’s IT industry to navigate the AI transition?
A3: The six strategic frontiers are:

1. Agentic AI Scaling: Developing ‘services as software’ (SaS) to deliver non-linear efficiency gains, targeting mid-market businesses and legacy modernisation.

2. Human-tech Hybrid: Designing workforce models that optimise partnership between humans and AI agents, deploying agents for routine tasks and humans for judgment and creativity.

3. Software Reboot: Expanding India’s SaaS market share from ~1% by rearchitecting existing software value pools (CRM, ERP) with AI-native approaches.

4. Infrastructure Hub: Expanding data centre capacity from 1.4 GW to 12 GW, establishing India as a global hub for AI-ready infrastructure and building sovereign cloud stacks.

5. Innovation Hub: Capturing share of the $1.1 trillion global R&D operations market through specialised Centres of Excellence in sectors like medtech and automotive.

6. India for India: Solving local challenges in healthcare, agriculture, and governance to create exportable AI models tailored to local contexts.

Q4: What reinvestments does the authors’ vision require from Indian IT firms, and why are these difficult?
A4: The vision requires firms to: shift from billing hours to pricing outcomes, fundamentally changing their revenue models and client relationships; invest 1-2 per cent of revenue into defensible intellectual property and R&D, moving beyond service delivery to product creation; and rebuild operating models around AI, rather than bolting it onto existing structures. These are difficult because they require courage to cannibalise existing revenue streams, patience to wait for new ones to mature, and discipline to maintain focus on long-term goals while managing short-term pressures. Firms that cling to outdated models will be commoditised; those that embrace reinvention must accept short-term disruption for long-term gain. The authors acknowledge this tension but argue that the alternative is not viable.

Q5: Why are the authors confident that the Indian IT industry can successfully navigate this transition, despite the unprecedented scale and speed of change?
A5: The authors’ confidence rests on the industry’s track record of reinvention. It has survived previous disruptions—automation, cloud computing, the rise of platforms—each time finding new sources of value, new markets to serve, and new ways to compete. This history demonstrates a capacity for adaptation that is not accidental but embedded in the industry’s DNA. The current moment is different in scale and speed, but the fundamental dynamic—technology creating disruption that forces evolution—is familiar. The authors are not promising that every firm will succeed; they are arguing that the industry as a whole can, and that the opportunity for those who seize it is enormous. Their vision of $750-850 billion in annual revenue by 2035, sustaining a 7-8 per cent share of GDP, is ambitious but grounded in a clear-eyed assessment of what reinvention requires. The obituaries are being written again. They are, as always, premature.