English only · Odia translation in progress

Chapter 4: The German Factory

In 1953, a delegation of West German engineers arrived in Sundargarh district, western Odisha, to survey a site for what would become Asia’s first Linz-Donawitz converter steel plant. They came from Krupp and Demag, companies headquartered in the Ruhr Valley — the industrial heartland of a country that had been rubble eight years earlier and was now rebuilding itself into an economic miracle. They carried with them blueprints for a steelmaking process so advanced that no plant in Asia had yet attempted it. The LD converter — named after Linz and Donawitz, the Austrian towns where it was perfected — could produce steel faster, cheaper, and at higher quality than the open-hearth furnaces that dominated world production. This was cutting-edge European metallurgy, the technology of the future, being planted in a district where more than half the population belonged to the Oraon, Munda, Kisan, Bhuyan, and Gond tribal communities. Where the dominant mode of agriculture was shifting cultivation. Where the nearest significant town, Sambalpur, was sixty kilometers away and connected by roads that turned to mud in the monsoon.

The Germans did what Germans do. They planned. By March 1954, Krupp and Demag had formed a dedicated consultancy company — Indien Gemeinschaft Krupp Demag, or IGKD — headquartered in Duisburg, to handle the design, procurement, erection supervision, and commissioning of the plant. They surveyed 7,700 hectares of land, roughly 19,000 acres, and drew up blueprints not just for a steel plant but for an entire township: residential sectors organized along a ring road, with housing ranked by employee designation, parks and green spaces, clubs and recreation facilities, a hospital, schools, a water supply system, a sanitation network. They designed a modern European industrial settlement and dropped it into a tribal forest.

On February 3, 1959, President Rajendra Prasad inaugurated the first blast furnace. They named it Parvati.

The tribal families whose land became the plant and the township were not, of course, consulted about the naming.

If you visit Rourkela today — sixty-seven years after construction began — you can still see the two cities that the German engineers created without intending to. Drive along the 16-kilometer Ring Road of the Steel Township: the sectors are numbered, the roads maintained, the trees mature, the parks manicured. There is Ispat General Hospital with its 780 beds, the DAV schools, the officers’ club, the stadium. The water runs, the drains function, the streetlights work. It is not beautiful in the way that planned European towns are beautiful, but it is functional in the way that German engineering is functional. It works.

Now cross the invisible line into the Civil Township — Uditnagar, Chhend Colony, Basanti Nagar, Jagda, Panposh. The grid dissolves. The roads deteriorate. The planning disappears. Here live the shopkeepers, the rickshaw pullers, the contractors, the domestic workers, the mechanics, the small traders — everyone who serves the steel plant’s ecosystem without being employed by it. The Steel Township is administered by SAIL. The Civil Township is administered by the Rourkela Municipal Corporation. The boundary between them is not marked by a wall or a gate, but you feel it in the asphalt under your feet. Two governance systems, two qualities of life, two Rourkelas, separated by the width of a road.

This physical partition is the metaphor made visible. And the metaphor is not about steel or cities or even inequality. It is about what happens when you build a monolith.

The Nehruvian Bet

To understand what Rourkela was supposed to become, you have to understand the bet that Jawaharlal Nehru was placing with it.

The year was 1954, and India was eight years into independence with a theory of development that had the force of religion. The theory went like this: India was poor because it lacked industrial capacity. The colonial economy had extracted raw materials and shipped them to British factories, leaving India with mines and plantations but no mills, no furnaces, no heavy engineering. The solution was obvious — build the mills, build the furnaces, build the heavy engineering — and the state would do the building because private capital in India was insufficient, the scale required was too large for private enterprise, and the commanding heights of the economy belonged in public hands by ideological conviction.

Steel was central to this vision. Nehru called the heavy industrial plants “temples of modern India,” and he was not being metaphorical. He meant it with the sincerity of a man who believed that a nation’s soul could be forged in a blast furnace. The Second Five Year Plan (1956-61) committed India to building three integrated steel plants simultaneously: Bhilai in Madhya Pradesh with Soviet collaboration, Durgapur in West Bengal with British collaboration, and Rourkela in Odisha with West German collaboration. Three plants, three foreign partners, three underdeveloped regions that would be transformed by the arrival of heavy industry.

The logic was not naive. It had worked elsewhere. Pittsburgh had been built around Carnegie Steel. Essen around Krupp. Sheffield around its cutlery and steel mills. And most relevantly for India, Jamshedpur had been built around Tata Steel — a single-company town created in 1908 that by the 1950s was a functioning city of several hundred thousand people with schools, hospitals, parks, and a growing ecosystem of ancillary industries. The Tata model was the proof of concept: anchor a heavy industry plant in an underdeveloped region, and the city will grow around it. The workers need housing, so housing gets built. The workers need food, so markets open. The workers need healthcare, so hospitals are established. The workers need schools for their children, so schools are built. The secondary economy serves the primary economy, and the tertiary economy serves both. This is the multiplier effect, and it is not theoretical — it had produced real cities in real places.

So Nehru placed the bet. The foundation stone was laid. The German engineers arrived. Construction began in 1956 on those 7,700 hectares. The first blast furnace was commissioned in 1959. The initial target was half a million tonnes of ingots per year, expandable to one million. By December 1961, Rourkela Steel Plant had achieved annual hot metal capacity of one million tonnes through the progressive commissioning of blast furnaces, coke ovens, and steel melting shops. The LD converter worked. The technology was world-class. Asia’s first LD converter steel was poured in Sundargarh district, Odisha.

Today, after decades of expansion and a major modernization completed in 2015-16, RSP has a hot metal capacity of 4.5 million tonnes per annum, crude steel capacity of 4.2 MTPA, and saleable steel capacity of 3.9 MTPA. SAIL has announced plans to boost this to 9 MTPA by 2030, though expansion projects have faced delays due to encroachments and rehabilitation failures — a phrase that contains within it the entire history of the plant’s relationship with the land it sits on.

