How AI is tilting India’s IT balance

An average Indian IT company with $1 billion in revenues employs nearly 30,000 people. But, in three years, a $1-billion IT services company could be staffed by 1,000 full-time equivalents (FTEs), predicts Ray Wang, CEO of Constellation Research. “A 100-person software company will take out a $100 billion (annual recurring revenue) company in the next three years,” he said.

The typical playbook of hiring software engineers offshore to drive technology projects at low costs is being rewritten rapidly with AI. With Agentic AI altering how software gets built and managed, the traditional labour-heavy model that built India’s software exports over the past two decades is being challenged.

Wang’s comments capture the seismic shift the industry will undertake, as AI makes companies chase ‘exponential efficiency’ in the services they offer, eventually having to cannibalise their own business models before someone else does.

Labour arbitrage

“Most of the profit pool of the entire industry is in the software development life cycle,” said Peter Bendor Samuel, CEO of Everest Group. “About 85% of profit in the entire IT services space is made there, and the key to that is through a labour opportunity.”

The traditional pyramid, led by thousands of low-level (L1 and L2) engineers delivering volume-based work, has been the bedrock of profitability. As AI takes over testing, coding and maintenance tasks, companies are already reporting productivity jumps of 45-50%, according to Samuel.
“The existential threat that AI poses is that we won’t have this L1 and L2 (engineers). This just blows up the entire TCS, Wipro, Infosys, TechM—to some degree Accenture’s—profit model,” he added.

In the short term, this means that how companies look at their deal structures and revenues will change. Sequentially, top IT firms like TCS, Wipro, Infosys, among others, have been reporting sluggish growth, with the $280-billion industry reporting a ‘flat’ second quarter in FY26, growing between 2% and 4%.

With AI, according to Phil Fersht, CEO at HFS Group, companies would not be looking at sudden collapses in revenues, but rather a reshaping of deal structures. “The classic $100-million transformation contract is harder to find because clients are using AI and cloud to take out cost in smaller increments,” he said. In practical terms, this means the next 12-18 months will see large, multi-year deals giving way to mid-sized contracts and shorter engagements.

Agile sustainability

While revenues, Fersht said, are “flattening in the short term”, growth will eventually come from new lines such as AI-first services, data modernisation and orchestration work. The impact, however, is non-uniform, subject to the nature of the deals and how much companies are willing to reinvent.

Among mid-tier firms, agility could be a differentiator. Whether it’s LTIMindtree’s digital focus, Persistent’s deep engineering, or UST’s track record of reinvention, the firms could face both sharper risks and opportunities, depending on their ability to reposition around AI-first services, Fersht added.

Despite the varied impact points, industry experts say no company is better placed than others when it comes to revenue compression driven by AI. As the cost of technology decreases, the industry is expected to see new demand, according to Samuel. “But whether you build more (software/AI) platforms or drive more volume on the platforms... both will drive you towards (vendor) consolidation,” he said.

And along with the impact on the delivery model, how deals get signed on in the first place is changing, too. AI-led efficiencies are now a “blanket driver” for reductions in contract values, according to Yugal Joshi, partner at Everest Group, with the price cuts across the board ranging between 20% and 30%. “One big reason is perception,” Joshi said. “Clients see providers as cost-reducing partners. This is how the entire industry has positioned itself, and this is how outsourcing works.”

Even for discretionary spends, service providers are under pressure to deliver faster and cheaper. “Service providers are committing to better time-to-market and reductions in contract value, thanks to their AI tools,” Joshi noted. “The unfortunate thing about services is that, unlike technology, where innovation makes clients pay more, for services it’s typically the other way round.”