Indian IT services companies are changing how they talk about artificial intelligence, shifting the focus from the risk of AI replacing existing work to the revenue opportunities that could emerge as companies deploy the technology at scale. But the transition is creating another risk for the sector as customers seek lower prices in return for AI-driven productivity gains.
The change emerged from commentary during the June-quarter earnings season, with IT companies increasingly discussing AI token consumption, inference costs, model selection and AI operations. At the same time, customers are demanding that service providers share the productivity benefits generated by AI, adding pressure on pricing.
Bank of America Securities said this combination could determine whether rising AI adoption eventually translates into higher earnings for Indian IT companies.
"We believe these pressures could continue to rise over the next one-two years, supported by the rollout of stronger vibe coding tools," BofA Securities analysts Kumar Tayal and Jatin Kalra said in a July 31 report.
The brokerage said the industry's earnings season was "again featured commentary around aggressive pricing and elevated competitive intensity", even as companies identified new opportunities from AI adoption.
AI As A Revenue Opportunity
Indian IT companies have started to position themselves around the infrastructure and operational requirements created by growing AI use rather than focusing only on the threat to their existing business models.
Rising capital expenditure and return expectations are bringing more attention to the hardware side of the AI value chain. Services companies, in turn, have shifted part of their narrative towards what happens as AI moves from experiments to deployment across businesses, according to BofA.
Token consumption is one area emerging as a potential source of revenue for IT services companies as enterprises increase AI adoption. Companies are also pointing to AI operations, including data enhancement, stronger IT security and running AI agents for businesses, as possible sources of demand.
HCLTech said token costs were becoming more relevant as customers moved from pilot projects towards scaled AI adoption. Customers were looking at different AI architectures, including enterprise-hosted small language models and policy-enforced gateways for frontier models, according to the report.
Infosys said its Topaz Fabric allowed clients to use different foundation models while retaining control over data and enterprise knowledge. Coforge highlighted service offerings around "token ops" that could help customers manage token economics.
Wipro discussed token economics as an emerging topic and pointed to AI adoption at scale as an opportunity across ModelOps, AI data centres and secure AI. TCS, meanwhile, identified opportunities across AI governance, lifecycle management and cost control.
BofA said the shift raised a longer-term question about how much of the AI value chain services companies could capture as enterprises use both open-source and lower-cost models.
AI Productivity Gains Put Pricing Under Pressure
The potential revenue opportunity comes with a competing challenge. Customers increasingly expect IT vendors to pass on productivity gains generated by AI, putting pressure on prices.
BofA said the sustainability of the industry's changing AI narrative would depend partly on whether productivity growth could improve from calendar year 2027 onwards.
Infosys said customer expectations on productivity remained high and competitive intensity had resulted in lower prices compared with initial expectations. AI benefits were being incorporated into large deals and non-large-deal work, while the company said it did not pursue a European manufacturing renewal beyond a certain price threshold.
TCS said it was sharing productivity benefits with customers during renewals and, in some cases, during ongoing engagements. Most gains passed to customers were in the 10% to 15% range, according to the report.
Capgemini said customers expected AI benefits to be reflected in prices, while Cognizant said more than 40% of software development was now AI-assisted. Cognizant also pointed to a shift from labour-based models towards agentic and platform-enabled work.
LTIMindtree said productivity-linked pricing discussions with some large clients were behind schedule, while Tech Mahindra said some traditional work would become more productive and require new commercial arrangements.
Mphasis said longer-term pricing would increasingly shift away from input or capacity-based arrangements towards outcome-based pricing where possible.
BofA said stronger coding tools could intensify these pressures further. The brokerage said it would look for evidence of an improvement in growth prospects to show that new earnings opportunities were offsetting the risks associated with AI-driven productivity.
Q1 Eases Near-Term Concerns
BofA said the June-quarter results from Indian IT services companies were broadly in line with expectations, with revenue modestly ahead in most cases.
The results eased some concerns about a sharp deterioration in demand after companies entered the earnings season against a backdrop of macroeconomic uncertainty and questions over tariffs and policy changes.
"Most results reported so far have been in line with, or modestly ahead of, revenue expectations, thereby disproving the more pessimistic scenarios prevalent going into earnings," BofA said.
However, the brokerage said the outlook for the rest of the year remained uncertain. It expects companies to become more selective about discretionary spending and said the environment could remain cautious.
BofA said the positive factors were "incrementally optimistic" about the pace of customer adoption of AI offerings.
The Next Big Triggers
BofA said investors would next focus on US policy developments and a potential Anthropic listing as possible triggers for the IT services sector.
Companies continue to cite policy uncertainty in their main markets as a constraint on demand, according to the brokerage. BofA said a split verdict in the US mid-term elections could change that backdrop, while a continuation of the current sweep could reinforce uncertainty.
A potential Anthropic listing later this year could also bring investor attention back to token economics, profitability and competition from open-source models, the brokerage said.
BofA said it would maintain a selective approach to the sector and retained Genpact, EXL, Coforge and LTTS as its top Buy ideas.
Essential Business Intelligence, Sharp Market Insights, Practical Personal Finance Advice, Daily Fuel, Gold and Silver Prices and Latest Stories — On NDTV Profit.