India’s Big IT Claims “AI-First”, Disclosures Show Otherwise

The Indian IT services sector has had a rough run into 2026. The Nifty IT index has fallen nearly 22% over the past year, more than the broader Nifty 50, and the price-earnings multiple has compressed to around 20 times against a ten-year median closer to 25. Sell-side notes have moved from "transition year" to "structural reset" since the December quarter.

Customers are deferring discretionary projects while they wait to see whether agents from Anthropic, OpenAI and Google can do the same work for less, and faster, and model creators have begun selling those agents directly into client organisations. Pricing power, billable hours and headcount-led growth are all under review at once, which is part of why the sell-side reframing has moved through the sector so quickly.

Every member of Indian IT’s Big 6—TCS, Infosys, Wipro, HCL Tech, LTI Mindtree, Tech Mahindra—has spent the past two earnings cycles describing itself as “AI-first”, with frameworks, marketplaces, agent counts and partnership press releases to match. 

Investors have heard these claims but are not yet paying the multiple that would normally accompany them, and the quarterly filings, AGM material and partnership disclosures help explain why.

Infosys carries the most disclosed AI revenue line of the six. Topaz, its enterprise generative-AI suite, contributed 5.5% of December-quarter revenue across 4,600 projects and more than 500 agents built, with Anthropic and OpenAI named as partners in the AI First Value Framework, though commercial terms are not disclosed. The investor and employee narratives still trail the framework's ambition.

LTIMindtree, now LTM Limited, has been the loudest voice of the six; the audited numbers appear to back the rhetoric: Revenue in 2025-26 up 11.3% year on year,  EBIT margin 14.3%, a $450 million agribusiness contract named, internal agentic AI platform BlueVerse hosting more than 1,500 internal agents; and 90% of an 88,000-strong workforce in AI courses. One gap: no named lab partner of the Anthropic-OpenAI-DeepMind tier.

Tech Mahindra has shipped Orion on NVIDIA's stack with 300-plus pre-built agents, is a Google Cloud Premier Partner, has contributed to the India AI Mission, trained 77,000 employees, and filed the VerifAI patent. That is a fuller AI build-out than Tech Mahindra's share of voice in the sector would suggest.

TCS owns the loudest single AI deal any Indian firm has made: the 1GW HyperVault-OpenAI build. It also disclosed $2.3 billion of annualised AI revenue for the January-March quarter of 2025-2026. But the number is top-down rather than product-attributed in the Topaz manner, the GCC channel is thinner than at HCL Tech or Wipro, and the 2% headcount cut is action without a public reskilling pitch attached.

Wipro's named GCC wins, including a Southeast Asian manufacturer for asset operations and the Olam strategic deal, and the new AI Native Business & Platforms unit pitched as a services-as-a-software pivot are the disclosed pieces. Beyond these, the rest is mostly framework slides such as named methodologies, capability maps, partnership logos without disclosed revenue lines, named contracts, or quantified agent counts of the Topaz / BlueVerse / Orion variety.

HCL Tech has the largest GCC footprint among the six with more than 200 disclosed engagements, a named NVIDIA-powered AI Lab built for a US telecom client, and annualised Advanced AI revenue of $620 million in the January-March of 2025-2026. That disclosure sits between Topaz's product-attributed share and TCS's top-down aggregate.

Beyond the disclosure question sits a structural one, which is that the Big 6 are no longer the default route between a global enterprise and an enterprise-grade AI deployment because model creators are increasingly selling that deployment as a product of their own.

For example, Google's Gemini Enterprise Agent Platform, launched at Cloud Next 2026, lets a company create, deploy and supervise autonomous software agents that act across its systems for hours at a time. The firms most exposed are the ones whose AI-first pitch leans hardest on the GCC channel: Wipro most of all, with HCL Tech also exposed though partly cushioned by its NVIDIA-anchored AI build pipeline. On the other hand, Tech Mahindra, a Google Cloud Premier Partner, sits closest to the Gemini stack and could yet flip the threat into channel revenue.

The compression in the Nifty IT P/E is the discount the market is applying to firms whose AI-first claims have not yet shown up in disclosures of the right kind. Infosys and LTM clear that bar most cleanly on product-attributed revenue and named contracts. Tech Mahindra and HCL Tech cover more ground than their brand voices suggest, the first on agents and partnerships, second on a $620 million Advanced AI line and a named NVIDIA build. TCS has the largest absolute AI number but the disclosure is top-down rather than product-attributed. Wipro has the most work to do between now and the first quarter earnings for 2026-27.

(This column reflects the author's personal views and is based on publicly available information. It is intended for general commentary and analytical purposes only and should not be construed as investment advice.)