The AI scare has come knocking on the doors of Indian IT sooner than expected. Spooked by rapid advances in agentic AI, the Nifty IT has erased 20% of investor wealth this month, even as the broader market has remained flat.
But as doomsday theories proliferate, the other side of the story is being ignored – that AI’s impact will be multi-layered and will depend on how India Inc adapts
. Let’s discuss.
The SaaS scare
Earlier this month, Palantir said it had cut enterprise resource planning (ERP) integration timelines from years to weeks using agentic AI. Soon after, Anthropic’s Claude Cowork announced open-source plugins that similarly compress workflows across legal, finance, sales and marketing. The global agentic AI market is set to expand more than six-fold to $43 billion over the next five years, according to a whitepaper by LTIMindtree.
Indian IT works on a headcount-based pricing model – the more the resources deployed on a project, the higher the billed revenue. The possibility that AI agents could replace the work of thousands of IT professionals threatens to disrupt this long-standing business model. Application managed services, which currently contribute 22-45% to the sector’s revenues, are the most at risk, according to a note by Jefferies.
Coforge, Hexaware, and Tech Mahindra derive the largest portion of their revenue from application services, according to aMint analysis of Gartner and Jefferies estimates. Their stocks have fallen as much as 33% over the past month. In contrast, giants like TCS, Wipro, and HCLTech were early movers in pivoting toward AI. Because they are less dependent on application services, their transition is expected to be smoother and less disruptive to their overall revenue, and their stocks have held up better amid the SaaS scare.
The SaaS catch
Generic LLMs are widely available, and differentiation is contingent on the quality of context provided to LLMs. Those with deeper domain expertise and access to high-quality proprietary data will be able to harness these models to improve the customer experience.
This explains why vertical SaaS, or SaaS that’s specialized for specific industries, has held up better than general (horizontal) SaaS. The Avenir Vertical SaaS Index has appreciated 1% since November 2021, while the horizontal SaaS index has shed 49%.
It is currently financially unfeasible for AI-native startups to achieve the same depth of business context and feature-completeness as legacy providers. Replicating the security, predictability, and vast ecosystem partnerships of established players while attempting to offer cheaper vertical SaaS simply requires more investment than the market can support.
Instead, AI finds better use when incorporated within the development processes of legacy SaaS players. While they have managed to stick to the much-talked-about “rule of 40” — maintaining revenue growth plus EBIT margin at 40%, GAAP calculations show bare breakeven. AI-led reduction in development costs can help them achieve “true” rule-of-40 results, even after adjusting for stock-based compensation and amortization of acquired intangibles.
Uphill battle for AI in mission-critical fields
Pureplay B2B providers of routine grunt work – low-value coding, testing, and customer care can be more easily replaced with AI. Recent developments at Anthropic also promise to disrupt investment management.
But in mission-critical business segments, where reliability is critical, AI autonomy is still a long way ahead. Even today, AI-generated research must be double-checked for hallucinations. Skipping this reportedly cost Deloitte AU$440,000 in fines last year.
So, in high-value fields such as banking, which demand auditable and explainable systems, true AI autonomy remains theoretical. Use of AI is “not realistic for the majority of high fidelity enterprise class platforms,” according to a note by HDFC.
AI will be what we make of it
Product development cycles are expected to shrink, resulting in faster go-to-market, quicker feedback loops, and possibly more competitive pricing. For example, Claude Code can read, build, test, and debug code-repositories in hours if not minutes, commoditizing the work of legacy software engineers. This could affect margins for businesses with large fixed employee costs.
But the opportunity is undeniable, and recent management commentary indicates that companies are trying to grab it with both hands. While almost all firms have been talking about their big plans for AI, a few have set them in motion.
TCS plans to build a 1GW data centre and has already signed OpenAI as an anchor client for 100MW. OpenAI’s ChatGPT will also be deployed across TCS’s employee base, while Infosys has partnered with Anthropic, prompting deeper AI integration across their services. TCS and Infosys have joined HCL Tech in reporting AI-led revenues, which have outpaced overall business growth, thanks to growing demand for AI transformation in cloud, cybersecurity and other enterprise workflows.
Deals requiring AI transformation are expected to scale faster than conventional IT transformation, as seen at TCS and HCL Tech, which have reported double-digit sequential revenue-growth for AI-led revenues against low single-digit growth for the overall business. Upskilling and agility will be critical as labour-cost arbitrage loses value and pricing models are forced from resource-based to outcome-based. Indian IT will need to move up the value chain, from providers of cost-effective coders to full-stack AI product developers.
How India can catch up: the hardware edge
AI capacity is hugely dependent on raw materials – chips, data-centres, power, land and water. The greater the capacity, the lower the costs and faster the adoption. Enterprise-wide adoption needs to pick up pace to justify further investments in AI.
India can catch up in the AI race in this next leg of expansion, by stepping in on data centres, thanks to ample land and water resources, and more recently, tax breaks for foreign firms operating from domestic data centres. Data centre investment commitments have run into hundreds of millions of dollars, with domestic conglomerates such as Adani and Reliance, as well as global tech majors including Microsoft, Amazon and Google jumping on the bandwagon.
Higher data-centre capacity is expected to cut latency and costs for local AI-integrated business processes including SaaS. It is also expected to attract model-ops workloads and help capture a share of inference revenues.
AI can redraw the software map
As AI democratizes coding, there’s no reason for software development to be restricted to large IT companies anymore. Startups and even IT-adjacent industries have started developing applications in their respective domains using generative AI.
Take for instance Tanla’s Wisely, an AI and blockchain-powered secure communications platform as a service (CPaaS), and Tata Communications’ backing of Commotion Inc, a leading AI-native enterprise startup that has partnered with Nvidia to develop an AI operating system for use across telecom, aviation, hospitality and other industries.
Tanla stock has bucked the sectoral trend by correcting only 8% in February, and Tata Communications stock has gained almost 6%.
IT companies, with their deep pockets and long client relationships, could acquire up-and-coming AI-centred businesses to catch up, much like Google acquired Android instead of building it in-house. L&T recently picked up a 21% stake in E2E Networks.
While regional regulations around AI will play a role, Indian IT has executed pivots before, such as when ERP and cloud disrupted their business models, and can do so again. But execution risks linger, and investors are paying the price for the near- to medium-term uncertainty.
Ananya Roy is the founder of Credibull Capital, a Sebi-registered investment adviser. X: @ananyaroycfa
Disclosure: The author does not hold shares of the companies discussed. The views expressed are for informational purposes only and should not be considered investment advice. Readers are encouraged to conduct their own research and consult a financial professional before making any investment decisions.