The conventional read on AI is that it raises productivity for IT services firms. AI tools let consultants ship faster, write code faster, deliver projects faster — which should translate to better margins, more output, more revenue. That read has been the consensus narrative across analyst coverage of TCS, Infosys, Wipro, Accenture, IBM, and the broader services category for the past 18 months.
The read is wrong by the time the OpenAI Deployment Company and Anthropic's $1.5B services partnership are stood up at scale. The structure of those entities is specifically designed to capture the implementation revenue that has historically flowed to services firms, not to make services firms more efficient. The frontier labs are not partners to the services market. They are now competitors in it, and Indian IT firms face the sharpest competitive shift because their structural position is the most exposed.
This is the battle lines piece. The math is rough but the direction is clear.
The Indian IT Revenue Mix
The structural exposure of Indian IT majors begins with the revenue composition. The largest five Indian IT firms (TCS, Infosys, Wipro, HCL Technologies, Tech Mahindra) collectively post roughly $90 billion in annual revenue, mostly from US and European enterprise customers. The revenue mix is roughly:
- Application development and maintenance: 40-50%. Building and maintaining custom software for enterprise customers. - Infrastructure services: 15-20%. Running and managing IT infrastructure on behalf of customers. - Business process services: 10-15%. Handling specific business processes (claims processing, customer service, finance operations) on outsourced basis. - Consulting and digital: 15-25%. Higher-value advisory and transformation work, growing as the digital share of customer spend increases.
The first three categories are the labor-arbitrage heart of the Indian IT model. The work is done by lower-cost engineers in India for premium hourly rates billed to US/Europe customers. The fourth category is a higher-value play that has been growing as a share of mix.
AI-native services delivery threatens the first three categories disproportionately, and the threat is structural rather than incremental.
The Compression Math
Take a representative enterprise application development project. Pre-AI baseline: 12 engineers for 18 months to deliver a custom enterprise application. Total project cost at Indian IT rates (assume $50 per engineer hour billed): $50 × 12 × 40 hours × 78 weeks = $1.87M. The customer pays this for a 78-week delivery.
AI-native delivery (using Claude Code, Codex, or similar tools at high productivity): 4 engineers for 9 months to deliver the same application. Total cost at the same hourly rates: $50 × 4 × 40 × 39 = $312K. Or at premium hourly rates ($100, reflecting AI-augmented engineers commanding higher prices): $625K.
The compression is real. The customer gets 75-83% cost reduction or some combination of cost reduction and speed improvement. The services firm captures 17-25% of the prior revenue at most.
This is the structural threat. If AI-native delivery becomes the default delivery model and Indian IT firms participate in the transition, their revenue per project compresses dramatically. If Indian IT firms do not participate in the transition fast enough, frontier lab services entities (OpenAI Deployment Co, Anthropic services partnership) capture the AI-native delivery share, leaving Indian IT firms with the shrinking non-AI legacy work.
Either path produces meaningful revenue compression. The first path produces gradual margin compression; the second path produces share loss to new competitors.
Why Indian IT Is More Exposed Than Western Consulting
Concede the case before the takeaway. Accenture, Deloitte, McKinsey, BCG, EY, and IBM Consulting face the same structural threat as Indian IT firms. Why focus the analysis on Indian IT specifically?
Two reasons.
Labor-arbitrage exposure: The Indian IT model is more dependent on labor-arbitrage economics than the Western consulting model. Western consulting firms charge premium rates for high-value advisory work that depends less on labor input volume. AI compression hurts labor-volume-dependent services more than premium-rate advisory services.
Strategic partnership opportunity: McKinsey, BCG, and EY are participants in the OpenAI Deployment Company cap table. This signals an intended partnership model where the Western consulting firms work with OpenAI's services entity rather than against it. Indian IT firms are notably absent from public cap table participation — one Indian IT major is reported as being in late-stage discussions but has not committed publicly. The structural positioning is different.
The exposure is therefore both quantitatively higher (more labor-volume revenue at risk) and strategically less protected (less partnership alignment with the entering competitors) for Indian IT.
The Indian IT Response Already In Motion
Indian IT firms have not been passive about the AI transition. The visible responses across the top five firms through Q1 2026:
TCS: Announced a $5B AI investment program in late 2025, focused on platform development, workforce upskilling, and customer-specific AI delivery capability. The program has been described as the largest in the firm's history.
