Tata Consultancy Services expects digital and AI-related skills to characterize roughly 80% of new recruitment, after a September quarter in which AI work exceeded 10% of revenue. The change makes AI readiness a broad workforce requirement rather than a small specialist track. It does not mean four out of five recruits will be foundation-model engineers, or that rising AI revenue has already proved higher profitability.
Chief Operating Officer Aarthi Subramanian said TCS added 4,500 people in its fiscal second quarter. About 25–30% were “pure AI” talent, while roughly 80% belonged to broader new-age skill categories. She expected the mix to continue. The specialist population overlaps with the larger digital group; the percentages should not be added as separate hiring buckets.
Hiring follows a larger AI revenue stream
TCS’s September-quarter results put annualized AI revenue at $3.1 billion for the period ending September 30, 2026. That follows more than $2.3 billion at fiscal 2026 year-end and $2.6 billion in the June quarter. Annualized revenue is the current pace expressed as a yearly run rate, not $3.1 billion recognized in one quarter.
Total quarterly revenue was $7.64 billion, up 2.4% year on year in reported dollars. Sequential constant-currency growth was 0.5%. Total contract value was $9.6 billion, excluding the announced Porsche and Best Buy transactions. AI is becoming material, but the much slower company-wide growth shows that it has not transformed the entire business at the same pace.
Subramanian described clients expanding initial AI projects into more functions after measuring their effectiveness. That changes delivery needs: a pilot can rely on a few specialists, while production needs architects, developers, data managers, security staff, domain experts and continuing maintenance. TCS’s banking, manufacturing, retail, healthcare, telecom and public-service engagements require integration with established systems, not chatbot familiarity alone.
Chief Executive K. Krithivasan emphasized willingness to learn rather than age alone. Experienced engineers can understand client systems and industry exceptions better than recruits who know new tools but lack operating context. The staffing challenge is to combine that institutional knowledge with current technical fluency, not choose one generation of workers over another.
A margin thesis still needs evidence
TCS reported a 24% operating margin and 19% net margin for the September quarter. Krithivasan said greater volumes of AI deals should benefit profitability. Reusable components, automation, shared platforms and shorter delivery cycles could reduce work per engagement and spread fixed investment over more revenue.
The fiscal 2026 results provide a comparison: operating margin was 25%, excluding one-time effects. The latest 24% therefore leaves recovery to achieve before AI clearly improves on that recent baseline. TCS has demonstrated AI revenue growth, not isolated its margin contribution after all relevant costs.
Clients may demand part of the savings. If coding assistants reduce required hours, competitive bids and contract terms can pass productivity gains to buyers rather than preserve them for TCS. Outcome-based contracts can reward delivery expertise, but also expose the provider to client data quality, adoption and decisions outside its control.
Costs include recruitment, specialist pay, training time, models, inference, security, evaluation, data preparation and integration. Early deployments often need considerable human review, monitoring and exception handling; regulated sectors add documentation and approvals. AI inside modernization or data work also complicates revenue classification. Management describes its classification as conservative, but an AI label alone does not prove stronger utilization, pricing or profit.
Training totals are preparation, not production readiness
In fiscal 2026 TCS reported 69 million learning hours, up 23%, 5.2 million acquired competencies and more than 270,000 employees with higher AI and machine-learning proficiency. It described role-based learning, an AI learning coach and internal hackathons, with more than 281,000 participants in its largest hackathon. These quantify activity and preparation, not financial return or independently established deployment skill.
The distinction between specialist and broad digital hiring matters here. Specialists may include machine-learning engineers, evaluators, data scientists and AI architects. The wider group includes software, cloud, cybersecurity, data and enterprise-platform professionals using AI in delivery. Enterprise competence requires knowing how generated code affects security, how output is evaluated and when a human must review a consequential decision.
Nikhar Arora of Bots.ai by HR Anexi warned in NDTV’s hiring coverage that adoption dashboards can overstate readiness without shared standards and defined outcomes. That critique matters for a company with 598,056 employees at September-end and trailing 12-month IT-services attrition of 13.3%. Course completions cannot establish reliable work across that workforce.
Client context must be assembled and governed. Retrieval systems need current documents, appropriate access, provenance and monitoring; generic models do not automatically know policies, products or regulatory obligations. Public tools can expose confidential information, and overprivileged agents can act across connected systems. Domain experience and security judgment are therefore complementary to prompt fluency.
A searchable engineering knowledge base can preserve decisions, evaluation results and client constraints across teams. It supports organizational memory, not proof that an AI deployment is safe or economically successful. A hackathon demonstration and a service handling patient data or financial transactions face different demands.
Competitors and concrete deals sharpen the test
Accenture reported about 77,000 skilled AI and data professionals at fiscal 2025 year-end. Its fiscal 2026 earnings disclosure lists $74.2 billion in revenue and $84.5 billion in new bookings across the whole business, not AI alone. HCLTech separately reports advanced-AI revenue, while Infosys organizes enterprise work around Topaz and broad training.
This competition increases demand for scarce specialists and gives clients alternatives when comparing prices and delivery claims. Training the much larger workforce also consumes time that could be billable, while assigning unprepared staff creates execution risk. Utilization and specialist retention, not hiring percentages alone, shape the economics.
The Best Buy agreement offers a concrete expansion model: the companies plan to move the retailer’s Indian global capability center to TCS and develop it into an AI capability center. The transaction combines people, operations and technology rather than selling a stand-alone model.
TCS also announced a five-year Porsche partnership and an agreement to acquire Porsche’s management and IT consultancy MHP, subject to approvals. It points toward AI-supported engineering and mobility with deeper industry context. Transition work, integration, duplicated operations and acquisition expenses may delay full revenue and margin contribution; announced scope is not realized profitability.
Management says enterprise demand is moving toward redesigned value chains and governed agentic systems, where software plans or executes tasks under controls. Yet discretionary programs without near-term value remain under scrutiny. Fast AI growth can partly replace existing services rather than create equivalent net-new company revenue, making the 0.5% sequential constant-currency growth especially relevant.
The economic test links three outcomes
The useful next pattern is rising annualized AI revenue, broader constant-currency growth and improving operating margin together. Only the first would demonstrate a larger AI business, not necessarily the wider productivity thesis. Flat margins would not instantly disprove the strategy, but could show that client pricing or implementation costs absorb much of the benefit.
Workforce evidence should also connect preparation to delivery: revenue per employee, utilization, internal mobility, specialist retention and assignment of trained staff to production work. TCS does not disclose every measure quarterly. Stable definitions of pure AI, broader new-age talent and internally trained employees are needed for comparisons.
For clients, the evidence is shorter implementations, expansion beyond pilot budgets, controlled errors and accountable handoffs when agents encounter exceptions. For employees, durable value combines software fundamentals, trusted data, security and business knowledge with changing AI tools. TCS’s 80% hiring ambition and $3.1 billion AI run rate establish scale. Whether they reinforce overall growth and margins remains the unresolved business question.