TCS, Infosys Hit 5-Year Lows as OpenAI Kills Headcount Model

The Nifty IT index is down 28% in 2026 against a Nifty 50 that’s off 4%. This is not a sector rotation. The options chain confirms institutional desks are positioned for more pain, not a recovery.

NiftyTrader Editorial & Trade Desk | May 20, 2026 | 08:10 IST | Not investment advice—editorial positions only

---

The Numbers the Market Is Already Acting On

Infosys closed at ₹1,135 on May 19, 2026, its lowest closing price since December 2020, per NSE end-of-day data. TCS settled at ₹2,310 the same session, a level last seen in August 2020. Wipro has shed 34% of its market value in five months. The Nifty IT index is down 28% year-to-date against a broader Nifty 50 that is off 4.1%, per NSE index data published May 19.

That 24-percentage-point spread between sector and index is not what you see in a demand slowdown. It is not what you see when clients delay IT spending because of macro uncertainty. It is what you see when the market is repricing the fundamental economics of a business model.

For historical context: the 2022 rate hike selloff, which at the time triggered every “Indian IT is cheap” call you remember reading, produced a 22% peak-to-trough decline in the Nifty IT index over seven months, per NSE historical data. The current drawdown is already deeper, and it has arrived in five months. The 2022 move was cyclical. Rates went up, deal closures slowed, and clients deferred spending. The IT stocks recovered because the delivery model was intact. This time, the delivery model is what the market is questioning.

Every mainstream outlet is writing “IT stocks fell on AI concerns.” That is a euphemism. What institutional positioning, visible in the NSE derivatives market, actually shows is a sector being repriced as something structurally different from what it was eighteen months ago.

What OpenAI’s Enterprise Push Actually Did to the Market

In May 2025, OpenAI formally expanded its enterprise deployment operations, moving beyond software licensing toward a model where OpenAI’s own technical staff work embedded inside client organisations, not as consultants who hand off a product but as the ongoing delivery team. In early 2026, this model was publicly confirmed to be operational at multiple Fortune 500 clients in financial services and healthcare, per reporting by The Information and Bloomberg Technology.

The market began pricing this seriously in late March 2026. The Nifty IT index accelerated its decline through April and into May as deal-level intelligence from US and European enterprise markets began suggesting that clients were explicitly asking incumbent IT vendors, including Infosys and Wipro, whether their engagement teams could be replaced by smaller, AI-native teams producing equivalent output.

Infosys dropped 4.1% in the week of May 12–16, 2026, per NSE weekly return data. TCS declined 3.6% the same week. HCLTECH fell 2.8%. The Nifty IT index moved to a new 2026 low during that week and has not recovered.

The market did not react to OpenAI’s enterprise expansion as a surprise. It reacted as confirmation of a thesis that institutional desks had been building positions around since Q1 2026. The derivatives data, which we will get to, makes this sequencing clear.

The Math Behind the Headcount Model’s Problem

Indian IT built its competitive advantage on one equation. Hire engineers in Bengaluru and Hyderabad at ₹6–12 lakh annually. Bill those engineers to US and European clients at $80–120 per hour. Pocket the arbitrage across a workforce of hundreds of thousands. At scale, it produced some of the most consistent margin profiles in global services.

The FY25 numbers, from company annual reports, show where that model stands today:

Stock Performance & Valuation Snapshot — May 19, 2026

Source: NSE closing data May 19, 2026; FY25 annual reports (revenue and margin); Bloomberg consensus estimates (forward P/E)

Also Check: NSE – TATA CONSULTANCY SERVICES Option Chain

Revenue per employee at TCS in FY25 was approximately $28,400, per the company’s FY25 annual report. At AI-native services firms operating today, like Cognition AI, Poolside, and Scale AI’s services arm, industry analysts at Bernstein and Jefferies have cited revenue-per-employee figures in the $500,000–$800,000 range in published research from Q4 2025 and Q1 2026. That gap does not close through upskilling. It exists because AI-native firms sell outcomes, not hours, and produce those outcomes with a fraction of the headcount.

Operating margins tell the same story from a different direction. TCS’s operating margin has declined from 26.8% in FY23 to 24.5% in FY25, per company filings. Infosys declined from 23.0% in FY23 to 21.1% in FY25 over the same period, per Infosys’s annual reports. That looks like a manageable drift on a quarterly slide deck. The mechanism behind it is not manageable. Clients are renegotiating toward outcome-based pricing structures. Under outcome-based contracts, the number of engineers deployed becomes irrelevant to the fee. If an AI agent produces the deliverable in one-eighth the time, the implied billing rate collapses even when the nominal contract value holds flat.

Testing services represent a significant revenue line across the top five Indian IT firms. JLL India’s Q1 2026 India IT Real Estate Demand Report notes that Bengaluru Outer Ring Road, the geographic heart of Indian IT delivery, recorded a 3.1 percentage point rise in office vacancy between Q3 FY25 and Q1 FY26, the largest two-quarter move since 2013. That is a physical-space signal of what the headcount freeze means in practice.

