The traditional Indian IT services model — high-value client engagement and consulting on US soil combined with large-scale execution from lower-cost Indian centers — has powered decades of growth, employment for millions, and substantial export earnings. That model is now under simultaneous strain from tighter US immigration rules and the rapid commercialisation of agentic AI. The pressures are real and consequential.
The Visa Squeeze Is Real — But Exposure Has Already Shrunk
In early October 2026 the Trump administration indefinitely suspended several major firms — including Tata Consultancy Services, Infosys, Wipro, HCL Technologies, Cognizant, Capgemini, Microsoft and Adobe — from the Permanent Labor Certification (PERM) programme used to sponsor foreign workers for employment-based green cards. Officials cited alleged abuse of the system and prioritisation of American workers. This follows earlier measures that raised costs sharply for certain new H-1B petitions.
Indian IT firms have already reduced dependence on the classic onsite-offshore pipeline. Industry and analyst data indicate that only about 20% of a typical large vendor’s workforce is based onsite in the US; of those, a minority hold H-1B visas, so overall H-1B exposure sits in the 3–5% range of total headcount. H-1B registrations by the major Indian IT companies fell roughly 92% for the FY27 cap. PERM filings by these firms had already declined to a very small share of the US total. Most software and IT-enabled services exports continue to be delivered from India.
The freeze complicates long-term residency pathways for existing H-1B holders and raises retention and local-hiring costs in the United States. It does not cancel current visas or stop offshore delivery. Some industry voices note that higher US costs and tighter mobility can accelerate the shift of more work to India rather than reverse it. A large-scale sudden return of “20% of employees” is not supported by the actual onsite/H-1B numbers; the pool is smaller and the industry has already localised and offshored more heavily.
Agentic AI Changes The Productivity Equation
Agentic AI systems that can plan, generate, test and iterate on code and workflows are maturing commercially. This compresses the volume of routine, entry-level and mid-level coding, testing and support work that historically absorbed large numbers of fresh graduates and junior engineers. Campus hiring has been subdued for several years as firms prioritise productivity and higher-skill roles. Demand for specialised profiles — Agentic AI engineers, orchestration and reliability roles, AI governance, data readiness — has risen sharply.
The economic logic is straightforward: one skilled professional supervising or directing capable agents can deliver output previously requiring many more people. Successful adoption therefore tends to reduce required headcount for a given volume of work even as it creates new categories of higher-value work. Overall industry revenue continues to grow, though more slowly than in earlier boom years, and the historical tight linkage between revenue growth and headcount growth has loosened.
This is a genuine productivity shock, not merely incremental tooling. Entry-level pipelines that once absorbed large numbers of engineering graduates are under pressure. At the same time, the technology does not eliminate the need for domain expertise, system design, quality oversight, security, client understanding or the orchestration of complex enterprise environments.
Absorption Challenges And India’s Broader Labour Reality
India already faces elevated youth and graduate unemployment relative to the headline national rate. Skills mismatch, the sheer volume of annual engineering and other graduates, and the quality of many jobs outside the formal high-skill sector are longstanding issues. Returning experienced professionals from the US can bring valuable client-facing, domain and process knowledge that benefits global capability centres (GCCs), product companies, startups and domestic digital transformation projects. They cannot, however, magically create millions of additional high-quality roles overnight in an economy already struggling to absorb new entrants.
The larger question is structural: an industry that once expanded employment roughly in line with revenue is shifting toward higher productivity and outcome-based models. Revenue can grow while net hiring slows or becomes more selective. This severs the earlier automatic link between IT export growth and middle-class job creation on the previous scale. Complementary growth in manufacturing, domestic deep tech, R&D, and other high-skill sectors will be necessary to absorb talent over time.
Adaptation Is Underway, Not Guaranteed
Indian IT companies are responding with localisation in the US, greater offshore delivery, heavy investment in AI platforms and large-scale reskilling programmes. GCCs continue to expand and increasingly seek AI, cloud, cybersecurity and analytics talent. The shift from pure labour arbitrage toward higher-value engineering, product work and AI-enabled services is visible in strategy statements and early revenue disclosures. Success is not automatic. It requires continuous upskilling, stronger alignment between education and industry needs, and policy support for domestic innovation ecosystems.
Projections of 80–95% unemployment across IT and related professions by the end of 2026, or of an irreversible macroeconomic catastrophe driven solely by these two forces, is supported by current evidence. The dual squeeze is serious. The traditional headcount-driven model is under lasting pressure. The outcome will depend on the speed and quality of adaptation — by companies, by educational institutions, and by the broader economy — rather than on the inevitability of mass displacement without replacement. The next few years will test whether India can convert a challenging transition into a more resilient, higher-value technology ecosystem.