Chapter 181: The Old Lion and the Young Tiger
26 January 1975 — Gorakhpur
The aircraft was a Tata-owned Hawker Siddeley — the specific aircraft that JRD Tata used for domestic travel, which said something about the man, because in 1975 the typical Indian industrialist who used private aircraft used whatever was available and serviceable. JRD used a Hawker Siddeley because JRD had been flying aircraft since 1929, had obtained India's first commercial pilot's license, and had specific opinions about what a properly maintained aircraft felt like, and his specific opinions were expressed in his ownership of one that felt that way.
The aircraft landed at the Gorakhpur airfield at ten-fifteen in the morning.
The airfield was the same airfield that the S-35 programme used for its test flights, and it had been expanded in the past two years to handle the volume of traffic that the Shergill complex generated — private aircraft, military transport, the occasional commercial charter. On the morning of January 26th, it was quiet. Republic Day was a national holiday, and the factory was operating at reduced capacity, and the airfield had the specific quality of an industrial space that was resting.
Karan was at the airfield when the Hawker landed.
He had been there since ten, which was earlier than necessary, which was the specific expression of the relationship rather than of any logistical requirement. JRD Tata at seventy-one years old did not need to be received early. JRD Tata, in forty years of running a complex of companies that employed hundreds of thousands of people across India, had been in more important meetings than any meeting at the Gorakhpur airfield. But the relationship between them was the relationship between friends rather than between counterparts, and friendship expressed itself in small excesses of courtesy.
JRD came down the aircraft steps with the quality that was entirely his own: seventy-one years old, slim in the way of men who had been slim since birth and who remained slim through some combination of genetics and the specific relationship to food that his cosmopolitan formation had produced, wearing a suit that was elegant without being conspicuous, moving with the particular alertness of a man who had been paying close attention to the world for seven decades and had not stopped.
He saw Karan.
He smiled.
It was the smile of a man who was genuinely pleased to see someone — not the professional warmth of a public figure greeting an audience, the specific smile that appeared when you arrived somewhere and found the person you were looking forward to talking to.
"Twenty-fifth Republic Day," JRD said, as he reached the bottom of the steps. "And I am spending it in Gorakhpur instead of Bombay or Delhi, which tells you what I think about the celebrations."
"What do you think about the celebrations?" Karan said.
"I think they are important and that someone else should attend them," JRD said. He shook Karan's hand — the handshake of two people who had shaken hands many times and for whom the gesture was genuine rather than formal. "And I think Gorakhpur is considerably more interesting than watching tanks drive down Rajpath."
"The tanks are interesting," Karan said. "Some of them are ours."
"I read about the Arjuna," JRD said. "Production order in November. Two hundred units."
"The first twenty are being built at Avadi," Karan said.
"Good." JRD looked at the airfield. At the S-35 hangar visible in the distance — the aircraft programme's infrastructure. "Show me what you've built," he said. "I haven't been since early 1974 and I've been reading about what's changed and reading is not the same."
"No," Karan said. "It isn't."
They walked.
Karan walked him through the complex for ninety minutes.
This was what JRD had asked for — not a formal presentation, not a guided tour with prepared talking points, but the specific kind of walk that people who understood manufacturing did when they visited manufacturing: the walk that covered the things that mattered, which was not the things that looked impressive but the things that functioned correctly. JRD had been walking through factories since the 1930s and he had the quality that long experience of manufacturing produced: the ability to look at a production process and see not the equipment but the system, not the individual operation but the flow, not the product but the decisions that produced the product.
He asked specific questions.
At the aerospace assembly bay, he asked about the composite materials process — the specific layup procedure for the S-35's wing sections, the cure cycle, the quality control protocol. He was not asking to understand the aeronautics. He was asking to understand the manufacturing management. He listened to the answers with the attention of someone who understood what correct manufacturing looked like and was assessing what he was hearing against that understanding.
At the ISMC facility, he looked at the reactor bay through the clean room window for a long time.
"The LED production line," he said.
"Three reactors," Karan said. "Running at full capacity since December 28th."
"The yield," JRD said.
"91%," Karan said.
JRD was quiet for a moment. "In December, on the first full production shift."
"Yes," Karan said.
"I've been running Tata Steel's facilities for thirty years," JRD said. "The yield on a new production line in the first month is — for reference — typically 60 to 70 percent at best, improving over six to twelve months to the designed yield. 91% in the first shift means the process engineering was completely worked out before the line ran."
"SPEI," Karan said.
"Tell me," JRD said.
Karan explained the Shergill Process Engineering Institute — the specific methodology, the knowledge base, the approach of solving the scale-up problem before building the production line rather than discovering it on the production line.
JRD listened.
"A meta-engineering institute," he said, when Karan finished. He said it with the tone of a man recognising something he had wanted to exist for a long time. "The specific institutional gap between what laboratories discover and what factories produce. I have been aware of this gap my entire career. I have not had a language for it."
"Pillai called it a meta-engineering institute," Karan said. "On the day it was established."
"Who is Pillai?" JRD said.
"BHU professor. He's running SPEI's educational programme. He came up with the name in the founding meeting."
"Good name," JRD said. He was quiet for a moment. "The Tata group would benefit from something like this. The steel production processes. The chemical plants. We have accumulated the engineering knowledge — it is in our files and in our engineers' heads — but it is not systematised." He paused. "Has Ranjit Kumar said anything to you about this?"
Ranjit Kumar was the Tata Group's chief engineer — the man who had been managing Tata Steel's technical programmes for a decade.
"Not directly," Karan said. "He may have been thinking about it."