Infosys: Restructured delivery model to embed AI tools across all project teams. Launched "AI-native delivery" branding for new project engagements. Reports suggest 60%+ of new projects use significant AI augmentation, though the productivity multiplier per project varies widely.
Wipro: Acquired several AI-focused services firms to build implementation capability. Restructured workforce with significant retraining programs. Margin pressure from the transition has been visible in 2025-Q4 and 2026-Q1 earnings.
HCL Technologies: Focused on enterprise AI infrastructure and managed services. Less aggressive on AI consulting positioning, more focused on infrastructure that enterprise customers use to deploy AI internally.
Tech Mahindra: Sub-scale relative to the top four. Has acquired several smaller AI services firms to compete but faces structural challenges of scale.
The responses are real and substantial. The question is whether they are fast enough to capture the AI-native delivery share before frontier lab services entities establish position.
The Three-Way Battle Lines
The enterprise AI services market is now a three-way competition.
Frontier lab services entities: OpenAI Deployment Company, Anthropic's services partnership, possibly future Google equivalent. Positioned on deep model expertise, AI-native delivery, single-vendor coherence. Competing on capability differentiation and speed.
Western consulting firms: Accenture, Deloitte, McKinsey, BCG, EY, IBM Consulting. Positioned on enterprise relationship depth, multi-vendor integration capability, global delivery scale. Competing on customer trust and breadth of service.
Indian IT majors: TCS, Infosys, Wipro, HCL, Tech Mahindra. Positioned on cost efficiency, large engineering capacity, application development depth. Competing on price-to-quality ratio.
The battle lines fall along structural advantages. Each player has a real structural strength, and the market may segment around those strengths rather than consolidating around one winner.
The likely outcome:
- Frontier lab services: Capture 10-20% of total enterprise AI services revenue over 24 months, concentrated in customers wanting deep single-vendor AI commitment and early-adopter strategy. - Western consulting firms: Hold roughly current share (35-45%) as enterprise relationships persist and they upskill on AI-native delivery. Some margin compression but limited share loss. - Indian IT majors: Face the largest share loss (10-20 percentage points), with the lost share roughly split between frontier lab services and the upskilled Western consulting firms.
This is the rough picture. Specifics depend on execution at every layer.
The Indian IT Hedge
Indian IT firms have one structural hedge that is worth flagging. The frontier lab services entities and the Western consulting firms both depend on Indian engineering capacity for their delivery operations. The OpenAI Deployment Company will hire engineers globally, but a significant portion of the 3,000-person headcount will be in lower-cost engineering locations including India. Accenture's $10B AI investment includes major hiring in India.
Indian IT firms could plausibly position to provide this engineering capacity on a wholesale basis to the new competitive entrants. Rather than competing for enterprise customers directly, they could supply engineers to the entities that are competing for enterprise customers.
The economics would be different (lower margins on engineering supply, but potentially higher volume), and the strategic positioning would shift Indian IT from end-customer-facing services firms to delivery-capacity providers. The structural change is significant but not necessarily a loss — it is a reposition.
Whether Indian IT majors choose this reposition or instead double down on direct customer service is a strategic choice that the firms have not yet made publicly. The next 12-18 months will reveal the direction.
The Counterfactual
If the AI productivity dividend went to enterprise customers as cost savings rather than being captured by frontier lab services entities, the IT services market structure would look different. Customers would shrink their implementation budgets, services firms would compress margins but hold share, and the savings would flow to enterprise IT budgets that could be deployed elsewhere.
This counterfactual is what the conventional analyst narrative implicitly assumed. The reason the counterfactual probably doesn't happen: frontier labs have the capability and capital to capture the dividend at the source, before it flows to customers. The Deployment Company structure and the Anthropic services partnership are explicit moves to capture the dividend.
For Indian IT firm equity holders, the implication is to discount the AI-productivity-tailwind thesis that has supported valuations through 2024-2025. The thesis was that AI would lift IT services revenue and margins. The more accurate reading by mid-2026 is that AI is restructuring the market in ways that compress the labor-arbitrage model that Indian IT depends on most.
We read the next 24 months as the period when the structural shift becomes visible in earnings. By 2027, the relative positioning of frontier lab services, Western consulting, and Indian IT majors will be settled. The battle lines are visible now. The outcome will play out across the next eight quarters.