NextFin News - India may be one of the large economies least exposed to immediate artificial-intelligence job displacement, but that apparent protection is also a test of whether the country can turn AI into faster productivity and broader formal employment. Goldman Sachs estimates that generative AI could substitute for 8% to 12% of India’s non-agricultural employment, while complementing 42% to 48% of those jobs; the balance would remain largely unaffected. The lower substitution share than in advanced, knowledge-intensive economies reduces the risk of a sudden unemployment shock. It does not remove the risk that AI weakens the labor-intensity of India’s most important export engine.
The distinction matters because exposure is not the same as disruption. Goldman’s range refers to employment whose tasks could be substituted, not a forecast that those workers will permanently lose their jobs. In the more aggressive technology scenario, 9% to 17% of tasks performed by India’s non-agricultural workforce could be automated. A task can disappear while the occupation survives, with workers producing more, handling more complex assignments, or moving to a different role. That is why the report’s central message is less an all-clear than a different transmission mechanism: India is more likely to experience job reallocation and productivity gains than an immediate collapse in headcount.
As of August 14, 2026, this is a labor-market analysis rather than a report of a specific trading-day reaction. The Economic Survey 2025-26 says agriculture and allied activities account for 46.1% of the workforce while contributing nearly one-fifth of national income. In urban India, services account for 61.9% of employment. India is therefore partly insulated from generative AI because a large share of employment remains in physical, agricultural, informal, or locally delivered work. Yet its growth strategy depends on expanding precisely the service and technology activities in which AI can perform more tasks.
The result is a two-speed labor market. AI can protect India’s competitive position by making engineers, analysts, service operators, and finance professionals more productive. The same tools can reduce the number of additional workers needed to deliver each dollar of export revenue. India’s lower current exposure is cyclical protection against a technology shock; the challenge of creating enough higher-value jobs is structural.
India’s Lower Exposure Is Real, but Narrow
The first question is simple: why does India appear less exposed than the United States and other advanced economies? The answer is the composition of work. Generative AI is strongest where work is digital, language-based, and organized around repeatable information tasks. It is weaker where work depends on physical presence, dexterity, local relationships, or unpredictable environments. India has a large employment base outside those highly exposed categories.
The Economic Survey’s 46.1% employment share for agriculture and allied activities provides the broadest explanation. Agriculture generates nearly one-fifth of national income but employs almost half of the workforce, a gap that signals lower average labor productivity and a high concentration of work that current generative systems cannot simply perform end to end. In the first two quarters of fiscal 2026, services represented 61.9% of urban employment, showing the other side of the split: cities are already much more exposed to digital tools than the national workforce is as a whole.
Goldman’s India estimates fit that structure. The 8% to 12% substitution-risk range applies to non-agricultural employment, not to every worker in the country. Its 42% to 48% complementarity range is larger because most occupations combine automatable tasks with work that still requires judgment, creativity, accountability, physical activity, or interaction with customers and institutions. The estimate is therefore a map of task exposure, not a headcount alarm.
That distinction also explains why the upper end of the task range, 17%, is not inconsistent with the lower employment-substitution range. A model may automate several tasks inside an occupation without eliminating the occupation. A claims processor may use a model to extract documents and draft a response while retaining responsibility for exceptions. A software engineer may generate routine code while spending more time on architecture, testing, security, and customer requirements. The productivity gain is real only if firms redeploy the saved time into additional output or higher-value services.
The contrast with advanced economies is consequently a composition effect, but not only a composition effect. India’s lower wages can delay substitution if the cost of deploying reliable systems exceeds the cost of human labor. A fragmented employer base, uneven digital infrastructure, and the need for local-language or domain-specific systems can further slow adoption. Those frictions buy time for workers and firms. They do not guarantee that the eventual path will be benign.
The immediate read is protective. The long-term implication is conditional.
The Transmission Runs Through Productivity, Prices, and Hiring
AI reaches the Indian labor market through unit costs before it reaches it through mass layoffs. That sequence is the central mechanism. If a technology-services company can complete more work with the same staff, it can lower prices, expand margins, win more contracts, or shift employees into new offerings. Each outcome has a different employment consequence, and the report’s complementarity estimate assumes that the expansion channel remains strong enough to absorb the substitution channel.
India’s technology-industry data show why the question is urgent. Nasscom’s Annual Strategic Review 2026 estimates that industry revenue will reach $315 billion in fiscal 2026, up 6.1% from $297 billion in fiscal 2025. Employment is expected to rise by 135,000 to 5.95 million, an increase of roughly 2.3%. Revenue is therefore growing nearly three times as fast as headcount. That ratio does not prove that AI caused the difference: demand mix, pricing, acquisitions, and business cycles also matter. It does show that revenue growth is no longer translating mechanically into large hiring gains.