The bet was that Rourkela would become a city. The steel plant would be the seed. The city would be the tree. Sixty-seven years later, the seed is still there — larger, modernized, productive. But the tree never grew. The question that matters is not what RSP produces (it produces excellent steel) but what RSP failed to produce: a city, an ecosystem, an economy that could exist independently of a single employer.

In software engineering, there is a name for a system that does everything, that cannot be modified without risking the whole, and that prevents the ecosystem around it from developing because it absorbs all resources. It is called a monolith.

Two Rourkelas

The Steel Township consists of 18 sectors organized along the Ring Road. It is classified separately in the Census of India — not as part of Rourkela city, but as “Steel Township,” a modern industrial habitation. This is not a bureaucratic accident. It reflects a genuine physical and administrative reality: the Steel Township is a self-contained system. SAIL provides housing to employees (ranked by designation — the size of your house is your rank made visible in concrete and paint). SAIL maintains the roads and greenery. SAIL runs Ispat General Hospital, a 780-bed referral facility that serves the entire region. SAIL operates DAV schools and community centers. SAIL manages water supply and sanitation.

The quality of life inside the Steel Township is not world-class, but it is dramatically better than anything else in western Odisha. This is a basic but important point. If you are an RSP employee, you live in a planned environment with functioning infrastructure, maintained by a well-funded organization with the institutional capacity to keep things working. Your children go to decent schools. You have access to a real hospital. Your roads have drainage.

Now consider the Civil Township. After the steel plant was established, government offices were set up at Uditnagar, named after Raja Udit Pratap Shekhar Deo, the last ruler of the Gangpur princely state whose territory this had been. Around these offices, and around the gravitational pull of the steel plant’s economy, the rest of Rourkela grew — organically, which in Indian urban planning means chaotically. Chhend Colony, Basanti Nagar, Koel Nagar, Jagda, Jhirpani, Panposh, and the surrounding areas sprawled outward without master plans, without zoning, without the systematic infrastructure that the German engineers had designed for the Steel Township.

The contrast is measurable:

In the Steel Township: planned sectors, grid layout, extensive managed green cover, well-maintained roads, company-managed water and sanitation, Ispat General Hospital, RSP schools, clubs and recreational facilities, all administered by RSP.

In the Civil Township: organic unplanned growth, sparse green cover, roads of variable quality under the Rourkela Municipal Corporation, government hospitals and private clinics, government schools of mixed quality, municipal infrastructure for water and sanitation, all administered by an under-resourced municipal corporation.

This is where the monolith metaphor begins to do real analytical work, and where it stops being decorative.

A software monolith is not just a large program. It is a program that has absorbed so many functions that nothing independent can develop beside it. Every feature is built inside the monolith. Every developer must understand the entire system to change any part of it. The monolith controls all the interfaces — the APIs, the data flows, the user interactions. If you want to build something that works with the monolith, you must build it on the monolith’s terms, using the monolith’s interfaces, dependent on the monolith’s continued operation. Nothing grows alongside it. Everything either lives inside it or feeds into it.

RSP is that monolith. Inside the monolith — inside the Steel Township — everything works because the monolith provides it. Housing, healthcare, education, recreation, utilities, governance. Outside the monolith — in the Civil Township — everything is worse because the monolith has absorbed the functions that would normally create demand for independent services. Why would a private hospital invest heavily in Rourkela when Ispat General Hospital, funded by SAIL, serves as the regional referral center? Why would a private housing developer build premium residential complexes when SAIL provides quarters? Why would a private school charge market rates when DAV schools, subsidized by the company, set the benchmark? The monolith does not just serve its own people — it crowds out the market mechanisms that would create an independent economy.

The population data tells this story quantitatively. The Rourkela Municipal Corporation had a population of 320,040 in the 2011 Census. The metropolitan area was 536,450. By 2024, the metro population is estimated at approximately 664,000. Compare this with Jamshedpur, the other eastern Indian steel city: its 2024 metro population is approximately 1,731,000. Both cities were created around a single steel plant. Both are in eastern India. Both had roughly comparable starting conditions. But Jamshedpur is 2.6 times larger. The gap is not in steel production — RSP at 4.5 MTPA is not dramatically smaller than Tata Steel Jamshedpur. The gap is in everything else.

The sex ratio tells another part of the story: 893 females per 1,000 males in Rourkela, lower than the national average. This is the demographic signature of a town that imports male workers and has not become a place where families settle permanently. The transfer culture of PSU employment — officers posted for three to five years, maintaining homes in their “native places,” sending children to boarding schools in other states — means that even the best-paid residents treat Rourkela as temporary. The monolith’s own employees do not invest in the city because they know they will be transferred. The monolith creates the city but refuses to commit to it.

The Jamshedpur Comparison

This is the analytical comparison that reveals everything. Rourkela and Jamshedpur are natural controls for each other. Both are steel cities in eastern India. Both were created around a single steel plant in a previously underdeveloped area. Both sit in tribal-majority districts. Both began as company towns. The differences in outcome are so stark that they function as a natural experiment in how institutional design determines urban fate.

Start with the employers.

Jamshedpur has Tata Steel — the anchor, India’s oldest private steel company, operational since 1912. But Jamshedpur also has Tata Motors (formerly TELCO), which manufactures commercial vehicles. It has Tata Tinplate Company of India, doing downstream steel processing. It has Tata Steel Processing and Distribution for value addition. It has Tata Pigments, Tata Bluescope, Jamipol. It has a TCS presence. And crucially, it has the Adityapur Industrial Area.

