After nearly five decades, the learned judges have to take into account new challenges
By T N Ashok
There is something almost poetic about India’s Supreme Court being forced, in 2026, to sit nine judges deep and argue about what exactly constitutes an “industry.” Not because the question is trivial. Precisely because it is not.
The definition they are wrestling with is nearly half a century old. It was written for a country that no longer exists — a nation of factories and furnaces, of public sector behemoths, of workers who punched in, punched out, and knew exactly who their employer was.
That India is gone. In its place stands a country where your grocery arrives in twelve minutes via an app, your cab driver is technically a “partner,” your HR manager works from a laptop in Pune for a company headquartered in Singapore, and an artificial intelligence tool has replaced three junior associates at a Delhi law firm. And yet, the legal architecture governing all of this labour — all of this human economic activity — still rests on a judgment delivered in 1978, when Indira Gandhi’s India was still trying to figure out what socialism meant in practice.
The question now before the Court is blunt: Does any of this count as “industry”? And if it does, who is responsible when it all goes wrong?
The Original Sin: A Definition Built for a Different Republic; The 1978 judgment — Bangalore Water Supply and Sewerage Board vs A. Rajappa — was a product of its ideological moment. Justice V.R. Krishna Iyer, one of the Supreme Court’s most brilliantly activist minds, crafted a definition so expansive it could swallow virtually any organised human activity. His “triple test” — systematic activity, cooperation between employer and employee, production or distribution of goods and services — was a deliberately wide net. Hospitals qualified. Universities qualified. NGOs qualified. Government welfare departments qualified.
The philosophy was unambiguous and unapologetic: labour protection must follow the worker, not the ledger. In a country where capital had historically crushed labour underfoot, this was a corrective of historic proportions. It was also, from the vantage point of 2026, a definition with no expiry date stamped on it — and that omission is now causing a constitutional crisis in slow motion.
Enter the Algorithm Economy; Consider what India’s economy actually looks like today.
Swiggy and Zomato employ — or rather, “partner with” — hundreds of thousands of delivery workers who ride through monsoons and summer heat carrying other people’s lunches. These workers have no fixed hours, no employment contracts in the conventional sense, no designated workplace beyond the streets of Indian cities.
Are they employees of an industry? Are Swiggy and Zomato industries at all, in the 1978 sense? Their product is convenience. Their factory is an app. Their assembly line is an algorithm that pings the nearest available human being.
Ola and Uber present an almost identical puzzle. The company insists it is a technology platform. The driver insists — when he bothers to — that he is effectively an employee, subject to platform-dictated pricing, platform-dictated performance metrics, and platform-dictated termination.
Amazon India warehouses goods, yes — but also intermediates between millions of third-party sellers and hundreds of millions of buyers, using predictive logistics, drone delivery trials, and AI-powered inventory management. Is it a shop? A factory?A marketplace?A tech company? The honest answer is that it is all of these simultaneously, and existing labour law was built to handle exactly none of them. These are not edge cases. They are the Indian economy.
Parliament’s Spectacular Abdication; Here is where the story takes a particularly damning turn.
Parliament was not oblivious to this problem. As far back as 1982, lawmakers recognised that the Rajappa definition had become unwieldy and passed an amendment to the Industrial Disputes Act that would have narrowed it. The amendment was sensible. The amendment was passed. The amendment was then simply never notified into force.
For over four decades, a legislative fix sat gathering dust — unused, inert, legally meaningless. Courts continued to interpret the 1978 definition, producing wildly contradictory outcomes. A forest department in one state was classified as an industry. An identical department in another state was not. Different benches of the same Supreme Court arrived at incompatible conclusions about the same statutory language. The doctrinal chaos accumulated, unresolved, year after year.
The result of Parliament’s inaction is that nine Supreme Court judges must now do the work that legislators refused to do — except that judges are constrained by precedent, by statutory text, and by the institutional limits of the judiciary in ways that Parliament simply is not. They are being asked to retrofit a half-century-old definition onto a 21st-century economy using tools designed for neither.
The Sovereign Function Trap; One of the sharpest tensions in this case involves the government itself.
The 1978 ruling brought state welfare activities within the definition of industry. In theory, this meant that workers in government hospitals and public universities had labour law protections — an admirable goal. In practice, it created an absurd situation where sovereign state functions could become subject to industrial disputes, strikes, and litigation.
Today the stakes are exponentially higher. India runs massive digital welfare programmes — direct benefit transfers, employment guarantee schemes, digital health missions. If these are classified as industries, every administrative dispute becomes a potential industrial strife. Every government functionary acquires the legal standing of an industrial worker. The state becomes simultaneously regulator, employer, and party to every labour dispute it is supposed to adjudicate.
This is not a hypothetical concern. It is a structural contradiction at the heart of the case — and the Court must somehow resolve it without either gutting worker protections or paralyzing government administration.
The deeper truth is that no definition crafted in this courtroom will be truly adequate. The economy will continue to mutate faster than any legal test can track. AI systems are already performing tasks that, a decade ago, required human employees. The “employer” of the future may be a machine-learning model that nobody fully controls.
What the Court can do — what it must do — is establish a framework flexible enough to absorb these disruptions without collapsing. The binary of “industry or not industry” may itself be the problem. A more honest legal architecture might acknowledge gradations: platforms as a distinct category, gig workers as a class deserving sui generis protections, sovereign functions as categorically separate from economic activity.
But flexibility cuts both ways. Every grey area in labour law is an invitation for exploitation. Every ambiguity is a gap through which a platform company will cheerfully drive a truck — arguing that its workers are partners, its operations are intermediation, and therefore none of this is its legal problem.
The workers who will bear the consequences of this ruling are not abstractions. They are the Swiggy delivery rider in the Chennai rain, the Uber driver calculating whether this week’s earnings cover fuel costs, the contract worker at an AI company who was never quite an employee and is never quite protected.
India’s Supreme Court is not merely interpreting a statute. It is deciding, for the next half century, whether the law will see these people — or look through them. That is what is actually at stake in the ghost of 1978. (IPA Service)
