A Broken Engine - India’s AI Moment

The Rigged Race

In 1957, B.R. Chopra released a film that captured the economic conflict of independent India. Naya Daur (The New Era) ended with a sequence the country never forgot: a race between a machine and a man.

On one side stood the “Bus”—a symbol of the mechanized age, industrial efficiency, and the modernity of the West. On the other stood the “Tonga” (horse cart), driven by the hero, Dilip Kumar. The Tonga was more than a vehicle; it represented the village, the dignity of manual labour, and the soul of the soil.

The laws of physics and economics dictate that an internal combustion engine crushes a horse cart every single time. A bus doesn’t tire. A bus doesn’t need empathy. But Naya Daur wasn’t a documentary; it was a mythology.

In the movie, the Tonga wins.

The audience cheered. A young nation, insecure in a rapidly industrializing world, found comfort in the fantasy that the sweat could outrun the engine.

We didn’t cheer a movie; we swallowed a lie. We coded an equation into the Indian psyche:

Soul + Sweat > Machine.

The victory of the Tonga wasn’t a plot twist. It was the alibi for our national policy.

The Fatal Conceit

Post-independence India didn’t choose manual labor out of ignorance; it chose it out of historical trauma. A nation colonized by a corporation viewed capitalism with terror. In the 1950s, the Western capitalist engine wasn’t just “stained.” It arrived dripping with the plunder of imperialism, dragging the corpses of two World Wars, and choking on the radioactive ash of a nuclear bomb.

The Soviet Union offered a seductive alternative. It presented a state-planned industrial revolution that promised to drag a starving, feudal society out of the dirt. But India’s leaders were not blind. They knew the secret of the Soviet miracle: mass murder fueled it. The USSR did not just replace its peasantry; it slaughtered millions, deliberately starving its own villages to feed its steel mills.

India, born of a non-violent freedom struggle, refused to pay that price in blood. For Jawaharlal Nehru, the goal was a paradox: Democratic Planning. He wanted the absolute control required to build massive industries, but he wanted it without firing a single bullet or crushing a single vote. He wanted the ruthless efficiency of the machine, but he refused to let it crush the romanticized village laborer.

Since the government could not rewrite the laws of physics to make this happen, it rewrote the laws of the market. To keep the fantasy alive, the State rigged the race.

Terrified of destroying the independent craftsman, we wrote labor laws that made scaling private factories impossible. If a company grew too big, it was taxed, regulated, and inspected into submission. Unlike the movie, where the hero wins through willpower, the Indian State ensured the survival of manual labor by placing legislative speed bumps in front of private enterprise.

But a modern military and a rapidly growing population could not survive on hand-spun cloth. The country still needed steel, power grids, and ports. Since the State viewed private businessmen as ruthless profiteers, it decreed that only the government was morally pure enough to build at scale. It created massive state-owned behemoths and called them Public Sector Undertakings (PSUs).

Nehru called them the “temples of modern India.” But temples are built for ritual, not invention. Their mandate was not to invent the future; it was strictly to execute foreign manuals under the dignified label of technology transfer. The government imported the heavy machinery, the blueprints, and the assembly lines from the USSR and the West.

To run these imported state factories, the government needed top-tier technocrats, not private innovators. They built the IITs.

This created a grotesque structural flaw. We poured our national wealth into a tiny, elite canopy—building just a handful of IITs for an entire subcontinent—while starving mass primary education at the roots. To manage the bottleneck, we turned education into an elimination tournament. We invented the high-stakes entrance exam—a funnel designed not to find thinkers, but to filter for the most obedient survivors.

In this system, our brightest students stopped looking into the deep waters of mathematics and proof; they became athletes of the shortcut, perfecting the art of the rote response to win a ticket to the elite institution. We didn’t build a culture of discovery; we built an ecosystem of execution.

When the Digital Bus arrived in the 1990s, the bill came due. We had spent fifty years protecting the horse cart. We didn’t know how to build the engine; we only knew how to labor.

The Sikandar Phase

In 1991, the Indian economy faced international bankruptcy. The republic was humiliated on the global stage. With barely three weeks of foreign exchange reserves left to pay for essential imports, we were staring into the abyss. We packed our national gold into airplanes and pledged it to avoid the default. We had to open our economy. The terror of liberalization loomed over every middle-class household. Our protected local companies were about to be pitted against global giants equipped with better engines, deeper pockets, and faster math.

