Upcoming AI revolution and the great Indian middle class

FOr 30 years, bharat’s IT manufacture has grown consistently, and so feature the salaries and stocks of the people it employs, contributing to the development of the bodoni font Indian middle class. According to the National Association of Software and Service Companies (NASSCOM), this year the sector will cross roughly $315 billion in revenue, which is about 10–12% of India’s gross domestic product (GDP) and nearly 25% of service exports.The emergence of artificial intelligence (AI) changes everything. The big question now is: will this sector continue to build and support the middle class in India?Co-founder of Hindustan Computers Limited (HCL) Ajai Choudhari described this concern as “a serious overreaction to AI.” Writing over email, he points out that much of the public excitement revolves around consumer applications. “Our companies address the enterprise market.” What this means is that Indian companies build and maintain software systems that run banks, insurers, manufacturers and governments. These systems process salaries, manage inventory, move money and store records.In that context, AI writes the first draft of code, accelerates testing and assists documentation. Engineers refine the output, adapt it to specific business needs and solve complex problems. Choudhari acknowledges there will be “some impact on entry level” roles and that the industry is entering a “different orbit of productivity.” Value shifts from “billable hours to faster, higher-quality outcomes.”When asked whether this means fewer engineers will be required for these jobs, he insists the impact will be only “at the entry level.” His optimism rests on substitution. New demand will emerge, he adds. “For example, there is a need to create our own chips for replacing Chinese chips for security reasons.” There is quantum computing under public missions, and quantum-based security systems that could be deployed globally. Companies such as HCL, Cyient, Persistent and TCS already possess relevant engineering depth. These domains require research investment and proprietary capability. They do not scale by multiplying entry-level billing hours.In summary, Choudhari’s counterargument is this: some kinds of work may shrink, but others will grow. The net outcome will become clear as companies decisively move up the value chain.When the same set of questions was sent to Nandan Nilekani, chairman of Infosys, he pointed to his comments at the recently concluded India AI Summit. There, he made the case that AI is a layer that cuts across workflows. It is not a tool that replaces workers. A bank processing loans may use AI to summarise documents and flag risks. But humans remain accountable for decisions. Nilekani emphasises redesign and reskilling, but he does not forecast mass elimination of roles.A recent HSBC report agrees with Choudhari and Nilekani. Systems that exist now are built for reliability and auditability. Instead of being removed, they are expected to incorporate AI to improve efficiency while keeping the backbone intact. HSBC expects these upgrades to begin contributing meaningfully to software revenues from this year onward.Few in India are predicting a collapse in jobs, but neither are they promising immunity from this tectonic shift. Where all of them converge is on the fact that productivity per engineer will rise. At the same time, data in the public domain says the cost per line of code will decline from $15 to $2. That is why Shrinath V, a Bengaluru-based technology consultant who advises firms on AI adoption, believes it is the beginning of the end for jobs in the IT industry.“The challenge they will face is that, as AI advances, more can be done per person. This is not just for IT companies, but for their clients as well.” As that happens, clients who used to outsource earlier will be unable to justify why they outsource. “They can get it done on their own from a small and capable in-house team,” says Shrinath.The second reason the job market will be impacted, he says, is that until now the business model for Indian IT services has been built around volumes. This means having a large number of people working on projects. But now, if you get more efficiency with fewer engineers, “how will you survive when you have smaller requirements? That is unclear.”The third issue Shrinath flags is around skills. “Many excel in project management and process. But AI requires a very different kind of thinking, which is around problem solving. Do they have the volumes of people who can build?”The HSBC report, Nilekani and Choudhari allude to the intake of engineers in new sectors to rebuild companies for the new age. This means reskilling existing ones. “This is what happened to design thinking,” Shrinath points out. “As curiosity around it built up, one of the largest IT companies retrained all its people on it. Nothing came of it; the CEO moved on; and with that any interest the IT sector had in it,” he points out.Between these two worldviews lies an inflection point. If AI accelerates delivery while complexity continues to rise, Indian IT will move up, revenue growth will continue, and the great Indian middle class will not fracture. If, however, AI compresses the need for outsourced labour faster than new high-value work emerges, the arithmetic will change. The answer will determine the trajectory of India’s next generation of engineers — and its middle class.