The Survivors Framework — Who Has a Real Pivot, Who Doesn’t

This is not a sector-wide extinction. But the firms that survive the next 24 months will look structurally different from the firms that struggle, and the distinction is clearer now than it has been at any point in the last decade.

Tier 1 — Defensible positioning

HCL Technologies’ revenue is more weighted toward infrastructure services, engineering R&D services, and IT products than its peers. Per HCL’s FY25 annual report, Engineering and R&D Services contributed 19.8% of revenue, and IT products (including the IBM partnership portfolio) contributed 7.4%. These segments have a longer automation horizon than application development and testing. Physical systems integration, embedded software engineering, and cloud infrastructure architecture are not yet agentic AI’s easiest targets.

Coforge, smaller, ₹18,500 crore market cap, is building an AI-native delivery model explicitly. Its Q3 FY26 earnings call (February 2026) included specific management commentary on outcome-based contract structures and a stated target of 40% of new contracts being outcome-priced by FY27. It trades at approximately 35x forward earnings, a premium to the sector that reflects the market’s recognition of the pivot. That premium will compress in a broad sector selloff; that is the entry point, not a reason to avoid it today.

LTIMindtree’s engineering services vertical, serving manufacturing, semiconductor, and hi-tech clients, has structurally stickier client relationships. Per LTIMindtree’s Q3 FY26 investor presentation, engineering and industrial segment revenues grew 11.2% year-on-year even as overall company growth slowed. This is the segment least exposed to pure-code agentic automation.

Tier 2 — Structurally exposed

TCS, Infosys, Wipro, and Tech Mahindra are most exposed for the same reason: their core revenue is concentrated in application development, maintenance, and testing, the three delivery categories most directly in the path of agentic AI automation. Tech Mahindra’s 8.9% operating margin in FY25 leaves essentially no buffer for pricing pressure. Wipro’s consecutive quarters of flat-to-negative constant-currency growth through FY25 and into FY26, per company disclosures, suggest the revenue problem is already present, not approaching it.

Check here: HCL TECHNOLOGIES Option Chain (NSE) — Live OI, PCR & Greeks

The BPO analogue—what it actually looked like

The closest structural analog is the BPO transition of 2014–2018, when automation tools began replacing the offshore voice and back-office processing that had been Genpact’s and WNS’s core business. The parallel is not perfect, but the de-rating sequence is instructive.

Genpact traded at approximately 22x forward earnings in 2014 at the peak of the offshoring narrative. By 2016, as automation adoption in back-office processing became measurable in client headcount reduction data, Genpact de-rated to approximately 14x forward earnings, per historical Bloomberg data. WNS followed a similar path from approximately 26x to 17x forward earnings over the same window, per MacroTrends historical P/E data. Both stocks eventually recovered, but only after the companies had demonstrably pivoted to analytics, digital-first delivery, and decision-support work. The recovery took 18–24 months from the trough.

Indian IT’s equivalent pivot is toward AI implementation architecture, sector-specific vertical software, and outcome-based managed services. A few firms are actively making this pivot. Most are still communicating to shareholders that the headcount model will coexist with AI tools.

What the Options Chain Is Saying — Institutional Positioning, Not Retail Noise

The most unambiguous read on institutional views is not in analyst notes. It is in the NSE derivatives market, where large desks build structured hedges that persist across multiple weeks. The following data reflects NSE open interest and implied volatility data as of the May 19, 2026 close.

NSE Options Chain — Institutional Positioning, May 19, 2026

Source: NSE open interest data, May 19, 2026. IV percentile calculated on 1-year trailing basis.

INFY’s put/call OI ratio has run above 1.4 for three consecutive weeks as of May 19. When this ratio sustains above 1.3 for more than two weeks on a large-cap Indian IT name, it reflects institutional hedging programmes, not retail momentum buying. Retail participants do not maintain that level of OI consistency, they chase momentum and roll positions weekly. Institutions build structured hedges and hold strikes. The put OI clustering at ₹1,050 and ₹980 on INFY tells you that large desks are positioning for another 8–14% decline from current levels before they consider a natural floor.

An IV percentile at the 74th percentile on INFY means the options market is pricing more daily volatility than 74% of sessions over the past year. This is not capitulation territory; capitulation would register at the 88th–92nd percentile, the range seen during March 2020 and May 2022. At the 74th percentile, the market is still in active price-discovery mode, not exhaustion. There is more movement being priced, not less.

HCLTECH is the clear outlier. Its put/call ratio at 0.94 is essentially neutral, and the heaviest put OI strikes imply only 3–5% downside from current levels. This is directionally consistent with the Tier 1 narrative above. The options market is already sorting this sector into survivors and names with further to fall. Coforge’s 55th percentile IV and 0.81 put/call ratio tell the same story in the small-cap tier.

This section is not a bottom call. The options chain is showing where institutional desks are hedged, not where they expect the floor. A floor call requires put OI to start rolling to meaningfully lower strikes, the ₹880–₹920 range on INFY or ₹1,900–₹1,950 on TCS. As of May 19, that has not happened.