Groups like the Swadeshi Jagran Manch were sounding the alarm across the country, warning of a new age of economic colonization. We desperately needed a narrative to convince ourselves we could survive the open market. Into this collective anxiety dropped a 1992 film that offered the exact fable the country was starved for: Jo Jeeta Wohi Sikandar.

The climax turned the macroeconomic dread of liberalization into a legible, inter-college bicycle race. On one side rode Shekhar, mounted on a state-of-the-art, multi-gear imported cycle. He represented the frictionless advantage of Western capital. On the other rode Sanju, pushing a heavy, single-gear iron frame. He was the indigenous underdog, equipped with nothing but deficit.

The laws of weight and gear ratios dictate that the imported machine wins. But the scriptwriter deployed a brilliant sleight of hand: confusing passion with physics. Sanju doesn’t win by engineering a lighter frame or a better gear mechanism. He wins through Jazba (passion) and Mehnat (sweat). Against a superior machine, he simply pedals harder, tearing his own muscles apart to cross the finish line.

The theaters erupted because Sanju vindicated our lack of capital. The film offered a seductive, dangerous logic: if sheer willpower could outrun imported technology, then our poverty wasn’t a structural failure—it was a moral advantage. We swallowed the comfort of a cinematic lie, mistaking human endurance for a development strategy. Thirty-five years after Naya Daur, we codified a new national delusion:

Effort > Equipment.

Digital Charkha

When the computer arrived in India, the State didn’t know what to do with it. But the market did. We resolved the paradox of Naya Daur: we used the machine to sell manual labor.

The rise of TCS, Infosys, and Wipro was an economic miracle, but it wasn’t a technological revolution. It was a Digital Charkha. A cottage industry with air-conditioning. We didn’t conquer the digital age by inventing new math; we conquered it by industrializing the night shift. We built massive campuses dedicated to clearing ticket queues, meeting Service Level Agreements, and migrating legacy databases.

It was a willing, highly successful transaction. For a generation suffocated by the License Raj, these jobs were a golden ticket. We gladly traded our Jazba for a clean, private-sector dignity. We sold our time and our resilience, holding the line while America slept. We poured our youth into night shifts to build the modern Indian middle class. We called this engine ‘IT Services.’

But this survival mechanism became a trap. The Sanju spirit calcified into a permanent economic blueprint. At a global scale, we didn’t build the operating systems, the search engines, or the silicon chips. We deployed a vast workforce trained to debug other people’s dreams and patch holes in their infrastructure. We became the people who keep other people’s machines alive.

India found its place in the digital age. Not as its author. As its ghostwriter.

Spontaneous Order vs. The State

The IT sector did not boom through strategic genius of the State. It boomed because the Indian State’s machinery belonged to the wrong century.

The bureaucracy was obsessed with the physical world. Even after the 1991 reforms, forty years of muscle memory refused to die. The State knew how to extract compliance through physical choke points—inspecting boilers, stalling shipping containers, and turning raw materials into a queue. Manufacturing remained a hostage situation.

But code is a fugitive.

It does not sit behind compound walls waiting for an inspector. It slips through telephone wires and migrates across satellite links. Software production minimized the traditional surface area of extortion. It offered no physical inventory to seize and no visible “asset” for a petty official to lean against with his hand out.

The invisibility was not absolute. The government still controlled the telecom gateways and eventually created export zones to manage the flow. But the State’s traditional interface of harassment—the grinding friction of the Inspector Raj—had no purchase on intangible exports. Power found it harder to strangle what it could not grasp.

By the time the government understood the scale of the Digital Charkha, the industry had bypassed the old geography of control. The lesson of the 1990s was brutal: in India, what the State can inspect, it will suffocate. What it cannot hold in its hands, it accidentally allows to live.

The Trap of the “Poor Cycle”

We survived the State, but we missed the trajectory of our own rescue. While we spent thirty years pedaling the iron cycle of IT services, California and Shenzhen laid down high-voltage rails. They built the platforms that own the distribution, and the silicon that runs the world. They didn’t ask whether we could endure. They asked whether they could make our endurance irrelevant.