Adityapur, a satellite industrial zone established in 1972, hosts approximately 1,500 industrial units. Of these, 85% manufacture auto components — a direct consequence of Tata Motors’ presence. These enterprises employ approximately 28,000 people directly, including an estimated 11,500 tribals and locals. Adityapur is one of Asia’s largest industrial estates, with infrastructure including a hi-tech laboratory with benchmarking center, a Common Effluent Treatment Plant, and Hazardous Waste Management facilities.

Now Rourkela. RSP is the anchor. Beyond RSP there is… NIT Rourkela. There is an RSP fertilizer plant that produces nitrogenous fertilizers using ammonia from the coke oven plant — an auxiliary of the steel operation, not an independent employer. There are some ancillary units in fabrication and machining, all dependent on RSP as their primary customer, lacking the scale and independence of Adityapur. There is an STPI earth station that has attracted a handful of IT firms. There are small-scale handicraft units — jute work, batik, applique, silver filigree — culturally significant but economically marginal.

That is it. That is the entire economy of a city with 664,000 people.

The numbers compress this into a devastating comparison:

Jamshedpur metro population (2024): approximately 1,731,000. Rourkela metro population (2024): approximately 664,000. Jamshedpur’s second large employer: Tata Motors. Rourkela’s second large employer: none. Jamshedpur’s SME cluster: 1,500+ units in Adityapur. Rourkela’s SME cluster: a handful of RSP-dependent ancillaries. Jamshedpur’s management school: XLRI, established 1949, one of India’s oldest and most prestigious, with over 30,000 alumni connecting Jamshedpur to corporate India. Rourkela’s management school: none. Jamshedpur’s city management: JUSCO (now Tata Steel Utilities and Infrastructure Services Limited), India’s first private integrated urban infrastructure services provider. Rourkela’s city management outside the Steel Township: the Rourkela Municipal Corporation.

The question is why.

The surface answer — that Tata is private and SAIL is public, and private companies are better than public ones — is true but insufficient. The deeper answer is about incentive structures, institutional design, and what happens when a monolith has no reason to create an ecosystem.

Tata Steel had commercial incentives to develop Jamshedpur. Jamsetji Tata’s founding vision explicitly included building “all the comforts and conveniences a city could provide.” This was not philanthropy separate from business. It was a business philosophy grounded in a practical understanding: a livable city attracts and retains talent, and talent produces steel. When Tata Motors established operations in Jamshedpur, it needed auto components, which created demand for suppliers, which attracted entrepreneurs, which spawned Adityapur. The ecosystem was not planned as a social project — it emerged from commercial logic. Each new company created demand that attracted the next company.

SAIL had no equivalent incentive. RSP was built by the Government of India under the Second Five Year Plan. Its mandate was steel production, full stop. The Steel Township was a welfare provision for employees — a benefit, like health insurance or leave encashment, not a city-building project. Everything beyond the township boundary was the state government’s problem, or the municipal corporation’s problem, or nobody’s problem. SAIL executives, posted on transfer for three to five years, had no personal stake in Rourkela’s long-term development. They arrived, did their tenure, and moved on. The institutional memory of the city sat in the filing cabinets of a Ranchi headquarters, not in the accumulated knowledge of people who had spent their careers in one place.

This is the critical structural point: Tata managers build careers in Jamshedpur. SAIL managers serve postings in Rourkela. The difference between building a career somewhere and serving a posting there is the difference between planting a tree and pitching a tent. The tree grows roots, provides shade, bears fruit, and attracts birds. The tent is functional, portable, and leaves no trace when it is moved.

XLRI is the proof of this difference. Xavier School of Management, established in 1949 — ten years before RSP’s first blast furnace — has been producing business leaders from Jamshedpur for seventy-seven years. Its 30,000+ alumni form a network that connects Jamshedpur to every corporate boardroom in India. When an XLRI alumnus rises to CEO of a major corporation, Jamshedpur’s reputation rises with them. When a company considers opening an office in eastern India, the XLRI network provides a reason to consider Jamshedpur. This is the compounding effect of institutional investment — an effect that requires decades to materialize and that no PSU transfer-posting culture would ever produce.

Rourkela has NIT. But NIT is an engineering college, not a management school. It produces engineers, not entrepreneurs. And its graduates leave — which brings us to the next paradox.

NIT Rourkela: The Brain Drain Engine

National Institute of Technology Rourkela is one of India’s 31 NITs and consistently ranks among the country’s top engineering institutions. In the 2023 placement season, over 330 companies visited campus. More than 1,500 job offers were extended. The highest package was Rs 1.2 crore per annum. The average package was Rs 14 lakh per annum. Google, Amazon, Microsoft, Qualcomm, Texas Instruments, Tata Steel, Vedanta — the recruiters read like a directory of corporate India and Silicon Valley’s Indian outposts.

The question that matters for Rourkela is not what NIT produces but where its products go.

The answer is: away. Virtually all high-package placements are in Bangalore, Hyderabad, Pune, Delhi NCR, Mumbai, or international locations. Computer science graduates go to IT companies in Bangalore. Electronics graduates go to chip design firms in Hyderabad. Even core engineering graduates in metallurgy or mechanical engineering are more likely to join Tata Steel in Jamshedpur, JSW in Karnataka, or JSPL in Chhattisgarh than RSP in the city where they studied.

NIT Rourkela has been noted to suffer from a “locational disadvantage” — the lack of connectivity makes it difficult for IT and software firms to visit campus during placement season. But this is symptom, not cause. The cause is that Rourkela offers zero employment for computer science, electronics, or IT graduates. A recruiter who travels to NIT Rourkela does so because the students are excellent, not because the city has anything to offer the company. The recruiter arrives, selects the students, and takes them to Bangalore.

Here is the circular trap, expressed with precision.

