Every day we're going to tear down a major Indian listed company and map the AI risk/opportunity around it - products, competitive shifts, key concall quotes. Day 4: Infosys (INFY) Peak share price: ₹1,999.70 (Dec 2024) Current share price: ₹1,320.90 (-34.0%) EV today: ₹4,89,217 crore Revenue Q3 FY26: ₹45,479 crore ( 9% YoY) Operating Margin: 21.2% (adjusted for labour code charges) EV/EBITDA: 11.57x Headcount: 3,37,000 ( 4.3% YoY) What Infosys does: Infosys manages the critical systems and digital transformation of 1,882 active enterprises across 50 countries, with 27% revenue concentration in BFSI and a 96.9% export revenue profile. Its Global Delivery Model deploys 3,37,000 professionals across application development, cloud migration via Infosys Cobalt, and AI transformation through its Topaz suite. As India's second-largest listed IT company, it is the default navigator for global corporations modernizing legacy stacks at scale - not just advising on transformation but executing it. AI bear case: The primary threat is disruption of the "pyramid" labor structure - junior-level coding, testing, and maintenance are high-volume, low-complexity, and directly in the automation crosshairs. GenAI enables clients to automate these workflows or shift them to captive Global Capability Centers in India, compressing volume in the core cost-of-effort segment that drives 55.4% of revenue. Management has acknowledged that AI productivity gains are frequently shared with clients, creating deflationary contract pricing that requires constant new scope to hold the top line. The labour code charges - ₹1,300 crore impact in Q3 FY26 alone - confirm that the cost of maintaining a specialized human talent pool remains high even as billable hours for entry-level work decline. Attrition at 14.3% adds further cost volatility to an already pressured margin structure. AI bull case: Infosys has delivered over 400 GenAI projects and carries 4,600 active AI engagements in pipeline - including 500 purpose-built agents for multi-step business process execution. The Topaz Fabric composable stack includes 50 purpose-built agents for IT operations and 20 managed connector platforms for enterprise data. Proprietary small language models for banking and cybersecurity reinforce the moat as an enterprise-grade integrator for the 90% of its top 200 clients already engaged in AI work. The $1.6 billion NHS contract is the clearest evidence that Infosys can embed AI into massive, long-term operational modernization programs where hyperscalers cannot simply walk in and displace the incumbent. Distribution advantage is real. AI traction: 4,600 active AI projects in Q3 FY26, up from 2,500 GenAI and 200 agentic AI projects just one quarter prior - the acceleration is disclosed, not extrapolated. Topaz Fabric launched with Cognition's Devin software agents integrated and deep co-innovation with Intel on Gaudi AI accelerators. 2,70,000 employees now AI-aware; 10% of top technology talent engaged in highly innovative AI solution building. Specific AI-attributed revenue is not yet isolated in filings - the one disclosure gap that matters most for the investment thesis. Competitive landscape: TCS has already disclosed annualized AI services revenue exceeding $2.3 billion - a concrete benchmark Infosys has not yet matched with a comparable disclosed number. Accenture and Capgemini are executing the same agentic AI pivot globally with larger consulting revenue bases. The structural threat from client-owned GCCs in India is accelerating independently of any competitor move - this is demand destruction that no IT major can outrun through better sales execution alone. The race to agentic frameworks is now table stakes across the sector. Management quotes: "So, what we have built within Topaz is fabric, Topaz Fabric, as we call it, is a set of purpose-built agents, which work with many different native AI companies interfaces and can support a lot of different functions within clients, horizontal and vertical." - Q3 FY26 Concall "If you are delivering better productivity, you would be able to keep some of that with you, which will get reflected in better pricing." - Q3 FY26 Concall "So on the headcount increase, I think it demonstrates that we have confidence in where the market is and what we are seeing in terms of the demand." - Q3 FY26 Concall Commentary: The ₹1,300 crore labour code hit distorted Q3 FY26 optics - the underlying story is 9% rupee revenue growth and a 21.2% adjusted margin holding firm. Infosys is defending territory effectively during a period of massive architectural upheaval. AI is a net tailwind over a 3-year horizon because orchestrating thousands of agents across fragmented legacy enterprise landscapes at the scale of 1,882 active clients is exactly what a large incumbent does and what an AI-native startup cannot. The unresolved question is whether the "sharing" of productivity gains triggers a race to the bottom in standard digital services contract values - that is the deflationary trap hiding inside every renewal. In the next two concalls, watch utilization rates and net additions to the "AI masters" talent tier. If the 20,000 fresher hiring target is replacing automated roles rather than generating new revenue, the pyramid model is already hollowing out faster than Topaz Fabric can fill it. Analysis by CompoundingAI Source : All figures sourced to BSE filings and Q3 FY26 earnings transcripts Note : No Buy/Sell Reco. Only for educational purposes.