Shares of Indian information technology (IT) services companies plunged on Wednesday, with the Nifty IT index sliding 6.3% before ending the day 5.87% down at 36,345 points, as investors reacted to a fresh and uncomfortable question: What happens to an industry built on selling human labour when software begins replacing that labour at scale? Meanwhile, Jensen Huang, Nvidia's chief executive officer (CEO), has dismissed fears that artificial intelligence (AI) will replace software and related tools. He called the idea 'illogical', following a significant selloff in global IT stocks.
The immediate trigger for the markets was the launch of new automation tools by US artificial intelligence company Anthropic, whose expanded Claude Cowork platform now automates tasks across legal, sales, marketing and data analysis. The announcement sparked a global sell-off in software and data services stocks overnight, with Indian IT companies among the hardest hit.
All 10 constituents of the Nifty IT index ended the day in the red, marking the sector’s worst single-day fall since March 2020. Infosys fell 7.3%, while TCS declined 7%, HCL Tech dropped 4.58%, Tech Mahindra by 4.52% and Wipro slid 3.8%. By mid-session, the combined market capitalisation of IT index stocks had fallen by about Rs1.9 lakh crore.
Market participants say the reaction goes deeper than another bout of AI anxiety. Instead, it reflects growing acceptance that the fundamental unit of value sold by IT services companies, the billable human hour, is being automated away in real time.
Anthropic, led by chief executive officer (CEO) Dario Amodei and backed by Amazon and Google, launched its agentic AI platform Claude Cowork in January. The system can plan, execute and review multi-step workflows with minimal human input.
While Anthropic’s developer-focused Claude Code tool had already pressured software stocks last month, it was the release of its legal automation plug-in on 3rd February that intensified the sell-off.
The legal plug-in is part of a broader set of 11 plug-ins, hosted on GitHub, covering legal, finance, marketing, sales, productivity and data workflows. These tools allow Claude to review contracts, triage (sorting based on urgency and likelihood of survival) non-disclosure agreements (NDAs), track compliance, write structured query language (SQL) queries, analyse datasets and generate reports, the tasks that form the backbone of outsourcing contracts globally.
Anthropic has stressed that the tools are not substitutes for professionals and that outputs require human review. But investors appear focused less on disclaimers and more on economics.
Indian IT services companies all sell variations of the same product: human hours billed at arbitrage rates. Engineers costing US$15–US$40 an hour are billed to global clients at US$50–US$150 an hour, depending on location and skill. That spread reportedly helped build more than US$500bn (billion) in combined market capitalisation across companies such as TCS, Infosys, Wipro, Accenture and Cognizant.
In a typical large IT services contract, hundreds of developers maintain legacy systems at blended rates of about US$50 an hour, translating into US$40mn (million) to US$60mn a year per contract. Tools such as Claude, Cursor and GitHub Copilot are already generating around 46% of developers' code, with acceptance rates exceeding 30%. GitHub says developers’ complete tasks 55% faster using Copilot.
The deeper problem for IT service-providers, however, is structural. Companies providing IT services cannot fully adopt AI without undermining their own revenue models. Every efficiency gain reduces the number of billable hours they can charge.
Most IT service providers, especially from India, are structurally incapable of benefiting from the technology that is replacing them. This is because headcount multiplied by billing rate is their business and AI attacks both these variables.
This creates a paradox: clients deploying AI will demand fewer people and lower costs, while vendors adopting AI internally risk shrinking their own top lines.
The challenge is especially acute in India, where the IT-business process management (IT-BPM) sector employs millions. Unlike past industrial disruptions, these are not easily re-deployable workers.
Over the years, software was considered a safe career. Now AI writes cleaner code than a three-year associate, debugs faster than a senior engineer and documents systems humans never bothered to document.
Valuations reflect this shift in perception. TCS and Infosys now trade at 20–25 times earnings, down from 35–40 times in 2021.
The global market reaction underscores the scale of the concern. A Goldman Sachs basket of US software stocks fell 6%, wiping out about US$285bn in market value, according to Bloomberg. Shares of Salesforce, ServiceNow, Adobe and Workday fell between 6% and 8%, while legal and data firms were hit harder.
Jefferies described the episode as a 'SaaSpocalypse', marking a shift from the belief that AI would help enterprise software firms to the fear that it could replace them.
Adding to investor anxiety are repeated warnings from
Anthropic CEO Amodei, who has said AI could lead to large-scale white-collar job losses within five years if left unchecked.
He has criticised governments and companies for downplaying the risk, arguing that entry-level roles in technology, finance, law and consulting are especially vulnerable.
Anthropic is entering a competitive legal AI space already populated by startups such as Harvey AI and Legora. Its advantage lies in owning its models rather than relying on third-party providers, allowing it to scale faster and price aggressively.
For Indian IT companies, the immediate impact of Wednesday’s sell-off may be sentiment-driven. But, for now, stock markets appear to be delivering a blunt verdict: IT services may be the sector AI was specifically designed to replace — not disrupt.