Rourkela needs an IT/startup ecosystem to diversify beyond steel. NIT Rourkela produces the people who could build such an ecosystem — over 1,500 graduates a year, many of them among the most talented young engineers in India. But those graduates leave because the ecosystem does not exist. And the ecosystem does not exist because the people who could build it keep leaving.

In software architecture, this is a dependency cycle — A depends on B, B depends on A, and the system deadlocks because neither can start without the other. The standard solution to a dependency cycle is injection: an external force provides the initial condition that breaks the loop. For Rourkela, that injection would be a deliberate intervention — an IT park, a startup incubator, a venture fund, tax incentives for knowledge-economy companies, quality housing outside the Steel Township — that gives some fraction of NIT graduates a reason to stay. Nobody has attempted this injection. Not SAIL (not their mandate). Not the state government (they treat Rourkela as “taken care of” because SAIL is there). Not NIT itself (which is an educational institution, not an economic development agency).

The irony compounds when you trace NIT’s own history. It was originally Regional Engineering College (REC) Rourkela, established in 1961, just two years after RSP’s first blast furnace. It exists because of the steel plant’s ecosystem — the plant created the demand for engineers, the township provided the infrastructure, the industrial culture provided the rationale for a premier engineering college in western Odisha. The monolith produced NIT. But NIT cannot produce anything for the monolith’s city, because the monolith’s city offers nothing for NIT’s graduates.

The monolith creates its own brain drain. It generates the human capital that could transform it and then watches that capital walk out the door.

Consider what NIT Rourkela could have been with different institutional design: an anchor for an innovation ecosystem, an incubator for materials science startups (given the proximity of a working steel plant), a research partnership hub that kept some fraction of its graduates as researchers and entrepreneurs. What it is instead is a pipeline — an efficient, well-functioning pipeline that extracts 18-year-olds from across India, gives them four years of excellent engineering education in Rourkela, and ships them to Bangalore permanently. The pipeline works perfectly. It is just pointed in the wrong direction.

The Hockey Stadium and the Smart City

On January 11, 2023, the newly built Birsa Munda International Hockey Stadium in Rourkela hosted its first match of the FIH Men’s Hockey World Cup, co-hosted with Bhubaneswar. The stadium, built on 6 hectares by IDCO at a cost of approximately Rs 261 crore, has a permanent seating capacity of 21,800 — recognized by Guinness World Records as the largest fully seated hockey arena in the world, certified on January 29, 2023. It is named after Birsa Munda, the tribal freedom fighter from what is now Jharkhand, whose Ulgulan rebellion against the British is one of the foundational stories of Adivasi resistance. That his name adorns a stadium built on tribal land acquired for industrial development is an irony the organizers did not acknowledge.

The Hockey World Cup brought temporary national attention to Rourkela. Television cameras showed aerial shots of the stadium against the Sundargarh forest backdrop. Social media celebrated the infrastructure. Smart City beautification projects — completed frantically ahead of the tournament — improved intersections, parks, and road surfaces. There was a traffic command centre, a tribal museum, a convention hall, a green buffer zone around DAV pond, Brahmani riverfront development, smart bus service, a Science Park, smart lighting. Rourkela had been selected in Round 2 of the Smart City Mission, and the hockey tournament accelerated the spending.

Within months of the World Cup, reports emerged that the beautification works were “ruining and wallowing in neglect.” The pattern is familiar to anyone who has watched Indian urban development: event-driven investment followed by maintenance decay. The investment is a one-time allocation tied to a deadline (the World Cup). The maintenance is a recurring obligation that requires institutional capacity, revenue streams, and political will — none of which the beautification program created.

This is the distinction between events and platforms, and it is the distinction that Rourkela has never understood — or more accurately, that nobody with authority over Rourkela has ever been incentivized to understand.

An event is a hockey tournament, a Smart City designation, a stadium inauguration. It has a date, a budget, a photo opportunity. It generates headlines and political capital. It can be completed, declared a success, and added to a list of achievements.

A platform is an institutional structure that generates compounding returns over time. It is a management school that produces entrepreneurs for forty years. It is an industrial estate that hosts 1,500 SMEs. It is a city management company that maintains infrastructure across decades. A platform is not completed — it is maintained, improved, and expanded. It does not generate headlines on a single day; it generates value on every day.

Rourkela has received events. It has never received a platform.

The comparison with Pittsburgh makes this painful. Pittsburgh, Pennsylvania, was a steel city — arguably the archetypal steel city. U.S. Steel’s dominance of Pittsburgh’s economy in the early twentieth century parallels RSP’s dominance of Rourkela. When the American steel industry collapsed in the 1970s and 1980s, Pittsburgh lost over 100,000 manufacturing jobs. The city could have died. Plenty of American steel towns did die — Gary, Indiana; Youngstown, Ohio; Bethlehem, Pennsylvania.

Pittsburgh did not die because it had platforms. Carnegie Mellon University and the University of Pittsburgh — both established decades before the steel collapse — became anchors for a new economy. UPMC (University of Pittsburgh Medical Center) grew into one of the largest healthcare systems in the United States. The combination of research universities and healthcare — “eds and meds” in the local shorthand — created the foundation for a technology economy. Google, Uber, Duolingo, and dozens of smaller tech companies established offices in Pittsburgh, drawn by Carnegie Mellon’s robotics and computer science programs. Pittsburgh’s transformation from steel city to technology city took roughly forty years of deliberate ecosystem building, starting in the 1980s with the recognition that steel was declining and culminating in the 2010s with a visible tech sector.

Rourkela has had the same forty years. NIT Rourkela has been producing engineers since 1961 — longer than Carnegie Mellon has been producing the computer scientists who attract Google to Pittsburgh. But Pittsburgh’s universities were embedded in a city with civic institutions, cultural infrastructure, diverse employment, and — critically — a population that was staying, not being transferred. Rourkela’s NIT is embedded in a monolith that absorbs everything and produces nothing beyond steel.

