Shares of information technology (IT) companies came under heavy selling pressure on Thursday, dragging the Nifty IT index down 6 per cent to 33,058.20 in intraday trade on the National Stock Exchange (NSE), amid concerns over the impact of artificial intelligence (AI) on business models and fading hopes of an early US Federal Reserve rate cut.
The Nifty IT index hit a more than nine-month low, trading at its weakest level since April 17, 2025. It had earlier touched a 52-week low of 30,918.95 on April 7, 2025.
By the close of trade, the index ended 5.5 per cent lower at 33,160.20, marking its second decline of more than 5 per cent in February. Earlier this month, on February 4, the index had plunged 5.9 per cent. In comparison, the benchmark Nifty 50 closed 0.6 per cent lower at 25,807.20. The combined market cap of Nifty IT index companies was down ₹1.6 trillion on Thursday, and has been down by ₹4.6 trillion since February 4.
Industry experts said the selloff reflects a correction in stretched valuations rather than any fundamental shift in business prospects.
Gaurav Vasu, founder and chief executive officer of UnearthInsights, said the Indian selloff mirrors trends in the US market. “Beyond the Anthropic impact, the US is expected to have a new Federal Reserve head. On employment data, expectations were that tech would not add many jobs. Given the layoffs we have been seeing, the employment data is not a surprise for the tech sector,” he said.
Index heavyweights Tata Consultancy Services (TCS), Infosys, Tech Mahindra, Coforge, Oracle Financial Services Software, and LTIMindtree declined between 6 per cent and 7 per cent. Among non-index stocks, Sonata Software, KPIT Technologies, Birlasoft, Cigniti Technologies, and Zensar Technologies fell 6-8 per cent.
Meanwhile, TCS, Wipro, Cyient, and Hexaware Technologies hit their respective 52-week lows. Infosys fell 6 per cent to ₹1,380.60 in intraday trade and is now 6 per cent away from its 52-week low of ₹1,307, touched on April 7, 2025.
Over the past two trading sessions, the Nifty IT index has declined 7 per cent, while losses over the last seven trading days stand at 14 per cent.
“Today’s decline in Indian IT stocks was driven by stronger-than-expected US employment data and a marginal drop in the unemployment rate, which has reduced expectations of an early rate cut by the US Federal Reserve. This pressure was further compounded by ongoing concerns around AI-led disruption in the sector,” said Vinod Nair, head of research at Geojit Investments.
Indian IT stocks were also weighed down by concerns that AI-native platforms could move up the value chain and automate outsourced work. Global software-as-a-service (SaaS) companies, including Salesforce and other enterprise software firms, also saw heavy selling as markets priced in faster AI-led disruption to legacy business models, ICICI Securities said.
Vasu said the impact of Claude Cowork is likely to be more pronounced in operations and business process outsourcing (BPO). “The impact will be felt in BPO, KPO, and legal services. The latest Anthropic release poses a threat to the traditional full-time equivalent (FTE)-based model. It could lead to a 5-10 per cent impact on core coding for generic tech services and a 10-15 per cent impact on operations such as finance, procurement, and HR,” he said.
However, analysts cautioned that enterprise adoption of such tools may not be straightforward.
Yugal Joshi, partner at Everest Group, said periods of intense hype make it difficult to distinguish between incremental change and genuine step shifts. “Claude Cowork appears to be a step change. While building plugins on enterprise data is not new, the quality of output could be a differentiator. However, what emerges from the tech world and what actually gets adopted by enterprises can be materially different,” he said.
Last week, Cognizant Chief Executive Officer Ravi Kumar said the assumption that new AI tools could be seamlessly plugged into enterprises to immediately replace large portions of IT services work is misplaced. “If a tool could simply be plugged into an enterprise and magically start delivering output, that value would have already shifted over the last three years. The reality is that the value is still sitting with infrastructure and has not fully migrated to enterprises,” he said.