Resilience is not a development plan. It is what you have when you are denied one. A mechanic is only valuable until the machine learns to service itself. Eventually, maintenance becomes an automated script, and human becomes an exception handler.

The Need for Speed

By 2004, the old metaphors were not just inadequate; they were laughable. Naya Daur warned us about the machine like a village elder. Jo Jeeta Wohi Sikandar offered the rusted bicycle as moral equipment. But as the millennium flipped, Dhoom arrived with a terrifying new unit of measurement: Velocity.

The antagonist doesn’t pedal. He rides a Suzuki Hayabusa. He is borderless global capital given headlights. He moves like an internet connection: frictionless, anonymous, and indifferent to what he crushes on the way.

On the other side is the Indian State. Jai, the police officer, lumbers into the frame on a standard-issue government motorcycle. He is the embodiment of paperwork trying to overtake fiber-optic cable. He is a file pushed from desk to desk, trying to catch a transaction that has already crossed three borders. The old machinery cannot close the gap.

So the State realizes it needs help. It turns to Ali, a street mechanic operating out of a cramped garage.

Ali is not a bureaucrat. He is the classic hacker—the raw, hungry Indian innovator. He doesn’t own the capital, but he knows the machine’s nervous system by touch. He is the only one who can build and tune a machine fast enough to catch the thief.

In the cinematic fantasy, the State recognizes this. It empowers the garage innovator, giving him the agency and the road to run down global capital.

Outside the theater, the bargain was very different.

The Haunting Question

This question haunts Bangalore and Gurugram—like a slogan, like a prayer, like a complaint: Why didn’t TCS, Infosys, or Wipro become Google?

They had the capital. They had oceans of engineers trained to worship logic and obey deadlines. They had the exact demographic that, elsewhere, was packaged into genius.

And here is the detail that cuts to the bone: the search engine, the cloud, and the algorithm were often written right here in Marathahalli and Manesar. They were coded in our offshore development centers. The tragedy is not that we couldn’t build the modern world. The tragedy is that we did build it—but we built it for someone else. We became the tailors of the digital age, stitching together the code for high-stakes blueprints we didn’t design, for brands we would never own. We chose the safety of the billable hour over the risk of the patent.

It wasn’t a failure of intellect. It was the trap of the guaranteed margin.

Building a product is a gamble. You are pouring money into a hunch that might never pay off. It requires the courage to fail loudly and expensively. To survive this phase, innovators need massive funding from investors who are comfortable with high risk. But even when global investors want to back Indian founders, our government panics. Through outdated foreign exchange rules, the State treats investment not as fuel, but as a threat. While Silicon Valley and Singapore let the money flow like water, we demand endless valuation certificates, strangling the money before it ever reaches the startup.

Trapped inside this regulatory fortress, our innovators are forced back to traditional banks. But local lenders demand guaranteed returns and physical collateral—house, farm or gold. You cannot pledge a family farm against an experiment that is mathematically likely to fail. Look at the financial craters on the global innovation frontier: Apple burned $10 billion over a decade on a self-driving car before quietly killing it. Taiwan’s TSMC routinely gambles $20 billion to build a single cutting-edge fab before knowing if the silicon physics will even work. Even OpenAI posted a $13.5 billion net loss in just six months. To play this game requires a cultural tolerance for watching fortunes go up in smoke with zero guaranteed return.

But renting out human time comes with cleaner arithmetic. When revenue scales directly with billed hours, the company’s innovation becomes hiring, not invention. The risk is outsourced along with the labor. At the end of the month, cash arrives like a salary. No imagination required.

That choice dictated the entire architecture the service giants. They built empires where a new headcount request sails through, but a product experiment requires five committees. They scaled certainty. They industrialized obedience. You cannot take a workforce of half a million people, rigorously trained to avoid errors, and suddenly order them to embrace failure. In this model, a failed product bet is career suicide; a delayed delivery is just an operational hiccup.

In optimizing for predictable margins, they recreated the very thing liberalization had promised to dismantle. They manufactured a bureaucracy in-house. Not the old one with dusty files, but a newer, shinier version: air-conditioned, obsessed with utilization metrics, fluent in PowerPoint, and terrified of actual risk. It is a system designed to reward “no escalations” over breakthroughs.