Pittsburgh decomposed its monolith. Rourkela built a hockey stadium.

I want to be fair here, because the comparison is not entirely equal. Pittsburgh had advantages Rourkela did not: a larger pre-existing urban base, proximity to the American East Coast corridor, decades of accumulated institutional wealth from the steel era (Carnegie’s philanthropy endowed the university; U.S. Steel’s profits funded civic infrastructure), and the structural advantage of operating within the world’s largest consumer economy. Rourkela sits in Sundargarh district, Odisha, India — a tribal-majority district in one of India’s poorer states, with limited connectivity, no proximity to a major economic corridor, and no accumulated institutional endowment. The confidence level on a direct Pittsburgh-Rourkela comparison should be discounted accordingly — perhaps 50-55% applicable rather than a clean parallel.

But the discount does not eliminate the analytical point. The point is not that Rourkela should have become Pittsburgh. The point is that Rourkela’s interventions — a hockey stadium, a Smart City designation, beautification ahead of a tournament — are events in a city that needs platforms. And nobody is building the platforms because the monolith makes them appear unnecessary. As long as SAIL is there, the thinking goes, Rourkela is “taken care of.”

What the Monolith Prevents

This is the deeper analytical section, and it requires being precise about what a monolith does to its environment. Because a software monolith does not just fail to scale — it actively prevents alternative architectures from developing. Understanding this mechanism is the key to understanding why Rourkela has remained frozen for six decades while cities with equivalent starting conditions have grown.

In software, a monolith absorbs functions. A feature that should be an independent service gets built inside the monolith because it is easier, because the monolith already has the data, because building a separate service would require defining interfaces and managing communication between systems. Each absorbed function makes the monolith larger and more complex, which makes it harder to extract any individual function, which means the next function also gets absorbed. The monolith grows not because anyone designs it to grow but because the path of least resistance always points inward.

Apply this to Rourkela.

Housing. In a normal city, housing is provided by the market — developers build, buyers purchase or tenants rent, and the quality and variety of housing reflect the diverse needs of a diverse population. In Rourkela, SAIL provides housing within the Steel Township for its employees. The housing is decent, maintained, and free (or heavily subsidized). This removes the highest-income segment of the population from the housing market. Private developers have no incentive to build premium housing because their natural customer — the well-paid industrial professional — already has company quarters. The housing market in the Civil Township serves the lower-income population that cannot access the Steel Township, producing the quality of housing you would expect when the highest-value customers are removed from the market. The monolith absorbed the housing function.

Healthcare. Ispat General Hospital is a 780-bed facility that serves as a referral hospital for the entire western Odisha region. It is funded by SAIL, staffed by SAIL-employed doctors, and available to SAIL employees and their families at minimal cost. It also serves the general public, but its primary identity and funding come from the steel plant. A private hospital competing with Ispat General faces an impossible economics: it must charge market rates to patients who have the alternative of a well-equipped, subsidized company hospital. The result is that private healthcare in Rourkela is underdeveloped relative to what a city of 664,000 would normally support. The monolith absorbed the healthcare function.

Education. RSP runs DAV schools and supports other educational institutions within the Steel Township. These are decent schools by the standards of western Odisha. A private school charging premium fees must compete with subsidized company schools — a competition that suppresses the development of the private education market. Again, the monolith absorbed the function.

Recreation. The officers’ club, the sports facilities, the parks and green spaces within the Steel Township — all company-provided, all available primarily to employees. The demand for commercial entertainment, restaurants, cultural venues, and recreational facilities that drives the service economy in normal cities is partially satisfied within the company boundary, leaving less demand in the open market. The monolith absorbed the recreation function.

Governance. The Steel Township is not governed by the Rourkela Municipal Corporation. It is governed by SAIL. This means the best-maintained, most functional part of the city does not contribute to or participate in municipal governance. The engineers and executives who would normally be the most demanding citizens — the ones who would insist on better roads, better sanitation, better public services, and who would pay the property taxes to fund them — live in a parallel governance system. The municipality governs the less affluent, less powerful remainder. It collects lower taxes, has less capacity, faces more problems, and receives less pressure for improvement. The monolith absorbed the governance function.

Each of these absorptions follows the same logic as a software monolith: it was easier, at the moment of decision, to build the function inside the company than to create an independent service. It is easier for SAIL to run a hospital than to negotiate with a private healthcare chain. It is easier to maintain company housing than to deal with a municipal real estate market. Each individual decision was rational. The cumulative effect is an environment where nothing independent can develop.

Academic research confirms this pattern at a theoretical level. Ashima Sood’s 2015 study in Urban Studies documented how industrial townships in India represent “corporate urbanisation” that creates planned enclaves while encouraging “patterns of unplanned and under-provisioned growth around the core.” The company town model, rather than seeding urban development, actually suppresses it by removing the constituency — employed professionals with purchasing power and civic expectations — that would normally demand municipal improvements.

This is the monolith’s deepest damage. It does not just fail to create an ecosystem. It removes the people and the demand that would have created one organically.

Tribal Displacement: The Foundation Cost

Before there was a steel plant, before there were German engineers, before there was a Ring Road with 18 numbered sectors, there were villages. Oraon villages, Munda villages, Kharia villages. Sundargarh district’s population is 50.75% Scheduled Tribe — primarily Kisan, Oraon, Munda, Bhuyan, and Gond communities. The land on which RSP was built — those 7,700 hectares — was not empty. It was inhabited, cultivated, and held by tribal communities under customary land tenure systems that the Indian state recognized in principle (through the Fifth Schedule to the Constitution) and violated in practice (through the Land Acquisition Act).