The first-order effect is higher output per worker. The second-order effect is a change in the bargaining relationship between Indian service providers and their global customers. If AI allows vendors to deliver the same process with fewer labor hours, buyers may demand lower prices rather than pay for the full productivity dividend. That can support India’s export competitiveness while limiting wage growth and entry-level hiring. In this scenario, India gains market share and foreign-exchange earnings without receiving an equivalent employment dividend.
The reverse is possible. If lower delivery costs stimulate demand for more software modernization, cybersecurity, data engineering, digital finance, and industry-specific AI systems, total workloads can rise faster than automation reduces labor hours. India’s advantage would then shift from low-cost execution to scalable AI-enabled production. The employment gains would be concentrated in workers who can supervise models, integrate them into business processes, validate outputs, and sell the resulting services.
“Although the impact of AI on the labor market is likely to be significant, most jobs and industries are only partially exposed to automation and are thus more likely to be complemented rather than substituted by AI,” Goldman Sachs Research wrote in its analysis of generative AI and economic growth.
The quote captures the mechanism but leaves open the distributional problem. Complementarity does not mean that every worker benefits equally. A senior analyst who uses AI to manage a larger book of work may become more valuable. A junior analyst whose routine work once served as training may find the entry point narrowed. The aggregate job count can remain stable while career ladders become steeper and wage gains concentrate at the top.
This is where India’s labor abundance changes the macroeconomic stakes. In a tight labor market, firms that automate can bid for scarce workers elsewhere. In a labor market with millions entering working age, automation can raise output without creating enough new positions for new entrants. Goldman’s global framework points to historical reallocation: Joseph Briggs says that around 85% of US job growth over the past 80 years was driven by the creation of new positions linked to technology, while about 30 million jobs are created and 29 million destroyed each year. But the historical lesson is not that adjustment happens automatically. It is that the speed of new-job creation determines whether the transition feels like productivity growth or unemployment.
India’s lower substitution exposure buys time for that adjustment. It does not settle the result.
Why the Services Model Is Both a Buffer and a Target
India’s services economy cushions the country from a near-term AI shock because its national employment base remains less digitized than that of richer economies. It also makes the country unusually sensitive to the next phase of adoption, because services are the channel through which India connects domestic labor to global demand.
The current industry numbers show the asymmetry. A technology sector with $315 billion in estimated annual revenue and 5.95 million employees is economically important, but it is not the whole labor market. Even a major slowdown in technology hiring would not translate one for one into national unemployment because agriculture, construction, retail, transport, public services, and household enterprises employ many more people. Goldman’s estimate that physical work is comparatively insulated reinforces that point.
But the technology sector has an option value beyond its direct payroll. Global capability centers, software exporters, business-process providers, banks, and multinational engineering teams create demand for office administration, transport, hospitality, security, and local services. The official Economic Survey’s urban-services share shows how much of the urban employment engine sits downstream from service concentration. If AI raises the productivity and scale of those centers, the spillover can broaden. If AI instead allows global clients to keep work in-house or consolidate vendors, the spillover can contract.
The second-order risk is not that AI eliminates India’s service exports overnight. It is that the technology changes the type of comparative advantage that customers buy. India built a large export business by supplying skilled labor at lower cost. AI makes raw labor hours less scarce. The premium moves toward proprietary data, domain expertise, trust, integration, computing access, and the ability to take responsibility for an automated process. Indian firms that remain pure capacity suppliers could face price compression even while AI-specialist firms capture higher margins.
India’s policy response is aimed at moving up that chain. The IndiaAI Mission carries a budget of about $1.25 billion, or ₹10,371.92 crore, over five years, covering computing capacity, datasets, skills, startups, and responsible AI. That spending is small relative to the scale of the broader technology industry, but its purpose is not simply to preserve existing service jobs. It is to build the infrastructure and talent needed for Indian firms to own more of the value created by the models.
Goldman’s broader India analysis also points to a talent bottleneck: more than 1.5 million engineering graduates enter the system each year, while the AI sector could create more than 2.3 million job openings by 2027 and the available talent pool may reach only about 1.2 million. Those figures are projections, not realized employment, and they should not be treated as a guaranteed hiring boom. They do show the policy tension clearly. India can have fewer jobs at immediate substitution risk and still face a shortage of workers qualified for the jobs AI creates.
