Accenture's Adjustment Hits TCS and Infosys, and India Discovers Offshoring Can No Longer Absorb the Slowdown
Five days after Accenture cut its guidance and lost nearly 20% in a single trading session, Infosys and TCS hit five- and six-year lows. The shock exposes a new problem: AI is suppressing service before offshoring can respond.
Accenture's 18% drop in stock on June 18, following the release of its third fiscal quarter for 2026, was described by Bloomberg as the company’s worst trading day since its IPO in 2001. The results themselves were not catastrophic: revenue of $18.7 billion, a 6% increase in dollar terms, an operating margin of 17%, and bookings of $19.3 billion, down 2% year-on-year. What sealed the fate of the trading session was the guidance: the company lowered its annual growth range to 3% to 4% in local currency, removing the ceiling that had been set at 5%, and cited persistent weakness in the U.S. federal business.
The effect reached Mumbai and Bombay the following morning. On June 19, Infosys fell 8.19% to Rs 1,034.85, a 52-week low, and TCS dropped 6.52% to Rs 2,060.50, nearing a six-year low. Capgemini closed Paris with an 8.9% loss. The market's immediate reading was correct in its short-term diagnosis: Accenture serves as a proxy for global discretionary spending in IT services, and when the company's guidance slows, the Nifty IT index bleeds alongside. The interesting question begins one level down.
Where the Structural Problem Lies
The traditional margin capture model in the sector has always operated in three stages: the American or European consultant sells the transformation, the Indian software factory delivers the code, and the margin accumulates in the spread between the two. This arbitrage worked for thirty years, survived the 2008 crisis, endured the pandemic, and financed the expansion of hubs in Bengaluru, Hyderabad, and Pune to nearly 5.9 million direct workers, according to Nasscom.
What changes in 2026 is that AI suppresses both sides simultaneously. At the top, coding agents and platforms like Claude Code, Cursor, and OpenAI's own Frontier reduce the time an engineer needs to deliver software functionality, according to benchmarks published by the laboratories themselves. At the base, the work supporting the bottom of the Indian pyramid—maintenance of legacy applications, manual testing, and configuration of SAP and ServiceNow—is the most rapidly automated with vertical agents. TCS laid off 23,460 people in the fiscal year ending in March, the largest cut in the company’s history, and Wipro reduced its hiring guidance for freshers to 7,500 to 8,000 positions, down from the historic 30,000. Infosys maintains its target of 20,000 freshers annually, but with a bias toward AI and cloud profiles.
Accenture is not immune to the same squeeze, and that is what makes the reading more uncomfortable. The company reported $1.5 billion in generative AI bookings just in the third quarter, bringing the fiscal year total to $4.1 billion. But GenAI replaces within the portfolio the work of larger teams on legacy projects. The consultancy is not losing the client; it is selling the same client a delivery that requires fewer people.
The Market Counterpoint, Before It Appears in a Goldman Window
The thesis of the "death of offshoring" has significant flaws. Aiman Ezzat, CEO of Capgemini, said in recent interviews that the demand for transformation services is growing at a pace faster than productivity cuts in existing teams, and analysts from Bernstein and HSBC argue that the current cycle resembles the slowdown of 2019 more than a structural break. The data supports part of the counterpoint: Accenture’s operating margins are expanding, not contracting, and revenue growth remains positive. Indian IT is also maintaining aggregate job growth, with the industry adding 140,000 positions to reach 5.9 million by 2026. The Indian IT sector is not shrinking; it is changing in composition.
What this optimistic view omits is the deadline. Even if transformation is inevitable, the gap between the decline in revenue per employee in traditional services and the maturation of the new outcome-based or agent-utilization models is what compresses the stock multiple. Buying Infosys today under the "cycle will return" thesis is betting against time, not against transformation.
The Reading in Brazil and Outside India
The impact is not limited to India. The Philippines, with about 1.5 million BPO workers according to IBPAP, and Poland, a European delivery hub for Accenture, Capgemini, and Deloitte, are experiencing a local version of the same compression. In Brazil, the four major consultancies maintain shared services hubs in São Paulo, Campinas, Salvador, and Porto Alegre, with at least 60,000 professionals collectively among Accenture, IBM Consulting, Capgemini, CI&T, and the Brazilian operations of TCS and Wipro. Falconi had been hiring for the AI-assisted transformation axis, but the pipeline depends on the same federal and banking budgets being cut at their American parent companies.
What Brazil lacked until 2024 was direct exposure to the global compression of discretionary spending. By 2026, with Accenture, IBM Consulting, and Capgemini reorganizing local practices and Itaú, Bradesco, and the insurance sector reviewing technology budgets for 2027, the import of the shock is immediate. The shock will not arrive via mass layoffs in Brazilian branches; it will arrive through the contraction of fresher hiring and the diversion of local client budgets to global agreements negotiated in Dublin or New York. This is hard to measure in headlines but will show up at the end of the year in the payrolls of the Faria Lima and Av. Eng. Luís Carlos Berrini towers.