The displacement was not presented as displacement. It was presented as sacrifice. The language of Nehruvian development framed it as a bargain: you sacrifice your land for the nation, and the nation builds your future. The steel plant will bring jobs, schools, hospitals, modernity. Your children will be steelworkers, not subsistence farmers. The nation asks for your land, and in return it offers you the twentieth century.

The bargain was never honored in its own terms. The compensation was inadequate by any standard — and this was a pattern across Nehruvian industrial projects. The Hirakud Dam, completed in 1957, two years before RSP’s first blast furnace, displaced 22,000 families from 325 submerged villages. Only 35% of the budgeted compensation was actually disbursed. Hirakud was the template; Rourkela followed it. The tribal families who gave up their land received cash payments that bore no relationship to the long-term value of what they had lost — not just the economic value of the land, but the social and cultural value of village communities, sacred groves, kinship networks, and an entire way of life organized around the forest and the river.

Some tribal people received SAIL jobs through reservation provisions. The exact numbers are not publicly documented in a way that allows precise claims, but the total was small relative to the displaced population. A tribal community of several thousand families cannot be absorbed by a steel plant that, at its peak, employed around 30,000 people (including contract workers) and reserved a fraction of positions for Scheduled Tribes. The mathematics of absorption never worked.

Decades later, the tribal communities of Rourkela occupy the lowest rungs of the city’s economy. They work as informal laborers, domestic workers, small shopkeepers on the periphery. The villages that existed before RSP are now slums within or adjacent to the city — settlements without planned infrastructure, without the Steel Township’s maintained roads and functioning drains, without access to the company hospital’s subsidized care or the company schools’ standardized education. They live in the shadow of the monolith that was built on their land, with no interface to the system that displaced them.

This connects to the monolith metaphor at the most fundamental level. In software, when you build a monolith, you destroy the existing system to create the new one. You cannot run both simultaneously — the monolith requires all the resources, all the interfaces, all the data. The old system is deprecated, its components scattered, its logic lost. The tribal villages of pre-RSP Sundargarh were the existing system. They had their own logic — land tenure, forest management, community governance, kinship networks, agricultural cycles. The monolith was built on top of this system, destroying it in the process. And the destroyed system’s inhabitants — the tribals who had built and maintained it for generations — have no interface to the new system. They cannot access the monolith’s services (company housing, company healthcare, company education) because they are not employees. They cannot rebuild their old system because the land is gone, the forests are gone, the village sites are under concrete. They exist in a liminal space: not inside the monolith, not outside it (because there is no longer an “outside” — the monolith has redefined the entire landscape), but beneath it. The foundation the monolith was built on, living in the rubble of what was demolished to lay that foundation.

I want to be careful with confidence levels here. The claim that tribal displacement for RSP was inadequately compensated and that displaced communities remain marginalized is supported by extensive evidence and documented patterns across all of India’s Nehruvian industrial projects. Confidence level: 85-90%. The claim that the specific mechanisms I have described — absorption of economic functions, crowding out of independent services — directly caused the tribal communities’ continued marginalization, as distinct from the general structural disadvantages that tribal communities face across India regardless of industrial displacement, requires more nuance. Tribal marginalization in Sundargarh would exist with or without RSP; the question is whether RSP’s specific monolith structure made it worse than it would otherwise be. On this narrower claim, my confidence is lower: 60-65%. The counterfactual — what would Sundargarh’s tribal communities look like without the steel plant — is genuinely uncertain. They might be marginalized in a different way, through mining or forest degradation or simple neglect, rather than through the specific dynamics of the company town.

What is not uncertain is the moral dimension: people were displaced with a promise that was not kept, and the institution that was supposed to justify the displacement has not produced the broad-based development that was promised. This is a factual statement, not a political one.

The RSP Workforce: The Monolith Shrinks

The monolith is not just failing to create an ecosystem. It is shrinking.

RSP currently employs approximately 11,800 permanent workers — roughly 9,350 non-executives, 1,700 executives, and 780 mine employees. An additional 12,000 people work on contract, in the precarious conditions that Indian contract labor typically endures: lower wages, no company housing, no pension, no job security, no membership in the community that the permanent employees constitute.

SAIL has directed RSP to develop a five-year strategic workforce roadmap for 2025-26 that targets a permanent workforce of 5,000-6,000 employees. The plant is automating. The blast furnaces that once required thousands of workers to operate can increasingly be run with hundreds. This is not a Rourkela-specific phenomenon — it is the global trajectory of steelmaking. Tata Steel’s Jamshedpur operations have undergone similar workforce reductions while increasing output. The difference is that Jamshedpur has Adityapur’s 1,500 SMEs, Tata Motors, XLRI, and a diversified economy that can absorb displaced workers. Rourkela has nothing.

When RSP’s permanent workforce drops from 11,800 to 5,000-6,000, roughly 6,000 permanent employees and their families will leave the Steel Township. These are the highest-income, highest-consumption, highest-tax-paying residents of Rourkela. Their departure will reduce demand for every local service — every restaurant, every shop, every school, every medical practice that serves the Steel Township population. The multiplier works in reverse: each departing employee removes not just their own consumption but the consumption of the service workers who depended on them.

This is what happens when a monolith shrinks without decomposition. In software, when an organization decides to reduce a monolith’s scope, the standard approach is to extract independent services first — to decompose the monolith into microservices that can operate autonomously before reducing the monolith itself. If you simply shrink the monolith without first building the independent services, everything that depended on the monolith collapses.

Nobody is decomposing the Rourkela monolith. Nobody is building independent economic services that can operate without SAIL. The monolith is being shrunk — automation will reduce it further — and the city that depends entirely on it has no alternative architecture to fall back on.