So no, they didn’t become Google.

The Retreat to the Safe Harbor

When the global tide finally turned—when the cloud arrived with its frictionless knife and software-as-a-service began eating the offshore model—the giants did not build a new future. They looked around for a customer who could not walk away.

They found the Indian State.

Instead of fighting for the future of the global internet, they signed contracts to build the plumbing of citizenship: tax portals, passport systems, and digital identity backbones. This was a change of customer— from a world that chooses to a world that commands. When software serves coercion, it stops being a product and becomes a gate.

But a gate is more than a metaphor; it is a business model. In the market, you survive through excellence. In the bureaucracy, you survive through entanglement. These systems are wired so deeply into payment rails, district offices, and legacy databases that replacing the vendor becomes more expensive than enduring the failure. The guaranteed margin is no longer earned through quality; it is secured through the fact that replacing you would mean rewriting the bureaucracy itself.

This creates a fatal shift in accountability. If an app fails, the company loses a user. If a State portal fails, the failure is socialized. When an old man’s fingerprints don’t register or a migrant’s document won’t upload, the vendor blames the server load, the department issues a circular extending the deadline, and the citizen is told to come back tomorrow.

The irony is absolute. The industry that was born by slipping through the gaps from the government’s compound walls ended up rebuilding those walls with APIs and biometric chokepoints. They did not merely become the IT department of the bureaucracy; they became the machinery that does not have to persuade—only to process.

The dangerous, competitive world was traded for the domestic tollbooth. The swagger of “world-class” was traded for the safety of building gates for people who cannot turn back. Not a leap forward, but a withdrawal from the open sea into the harbor where the boats are chained—and the State owns the lock.

The Death of Ali

If the IT giants had retreated to the safe harbor, chaining their boats to the State, why didn’t a hungry rebel rise to challenge them? Why didn’t a real-life Ali build the Superbike in a Koramangala garage?

The answer is buried in the graveyard of unwritten code.

Consider a startup—call it Akshar AI, born in Bengaluru in 2023. There are no IIT pedigrees here, no Silicon Valley returnee badges. Just two self-taught coders from a Tier-2 city, armed with cheap laptops and a singular ambition: to build a radically new, compute-efficient architecture for generative AI. They want to prove you don’t need a billion-dollar data center to build a frontier model. They have the talent, and they have the vision. They scrape together a small seed round from friends and family to lease a cluster of GPUs.

And then the Indian operating system boots up.

• The Angel Tax: Before they could write a single line of code, the Income Tax Department arrives like a priest at a bedside—not to bless a birth, but to certify a sin. The seed money is treated as “income,” and they are taxed on it. Their bank account is frozen pending an appeal, because in this bureaucracy, suspicion is a kind of due process.
• The GST Trap: Desperate for revenue, they win a small government tender to pilot their AI. The government payment is delayed by eighteen months—a national administrative habit. However, the law demands they pay the GST on that invoiced amount immediately. They are forced to pay tax on a promise. They must pay tax on an absence. Their working capital bleeds dry.
• The Paperwork: Instead of training neural networks, the founders spend forty percent of their week filing 23 different returns—ROC filings, TDS reconciliations, and municipal trade licenses. These are not just administrative chores; they are twenty-three different acts of obedience, each one a small kneeling.

By 2024, Akshar AI runs out of cash. They didn’t fail because their tech was bad; they failed because they couldn’t survive the institutional sandpaper that smooths ambition into compliance. The company shuts down quietly. Ali was not defeated by a rival with a better engine. He was defeated by forms. By delays. By the small, relentless violence of a system that only loves innovation if it is safely chained to a government desk.

The Simulacrum

Having killed the mechanic, the State realizes it cannot build the Superbike. But the political cycle still demands the optics of velocity. If it cannot engineer the reality of technological dominance, it must manufacture the simulation of it.

This explains the government AI summits with three thousand speakers, the “Make in India” lions roaring slogans into the void, and the glossy brochures promising “Sovereign AI”—as if sovereignty were a downloadable patch. These are not strategies; they are props designed to mask a void.