12 Lakh Jobs and What the Freeze Looks Like on the Ground

TCS, Infosys, Wipro, HCL Technologies, Tech Mahindra, and LTIMindtree together directly employ approximately 12.2 lakh people in India, per FY25 annual report headcount disclosures. The concentration is not even; Bengaluru, Hyderabad, and Pune account for the majority of delivery headcount.

The NASSCOM Strategic Review 2025 flagged that the Indian IT sector added net-zero jobs in FY25 for the first time since FY09 during the global financial crisis. That is not a statistic that typically appears in quarterly investor presentations.

Attrition across the top six IT firms ranged from 21.5% to 28.4% at peak in Q1–Q2 FY23, per individual company disclosures from that period. By Q3 FY26, the same firms are reporting attrition between 10.8% and 13.2%, per Q3 FY26 earnings disclosures. That collapse in attrition is not a sign that the sector has become a better employer. It is a sign that the external market for these skills has deteriorated faster than internal satisfaction has improved. Engineers are not leaving because there is nowhere better to go.

The campus hiring situation is more acute. Infosys’s Q3 FY26 earnings call transcript (available on the company’s investor relations page) includes management commentary acknowledging “extended onboarding timelines for campus hires” without specifying a number. TCS’s FY25 annual report noted that fresher hiring was “calibrated to demand conditions.” Industry estimates from NASSCOM’s FY26 interim data suggest that approximately 60,000–80,000 campus hires across the sector are in extended pre-joining status, meaning they have offer letters but no start dates. NASSCOM has not published a final figure; this range reflects their Q2 FY26 member survey data.

The downstream effects are visible in commercial real estate. JLL India’s Q1 FY26 India Office Market Report (published April 2026) recorded a 3.1 percentage point rise in Bengaluru Outer Ring Road vacancy between Q3 FY25 and Q1 FY26. This is the largest two-quarter move in that corridor since 2013. The IT sector’s expansion absorbed virtually all Bengaluru Grade A office supply between 2015 and 2023. That absorption has stopped.

Brokerage Downgrades — Last 90 Days

Kotak Institutional Equities cut its Infosys target price from ₹1,480 to ₹1,220 in March 2026, citing “structurally lower deal-level headcount intensity” as the primary driver, not macro softness. This language is specific: Kotak is saying the problem is inside the delivery model, not outside it.

Macquarie downgraded Wipro from Neutral to Underperform in April 2026, reducing its target from ₹520 to ₹390. The note, portions of which were reported by Bloomberg on April 14, 2026, cited Wipro’s lack of a credible AI-native delivery narrative relative to peers as the key differentiating risk.

Jefferies maintained a Buy on HCLTECH in April 2026 with a target of ₹1,680, explicitly citing the infrastructure services mix as the differentiating factor. This is the only major brokerage upgrade or maintained Buy on a top-six Indian IT name in the last 90 days.

Consensus FY27 EPS estimates for Infosys have been revised down 11.3% since January 2026, per Bloomberg consensus as of May 15, 2026. TCS FY27 EPS consensus is down 8.7% over the same period. These estimate cuts, not the stock price moves, are the structural signal. When analysts cut earnings estimates, not just targets, they are saying the revenue model is producing less, not that the stock was temporarily overpriced.

Also Check: FII DII Data Today 

Frequently Asked Questions

Is TCS at 22x forward P/E actually cheap right now?

No. IBM, during its mainframe-to-cloud transition, a structural disruption with broadly comparable dynamics, de-rated from approximately 20x earnings to 9–11x forward earnings before stabilising, per Bloomberg historical P/E data for IBM from 2013–2019. If Indian IT follows a comparable path, TCS at 22x forward earnings still has meaningful downside before reaching structurally supported valuations. The historical range for market-leading services firms undergoing business model disruption is 10–14x forward earnings. The sector is not there yet. Conflating “lower than it was” with “cheap” is the most common error in the current discussion.

Which Indian IT stock carries the least structural risk right now?

HCLTECH and Coforge carry the lowest structural exposure to the headcount model disruption. HCLTECH’s infrastructure and engineering tilt, 19.8% Engineering and R&D Services per the FY25 annual report, provides genuine insulation. Coforge’s explicit move toward outcome-based contract structures gives it a credible pivot narrative, confirmed in Q3 FY26 management commentary. Both stocks still face sector-wide selling pressure in a broad drawdown; correlation rises in a selloff regardless of fundamentals. HCLTECH below ₹1,320 is where the valuation-to-business-quality ratio becomes genuinely interesting. Coforge below ₹5,100 is the level where the growth premium compresses to a defensible entry point.

When is the next hard data catalyst for this sector?

Infosys Q1 FY27 results, expected mid-July 2026. The specific language to watch in management commentary: “outcome-based contract mix,” “AI-adjusted utilisation rates,” and “headcount optimisation targets.” If Infosys management uses any of these phrases for the first time in a quarterly call, it signals the structural thesis has moved from market pricing into official company acknowledgement, and the ₹980 INFY put strikes get tested within the same session. TCS Q1 FY27 follows approximately one week after Infosys. These two results are the first scheduled data points where the structural thesis shows up, or doesn’t, in audited numbers.

---

The Bottom Line Trade