That is the structural fork. The country’s exposure is low because much of its workforce is outside the frontier. Its opportunity is large only if education, training, and firms move workers toward the frontier.
The Counter-Thesis: Lower Exposure Could Mean Lower Productivity
The strongest argument against the optimistic interpretation is that India’s lower AI exposure may be a symptom of low productivity rather than a durable advantage. If nearly half of employment remains in agriculture and much urban work is informal or physical, then AI may have limited immediate scope to raise output where most people work. The economy could avoid mass displacement simply because too few workers perform tasks that advanced models can reach.
That counter-thesis matters for investors and policymakers because a productivity shock concentrated in a small, highly educated segment can widen the gap between GDP growth and broad income growth. Technology exporters may become more competitive, but workers in exposed entry-level occupations may face weaker wage progression. Meanwhile, workers in less exposed sectors may see little direct productivity gain. The result would be a bifurcated economy: high productivity in globally connected firms and low productivity in the larger employment base.
Goldman’s own historical argument is the best answer, but only under a condition. Briggs says the long-run problem is not permanent joblessness; it is whether new jobs are created fast enough to offset transitional headwinds. India’s demographic advantage makes that condition more demanding than in an aging economy. A five-year delay in reskilling can matter more when the labor force is expanding and millions of young people need first jobs.
The falsifying signal for the constructive view is specific: if India’s technology-industry revenue continues to grow at roughly 6% while headcount growth falls below 1% for two consecutive fiscal years, and if the urban services employment share rises without a corresponding increase in regular salaried work, the complementarity thesis would be weakening. That combination would suggest AI is improving vendor productivity without generating enough new demand or formal employment.
The bullish case would be strengthened by the opposite pattern: technology revenue growth above 6%, headcount growth returning above 3%, and a measurable rise in AI-related roles outside a narrow group of elite firms. The key is not the number of AI tools deployed. It is whether deployment expands the market for Indian labor at the same time that it reduces labor hours per task.
India therefore sits in a more favorable position than economies with a higher concentration of automatable knowledge work, but the advantage is a runway, not a destination.
What the Outlook Means Across Time Horizons
Over the short term, the Goldman assessment should reduce the probability of an abrupt India-wide employment shock from generative AI. Physical and agricultural work remain less exposed, and most non-agricultural occupations contain a mixture of automatable and human-dependent tasks. The immediate winners are likely to be firms that can use AI to compress delivery times, improve quality, and win additional contracts without cutting the workforce at the same rate as task automation.
The exposed group is narrower but economically important: entry-level workers in information technology, business-process management, routine finance, content production, and administrative roles. The risk may appear first in hiring, training, and wages rather than in headline layoffs. Nasscom’s projected 2.3% headcount growth against 6.1% revenue growth is consistent with that early phase, although the industry data cannot by themselves identify AI as the cause.
Over the medium term, the outcome depends on demand elasticity. In a base case, AI lowers the cost of Indian services, global demand expands, and the technology industry continues to add jobs, but at a slower rate and with higher skill requirements. The trigger would be continued revenue growth above headcount growth alongside rising hiring in AI integration, cybersecurity, data, engineering, and domain-specialist roles.
In an upside case, India captures more of the value chain. The IndiaAI Mission improves access to compute and data, domestic firms build reusable models and applications, and global capability centers expand higher-value work. The trigger would be sustained growth in AI-related exports and headcount growth above 3% in the technology industry, not merely a rise in AI investment.
In a downside case, global clients use AI to reduce outsourced labor hours faster than they expand total technology spending. Indian vendors face price cuts, entry-level hiring contracts, and the spillover into urban services weakens. The trigger would be two years of technology revenue growth near 6% with headcount growth below 1%, combined with a decline in regular salaried employment in major urban service hubs.
Over the long term, India’s lower exposure can become an advantage only if it converts time into capability. The country has a large labor pool, a substantial technology-export base, and a public AI program with a $1.25 billion budget. Those are inputs, not outcomes. The decisive variable is whether AI complements enough workers to create more output, new products, and new occupations faster than it removes routine tasks.
Goldman’s estimate is therefore best read as a warning against a simplistic AI-job narrative. India is less exposed to immediate substitution than many peers because its economy is less concentrated in automatable knowledge work. But the same structure limits the reach of the productivity boom. India’s AI story will be judged not by how few jobs disappear, but by whether the next productivity cycle creates enough better jobs to make the lower exposure economically meaningful.
India is not insulated from AI; it is earlier in the transition. Its lower displacement risk is valuable only if the country uses that time to build the jobs that automation will demand.
Explore more exclusive insights at nextfin.ai.
