India's Outsourcing Giants Face a Dual AI Reality

The Contradiction at the Core

Tata Consultancy Services and Infosys published third-quarter results that illustrate a fundamental tension reshaping India's IT services sector. Between July and September, TCS recorded annualised AI revenue growth above 19 per cent, while Infosys saw AI-related income climb 29 per cent quarter-on-quarter. Yet overall revenues remain under pressure, constrained by the same force driving the new growth.

At Opentechwire, we've tracked this contradiction across Bengaluru's largest employers for eighteen months. Automation and generative AI tools are simultaneously creating lucrative transformation contracts and reducing the billable hours needed to maintain existing software estates. The result is a business model in flux, where firms must chase new high-margin work fast enough to offset margin compression in their legacy book.

Where the Growth Is Coming From

The AI revenue spike reflects a specific category of engagement. Enterprises in North America and Europe are paying TCS, Infosys and peers to embed large language models into decades-old ERP systems, customer databases and supply-chain platforms. These projects command premium rates because they require both domain expertise and familiarity with fragile legacy code that in-house teams are reluctant to touch.

Infosys reported that its AI practice grew nearly 30 per cent in the three months to September, driven by clients in financial services and retail who want conversational interfaces layered over core transaction systems. TCS attributed its 19 per cent annualised increase to a pipeline of generative AI proof-of-concept deployments that have moved into production, particularly in banking and insurance.

Both firms have built dedicated AI units over the past two years, staffing them with machine-learning engineers and product managers who can scope use cases, fine-tune foundation models and integrate APIs from providers such as OpenAI, Anthropic and Cohere. The work is technically distinct from traditional application maintenance, and clients are willing to pay a margin premium for speed and specialised skill.

The Squeeze on Traditional Segments

The same technology delivering those premium contracts is eroding the base. AI-led productivity gains mean fewer engineers are required to perform routine coding, testing and helpdesk tasks that have historically formed the bulk of India's IT export revenue. A maintenance contract that once needed a team of twelve can now run with eight, and clients are renegotiating terms to reflect the reduced headcount.

TCS and Infosys have both acknowledged in recent earnings calls that automation is compressing billable hours in application management services, their largest revenue segment. Clients are using code-generation tools to handle repetitive tasks, and vendors are deploying internal AI assistants to accelerate ticket resolution and regression testing. The efficiency is real, but it translates directly into lower contract values when deals come up for renewal.

This dynamic is not unique to India. Global IT service providers from Accenture to Cognizant face the same pressure. But the exposure is acute in Bengaluru, Pune and Hyderabad, where employment models have been built around large teams performing high-volume, moderate-skill work. As that work becomes automatable, the competitive advantage shifts toward firms that can deliver transformation projects rather than steady-state operations.

Margin Structure Under Revision

The dual effect is reshaping the economics of India's IT services sector. High-margin AI transformation contracts are growing, but they are smaller in absolute revenue than the traditional segments they are meant to replace. A generative AI implementation might generate USD 5 million in fees over six months at a 40 per cent margin, while the application maintenance contract it displaces was worth USD 12 million annually at 25 per cent margin. The new work is more profitable per dollar, but the total dollar pool is shrinking.

Analysts in Mumbai and Singapore have noted that TCS and Infosys are managing this transition better than mid-tier competitors, thanks to scale, brand recognition and early investment in AI capability. Both firms began hiring machine-learning specialists and building partnerships with foundation model providers in late 2024, giving them an eighteen-month head start on smaller rivals. That timing is now visible in their AI revenue figures.

Still, the overall revenue outlook remains cautious. Neither company has forecast a return to the double-digit growth rates that characterised the sector before the pandemic. Instead, management commentary emphasises "resilience" and "strategic positioning", language that reflects an industry recalibrating expectations in the face of structural change.

What the Workforce Transition Looks Like

The employment implications are significant. TCS and Infosys have slowed campus hiring over the past year, even as they continue to recruit AI and cloud specialists through lateral channels. The firms are also running large-scale reskilling programmes, attempting to retrain application developers and support engineers in prompt engineering, fine-tuning workflows and API integration.

The success of these programmes will determine whether the workforce transition is managed smoothly or results in a wave of involuntary attrition. Historically, India's IT sector has absorbed technology shifts by moving people from declining skill areas into emerging ones. But the current transition is faster and more technically demanding than previous cycles, and not every mid-career developer will make the leap.

Trade unions and labour researchers in Karnataka have raised concerns about the social cost of automation in a sector that employs more than five million people nationwide. So far, the major firms have avoided large-scale layoffs, preferring to manage headcount through attrition and selective hiring freezes. But if AI-driven productivity gains accelerate faster than new project wins, that calculus may change.

The Regional Context

India's IT services sector is not the only one navigating this tension. Across Asia, software and business process outsourcing providers are grappling with the same dual reality: AI creates new revenue opportunities while eroding the economic foundation of their existing business. Firms in the Philippines, Vietnam and Malaysia that handle customer support and back-office work are seeing similar margin pressure as conversational agents automate tier-one support and document processing.

The difference is that India's largest players have the capital and technical depth to pivot toward higher-value AI services, while smaller regional competitors often lack the scale to build dedicated practices. TCS and Infosys are investing hundreds of millions of dollars in training, partnerships and product development, betting that they can capture a disproportionate share of the transformation market even as the overall revenue pool contracts.

Whether that bet pays off will depend on how quickly enterprises move from proof-of-concept AI projects to production deployments at scale. If adoption accelerates, the pipeline of high-margin transformation work could sustain growth for several more quarters. If it stalls, the revenue squeeze from automation will dominate, and the sector will face a prolonged adjustment.

A Sector in Transition

The third-quarter results from TCS and Infosys offer a snapshot of an industry caught between two futures. One is defined by premium AI transformation contracts that leverage India's engineering talent and enterprise relationships. The other is defined by relentless automation that reduces the need for the labour-intensive services that built the sector in the first place.

For now, the firms are managing both realities simultaneously, using AI revenue growth to offset declines in traditional segments and buying time to retrain workforces and reposition their offerings. But the underlying tension is structural, not cyclical, and it will shape the trajectory of India's IT sector for years to come. The question is not whether AI will transform the business model, but whether the transformation happens fast enough to sustain employment and revenue at current levels.