The contract workers — those 12,000 people who constitute the flexible labor force — are even more vulnerable. They have no Steel Township housing to lose because they never had it. They have no pension to worry about because they have none. When RSP reduces its contract workforce (as automation allows), these workers will simply vanish from Rourkela’s economy, returning to the districts they came from or migrating to another industrial town. Their departure will be economically significant but statistically invisible.

Can the Monolith Be Decomposed?

In software engineering, decomposing a monolith into microservices is acknowledged as one of the hardest architectural problems in the field. It is not hard because anyone disagrees with the goal — everyone agrees that monoliths should be broken up. It is hard because the decomposition must happen while the system is running. You cannot shut down the monolith, redesign it from scratch, and restart. The users depend on it. The data lives inside it. The business processes are encoded in its logic. You must extract pieces one at a time, build independent services around them, connect them with APIs, and verify that each extracted piece works independently — all while the monolith continues to serve its daily traffic.

Three conditions must be met for successful decomposition:

First, you must identify which parts of the monolith can function independently. Not everything can be extracted. Some functions are so deeply entangled with the monolith’s core that they cannot be separated without breaking the system. The skill is in finding the seams — the natural boundaries where a function could be cut away and given its own infrastructure, its own database, its own team.

Second, you must build interfaces between the extracted parts and the remaining monolith. The newly independent service still needs to communicate with the monolith — to receive data from it, to send data to it, to coordinate actions. These interfaces (APIs, message queues, event streams) must be designed carefully, because a badly designed interface creates its own dependency and the extraction fails.

Third, the whole system must keep running during the transition. This is the hardest constraint. You cannot tell the users, “The system will be down for six months while we rebuild it.” The monolith must continue to serve traffic while you extract pieces from it, like performing surgery on a patient who refuses to lie down.

Apply these conditions to Rourkela.

What could function independently?

NIT Rourkela could anchor an IT and startup ecosystem — if there were housing and quality of life outside the Steel Township that would make graduates willing to stay. NIT already has the talent, the research output, the institutional reputation. What it lacks is an environment in which its graduates could build careers without leaving. An IT park with reliable power and internet, startup incubation facilities, a venture fund with patient capital, and — critically — residential neighborhoods with the quality of life that NIT graduates have come to expect from Bangalore and Hyderabad. This would require investment in the Civil Township that matches or approaches the Steel Township’s standards. It would require treating the non-SAIL part of Rourkela as a city worth building, not a residual space for people who did not make it inside the monolith.

The tribal heritage and Sundargarh’s forests could anchor eco-tourism and cultural tourism — if there were infrastructure. Sundargarh is one of Odisha’s most forested districts, with significant biodiversity and tribal cultural heritage that, properly presented, would attract domestic and international tourists. The Birsa Munda Hockey Stadium could be part of a sports tourism package. But tourism requires reliable transport, quality accommodation, trained guides, food safety, and marketing — none of which exist at scale.

The industrial skill base accumulated over six decades could support auto-component or precision engineering SMEs — if there were an industrial estate with support services comparable to Jamshedpur’s Adityapur. RSP has produced generations of skilled metalworkers, fabricators, and machinists. Some fraction of them, combined with engineering graduates from NIT, could form the workforce for a manufacturing cluster. But a cluster requires anchor demand (a Tata Motors equivalent), industrial infrastructure (plots, power, water, effluent treatment), and financial support services (working capital loans, export facilitation).

What interfaces would be needed?

Each of these independent services — an IT ecosystem, a tourism economy, a manufacturing cluster — would need to interface with the existing RSP-dominated economy. They would need shared infrastructure (roads, power, water) that currently serves primarily RSP. They would need a governance framework that gives the municipal corporation authority and resources to manage urban development beyond the Steel Township. They would need a labor market that is not entirely captured by RSP’s wage structure (which, being PSU scale, is higher than what a startup or SME can offer, creating a wage floor that makes private sector employment unattractive).

Who would architect the decomposition?

This is where the analysis terminates in a governance vacuum.

SAIL will not do it. Ecosystem building is not SAIL’s mandate. SAIL is rewarded for steel production, not urban development. No SAIL executive’s performance review includes “percentage of Rourkela’s economy that is independent of SAIL.”

The state government will not do it. The state treats Rourkela as “taken care of” because SAIL is there. When a state government has competing demands — Bhubaneswar’s growth, Puri’s temple corridor, Paradip’s port development, tribal welfare across 118 blocks — the city with a Navratna PSU and a Smart City designation goes to the bottom of the urgent list. Rourkela does not look like a crisis. It looks like a success. The steel plant produces steel. The NIT produces engineers. The hockey stadium won a Guinness record. What is the problem?

The Rourkela Municipal Corporation will not do it. The RMC does not have the capacity, the resources, the expertise, or the jurisdictional authority to undertake the kind of strategic economic planning that decomposition requires. It can barely maintain the Civil Township’s roads.

The result is that no one has the authority, the incentive, or the capacity to decompose the monolith. And so the monolith persists — not because it is the best architecture, but because the cost and complexity of changing it exceed what any single actor is willing to bear. In software, this is called “technical debt” — the accumulated cost of expedient decisions that become progressively harder to reverse. Rourkela’s technical debt is sixty-seven years deep.

There is one possible exception to this governance vacuum, and I want to flag it with appropriate uncertainty. Tata Steel’s Kalinganagar plant in Jajpur district — a Rs 27,000 crore expansion that has made Odisha Tata Steel’s single-largest investment destination — has been accompanied by Tata Steel’s stated intention to “create another Jamshedpur at Kalinganagar.” If Tata Steel follows through (confidence: 40-50%, given that Kalinganagar’s history includes the 2006 killing of 13 tribal protesters and the fundamental structural differences between a greenfield 21st-century plant and the century-old Jamshedpur organic ecosystem), Odisha may get one example of how private corporate culture builds an urban ecosystem. But Kalinganagar would not solve the Rourkela problem. It would only demonstrate, painfully, what Rourkela could have been with different institutional DNA.