In the bureaucratic imagination, the poster of the Superbike becomes more real than the empty data center. The State celebrates the aesthetics of innovation on stage, building podiums faster than it builds capacity, while quietly importing the actual intelligence from California.

Smart City 2.0

This is a familiar playbook. We watched this performance during the “Smart Cities Mission,” where gleaming Command and Control Centres were inaugurated in cities that still lacked basic sewage drainage. We strapped RFID tags to overflowing garbage bins and called it “Smart Waste Management.” We mistook a control room for governance; we mistook a spreadsheet for a street.

Today, the “AI Mission” is merely the new RFID tag: a shiny label slapped onto a rusting machine. We paste “AI” onto broken industrial plumbing the way we once pasted “Smart” onto broken municipal plumbing. We add 2047 targets and digital slogans—grand, breathless, and patriotic—and pretend the future can be summoned by branding.

But stickers do not mend engines. A dashboard cannot fill a pothole, and a national mission cannot unclog the regulatory friction—the procurement lockouts that demand decades of turnover from companies that have only existed for months. What the State is doing is not building capacity; it is decorating collapse. The sticker is not the machine. And the machine is broken.

The Delusion of High Modernism

The NITI Aayog report, with its precise revenue targets for the year 2047, isn’t a plan. It is a prayer dressed up as an Excel sheet—a state-issued hallucination with footnotes. It is built on the arrogant conviction that a central authority can flatten a living, chaotic system into a diagram—one that confidently predicts, down to the decimal point, that AI adoption in Indian agriculture will reach exactly 49.5% by 2035, while construction will hit 49.6%. A 0.1% difference between two vastly different, overwhelmingly informal sectors. What does that 0.1% even mean? They arrived at them by feeding imported Oxford Economics data into a formula.

Under this bureaucratic gaze, forests become timber inventories. People become “beneficiaries.” And now, intelligence—wild, mutating intelligence—becomes a “sector” with milestones.

You cannot centrally plan for 2047 when the underlying technology reinvents itself every three to six months. True innovation is not happening in state-mandated task forces; it is happening at the absolute edge, driven by a private sector navigating chaos. Yet, the State treats AI like a municipal plumbing project—promising 38,000 state-backed GPUs and “smart corridors” while the basic physical and digital infrastructure beneath it remains broken.

This is the State’s old addiction to procedure: it wants the future to queue up at a counter with documents attached. It is like issuing a traffic challan to a cyclone. But the spreadsheet is not the world. And while we are ceremonially ticking boxes for a fabricated 2047, the world is rewriting the laws of physics in 2026.

Naya Daur

We have finally hit the hard wall of reality. The Naya Daur fantasy is dead. The reason is not political; it is physical.

For forty years, India played a “Low Entropy” game. Software services were light. We could win the IT era with Jazba because human labor was the primary fuel. We burned calories for the output. But Artificial Intelligence is a “High Entropy” engine. It does not run on man-hours; it runs on megawatts. You cannot “hustle” your way into training an AI model. The Digital Charkha is structurally obsolete.

To survive this, we must finally drop the romance of the underdog. We must accept that relentless sweating without a superior engine is a trap. The tragedy is that while the world races ahead, the State is busy performing the aesthetics of innovation on a brightly lit bureaucratic stage. Innovation, here, is not built. It is announced.

The State must stop playing Venture Capitalist. When the government announces a ₹10,000 crore AI fund, the public cheers. But that money rarely travels to the edges where real inventions live; it funds a lobbying tournament. It flows to the people who already possess the access codes to power—the incumbents who have mastered the art of translating ambitions into polished PPTs.

When the State creates a massive domestic pie, it actively shrinks the nation’s ambition. Instead of building for the world, our brightest companies turn inward, cannibalizing each other for safe government contracts.

The State must stop trying to pick winners from this deck of compliant insiders. Its job is not to build the Superbike, nor is it to centrally plan the road to 2047. We do not need another masterplan; we need a second 1991. But a second liberalization cannot be drafted by yes men. The State must abandon its suspicion of intellect and start listening to the domain experts. The State must step aside, dismantle the tollbooths, fix the broken education system that still teaches rote syntax instead of deep mathematics. And finally, it must unshackle the innovator in the garage.