AI growth has not lifted the whole sector
India’s largest IT services companies produced a divided scorecard in the latest reported quarters. TCS grew revenue by 0.5% sequentially in constant currency, HCLTech by 2.3%, while Wipro was flat. Cognizant’s June-quarter revenue rose 4.1% from a year earlier in constant currency. A published summary of Infosys’s September quarter put its sequential constant-currency growth at 1%.
The figures support a structural slowdown thesis, but not a uniform decline. Clients are buying artificial intelligence work, cloud modernisation and cost reduction at the same time as they restrain discretionary programmes. The resulting contracts can be large without producing the rapid revenue growth that followed the pandemic-era technology expansion.
HCLTech offers the clearest evidence that AI spending can become incremental revenue. TCS is building a sizeable AI practice, but its group growth remains subdued. Cognizant is executing earlier large deals, while Wipro’s weak bookings show that an AI-labelled pipeline does not guarantee growth. Infosys sits between those cases. The industry is evolving, but most providers remain outsourcing businesses using AI rather than AI-led businesses with distinct economics.
The latest results show that AI demand is growing much faster than the companies built to serve it, but is not yet large enough to restore broad-based growth.
AI is becoming material inside delivery before it becomes decisive in company-wide growth.
AI revenue is real but poorly comparable
TCS said annualised AI revenue reached $3.1 billion in the September quarter, up from $2.6 billion in the preceding quarter and above 10% of revenue. HCLTech reported $219 million of quarterly “advanced AI” revenue, up 28.4% sequentially and 78% year on year in constant currency. The available Infosys results summary put AI revenue at 8.2% of revenue in the June quarter.
Those disclosures are progress, but they do not establish a common market. TCS uses an annualised run rate, HCLTech reports quarterly revenue under a narrower label, and other providers disclose engagements, partnerships or productivity indicators. Wipro did not attach a comparable revenue figure to the AI projects described in its September-quarter materials. Cognizant said more than 40% of its software development was AI-assisted and cited more than 8,000 AI engagements, but did not report AI revenue separately.
Investors therefore cannot yet separate new AI budgets from cloud, data and application work that has been reclassified or expanded to include AI. AI is becoming material inside delivery before it becomes decisive in company-wide growth.
TCS and HCLTech have put revenue against their AI claims, but differing definitions prevent a clean comparison and leave open how much work is genuinely new.
Large deals do not settle the growth question
HCLTech paired its AI revenue increase with $3.26 billion of new bookings, up 27% from a year earlier. TCS recorded $9.6 billion of total contract value. Its announced work with Porsche and Best Buy also links AI to wider operating and technology change, rather than a stand-alone experiment. These are the strongest signs that AI can enlarge contracts.
The opposing evidence is substantial. Wipro’s total bookings fell 22.9% sequentially in constant currency to $2.60 billion, while large-deal bookings dropped 58.6% to $674 million. Cognizant’s trailing 12-month bookings rose 5% to $29.1 billion, but its quarterly bookings declined 6% from a year earlier. Infosys’s published Q2 summary showed $3.6 billion of large-deal value, without proving that AI had increased the total.
Contract value is not revenue, and providers define bookings differently. Large contracts often consolidate suppliers, renew existing work or promise clients savings. AI may help a provider win such a contract while reducing the labour and revenue needed to deliver it. The booking can defend market share without expanding the industry’s revenue pool.
Bookings favour HCLTech, but deal values across the sector include renewals and cost-saving work that can replace, rather than add to, existing revenue.
The case against structural decline
The best case against the structural slowdown argument is that the industry is between investment and conversion. Enterprises first need data preparation, cloud infrastructure, governance and security before deploying AI at scale. Much of that work resembles traditional services and enters reported revenue gradually. TCS’s AI run rate and HCLTech’s 78% year-on-year advanced AI growth suggest that conversion has started.
Large contracts can also create durable revenue even when their initial aim is efficiency. Cognizant’s trailing book-to-bill ratio was about 1.3 times, while seven June-quarter deals exceeded $100 million each. HCLTech combined record first-half bookings with a 131 basis point sequential margin expansion to 18.2%. On these figures, AI is not merely cutting billable effort. It is also helping some companies win and deliver broader transformation work.
That argument weakens, rather than overturns, the structural concern. TCS’s group growth remained slight despite its AI scale. Wipro guided to between a 1.5% decline and 0.5% growth for the December quarter. If AI revenue mostly substitutes for older application and support work, rapid growth in the new category can coexist with weak growth for the company.
The strongest argument against a structural slowdown is that productivity-led deals take time to convert, and early revenue measures at TCS and HCLTech are accelerating.
Productivity is not a new business model
Traditional outsourcing links revenue to people, effort and contract duration. Generative AI can produce code, test software and handle support requests with fewer hours. Providers may retain part of the saving through higher margins, but clients will seek lower prices when contracts renew. Revenue can therefore lag productivity even when demand for technology work remains healthy.
The five companies are responding through platforms, partnerships and acquisitions, yet their disclosed economics still depend mainly on services contracts. TCS and HCLTech have moved furthest in measuring AI revenue. Cognizant has shown AI-assisted delivery and improving revenue per employee. Infosys has disclosed a growing AI share, but the available evidence does not isolate how much is incremental. Wipro has described AI-enabled projects without matching them with stronger bookings or near-term growth.
A genuine shift would require three results at once: AI revenue rising faster than the old business, group growth accelerating, and margins holding after clients receive productivity savings. No company has yet demonstrated all three consistently. Cognizant’s next quarterly release will test whether its large-deal execution can sustain organic growth. The Indian providers’ December-quarter results must then show whether rising AI revenue adds to total revenue rather than merely replacing work that automation has made cheaper.
The sector can become more productive without acquiring the recurring, proprietary revenue or pricing power associated with an AI-led technology business.