The Compound Failure

Step back from the specifics and look at the structural pattern.

The Nehruvian bet was that a heavy industrial plant would seed a city. The bet worked in Jamshedpur because the institutional structure — private ownership, long-term career culture, commercial incentive to build an ecosystem, a visionary founder’s explicit city-building philosophy — created the conditions for the seed to germinate.

The bet failed in Rourkela because the institutional structure — PSU ownership, transfer culture, production-focused mandate, no incentive for ecosystem building — created a monolith that absorbed all the functions a city needs without allowing any of them to develop independently. The monolith provided housing instead of creating a housing market. It provided healthcare instead of creating a healthcare market. It provided education instead of creating an education market. It provided governance instead of developing civic institutions. And when NIT Rourkela produced the human capital that could have changed everything, the monolith’s city offered no reason for that capital to stay.

The tribal displacement compounds this failure. The monolith was built on a foundation it destroyed, and the destroyed foundation’s inhabitants have no interface to the system that displaced them. The compensation was inadequate. The promised modernization materialized for the people inside the monolith but not for the people beneath it. The Fifth Schedule protections that should have prevented the worst outcomes were violated in the name of national interest, and the national interest was served — India got its steel — while the local interest was not.

The workforce reduction accelerates the urgency. When RSP shrinks from 11,800 permanent workers to 5,000-6,000, the monolith’s already-limited economic contribution to the broader city will decrease further. Automation will produce more steel with fewer people. This is economically efficient and urbanistically catastrophic. The monolith will continue to produce steel — it may even produce more steel — while supporting fewer people, fewer families, fewer consumers, fewer taxpayers. The Steel Township will have more empty houses. The Civil Township will have fewer customers.

And the Smart City designation, the hockey stadium, the beautification projects — these are patches applied to a monolithic architecture. In software, “patching the monolith” is the opposite of decomposition. It makes the monolith slightly more presentable without changing its fundamental structure. It buys time without addressing the underlying problem. A new coat of paint on a building whose foundation is cracking.

The question for Rourkela is not whether the monolith should be decomposed. Everyone who looks at the situation honestly would agree that it should. The question is who will do it, with what resources, under what authority, and on what timeline.

The honest answer, as of 2026, is: nobody, with nothing, under no clear authority, on no timeline.

This is what a monolith looks like when you wait sixty-seven years to address it. It is not a technical problem. It is not an economic problem. It is a governance problem — the absence of an institution with both the authority and the incentive to redesign the architecture. SAIL has the resources but not the mandate. The state government has the authority but not the attention. The municipality has the proximity but not the capacity. The result is stasis: a steel plant that works, a city that does not, and a gap between them that grows wider every year as the steel plant automates and the city hollows out.

In the end, the German engineers who arrived in Sundargarh in 1953 built exactly what they were asked to build: a world-class steel plant. The LD converter worked. The blast furnaces produced. The Steel Township functioned. They were not asked to build a city, and they did not build one. The failure is not theirs. The failure belongs to the institutional design that treated a steel plant as a city-building project when it was, in fact, only a steel-plant-building project. The monolith did what monoliths do: it performed its core function excellently and prevented everything else from developing.

Rourkela Steel Plant produces 4.5 million tonnes of steel a year. What it does not produce — what it has never produced, what its institutional design prevents it from producing — is a city.

Sources

Rourkela Steel Plant History and Operations:

Rourkela City and Demographics:

Steel Township and Civil Township:

RSP Workforce:

Jamshedpur Comparison:

NIT Rourkela:

Smart City and Hockey Stadium:

Sundargarh District:

Company Town Research:

Kalinganagar:

  • “Kalinganagar.” Wikipedia. https://en.wikipedia.org/wiki/Kalinganagar
  • “At Kalinganagar, Tata Steel to copy Jamshedpur model.” Business Standard.
  • “Tata Steel inaugurates Phase II expansion of Kalinganagar operations.” Tata Steel.

Cross-References

The Leaving (full_read/the-leaving/): Chapter 4, “The Skilled Departure,” documents the brain drain from NIT Rourkela and Odisha’s engineering institutions as part of the broader pattern of skilled migration. The NIT paradox — producing talent that immediately leaves — is a node in the larger network of departure that defines Odisha’s relationship with its own human capital.

Tribal Odisha (full_read/tribal-odisha/): Chapters 1-2 document the Oraon, Munda, and Gond governance systems that existed in Sundargarh before RSP. Chapter 5, “The Mountain and the Mine,” places Rourkela’s tribal displacement within the larger pattern of industrial displacement across Odisha — Hirakud, Kalinganagar, NALCO Damanjodi, POSCO. The foundation cost of the monolith is one instance of the repeated pattern.

The Long Arc (full_read/the-long-arc/): Chapter 3, “The Cathedral in the Village,” covers Rourkela Steel Plant as part of the Nehruvian industrial policy and asks why Rourkela never became Jamshedpur. This chapter builds on that analysis with the monolith framework, adding the institutional design explanation.

Political Landscape (full_read/political-landscape/): Smart City designation and SAIL as political patronage operate within the broader dynamics of how Odisha’s political class uses central government programs and PSU presence as substitutes for genuine development strategy.

The Churning Fire (full_read/the-churning-fire/): Chapter 4, “The Inner Fortress,” and Chapter 8, “What Remains,” explore institutional design as the mechanism through which consciousness shifts become durable or fail. RSP’s institutional design — PSU culture, transfer posting, production mandate without ecosystem mandate — is a case study in how institutional DNA determines outcomes regardless of individual intention. The monolith is not malicious. It simply lacks the institutional code for city-building.

Source Research

The raw research that informs this series.