"I'll speak with him," JRD said. He said it the way he said things that were going to happen: as a statement rather than an intention.
They completed the walk at the Siddharth-1 production section.
JRD looked at the production line — the chassis moulding station, the keyboard assembly, the circuit board insertion, the testing station, the packaging line.
He looked at the machines being assembled.
He looked at the testing station, where a finished machine was being connected to a monitor and booted and put through the standard test protocol.
The boot screen appeared.
SHERGILL SIDDHARTH-1ISMC BASIC v1.032KB READY
>
JRD looked at it.
"I read Subroto's story this morning," he said. "In the Times of India." He paused. "He wrote: India has not followed the world into the age of personal computing. India has arrived first."
"Yes," Karan said.
"He is right," JRD said. "The Popular Electronics magazine — January 1975, the Altair 8800 cover — I read it also." He paused. "A kit that requires assembly, costs more than your machine, and has no keyboard or screen." He looked at the boot screen. "The comparison is embarrassing for the Americans."
"The Americans had the same idea at the same time," Karan said. "They had the parts. They had the processor. They had the knowledge. They built a different machine because they made a different decision about who the machine was for."
"They built it for engineers," JRD said.
"They built it for themselves," Karan said. "The people who built it were engineers who built it for the pleasure of building something technically challenging. The machine they built was technically challenging to use. It was for people like them." He paused. "We built it for everyone who has a problem that calculation can solve."
JRD looked at the boot screen.
"The cursor is blinking," he said.
"Yes," Karan said.
"It's waiting," JRD said.
"Yes," Karan said.
JRD was quiet for a moment.
"A machine that waits for the person to be ready," he said. "That is the correct design philosophy."
He turned away from the production line.
"Let's have lunch," he said. "I want to talk about computers."
The lunch was at the villa.
Leela Devi had been informed of the guest the previous week and had been, in her characteristic way, simultaneously dismissive of the significance ("I cook the same for everyone") and completely specific about the preparation ("JRD Tata is Parsi, they eat differently, I've talked to Roshan Irani from the Surat community about what's appropriate"). The result was a meal that was entirely Leela Devi in its generosity and exactingness and that was also, through whatever consultation had happened, appropriate for a Parsi guest who ate what was placed in front of him with the specific gratitude of a man who had had many excellent meals and still appreciated each one.
Arjun was at lunch.
JRD and Arjun had met before — twice, during the pipeline contract negotiations of 1972, and once at a Delhi event in 1973. They had the relationship of two older men who had been introduced through someone they both respected and who had found, in each other, the specific quality that older men found valuable in other older men: the settled perspective that came from having seen a great deal.
JRD said to Arjun: "Your son is building things that I didn't believe were buildable."
Arjun said: "I know. I watch it happen."
"How does that feel?" JRD said.
Arjun was quiet for a moment, with the quality he brought to questions that deserved honest answers. "It feels," he said, "like watching something grow that you planted without knowing what the seed was."
JRD looked at him.
"That is a specific feeling," JRD said.
"It is a specific situation," Arjun said.
They ate.
Leela Devi sat at the corner of the table and did not speak much, which was unusual. She was watching JRD Tata with the assessment of a woman who made assessments of people and who found, in this specific person, something that warranted more looking than usual. At one point she said: "Your father was in the textile business."
JRD looked at her. "Yes. My father founded the group."
"And you expanded it," she said.
"I tried to," he said.
"Don't be modest," she said. "It doesn't suit you."
JRD looked at Karan.
Karan said: "She does this to everyone."
"I do this to people who deserve to be told things plainly," Leela Devi said. "You took your father's textile business and made it into steel and chemicals and airlines and hotels and automobiles. That is not trying. That is doing."
JRD looked at Leela Devi for a moment.
"You are absolutely right," he said.
"I know," she said. "Eat the dal."
He ate the dal.
After lunch, Karan and JRD went to the study.
The study was a room in the villa that Karan had claimed as the specific space for work that was not the factory and not the design bureau — the work that happened at a desk with a notebook rather than at a drawing board with engineering documents. It had books, which Karan read with the specific attention of someone who read to understand rather than to accumulate, and it had the notebooks, the accumulated record of decisions and observations and the particular quality of thinking that Karan did when he was processing something he had not yet decided.
JRD sat in the chair that had been Arjun's chair in the study until Arjun had started preferring to read in the sitting room.
Karan sat at the desk.
Between them, on the desk, was the Siddharth-1 demo unit — the same one that had been in the demonstration area the week before, brought home because the study was where it made sense for it to live. It was connected to a small monitor that Karan had acquired from the factory's testing inventory.
JRD looked at it.
"The machine," he said.
"Yes," Karan said.
"I want one," JRD said.
Karan looked at him.
"Not for show," JRD said. "Not to put on a desk and call modern. To use." He paused. "I want to understand what it is by using it. That is how I understand things."
"It's yours," Karan said. "I'll have one packaged for you before you leave."
"Thank you." JRD settled into the chair. He had the quality in this room that he had in all the rooms he occupied — the quality of a man who was comfortable in his own presence and who made the space comfortable by being in it. "Now. The computer. Not the Siddharth. The other computer. The Ganesha."
"Tell me what you want," Karan said.
"TCS," JRD said. "Tata Consultancy Services. You know what TCS is."
"Yes," Karan said.
"TCS was founded in 1968," JRD said. "Seven years ago. The specific vision was: India has a substantial supply of technically educated people who are not being used to their full capacity in the domestic economy because the domestic economy does not have sufficient demand for their skills. The international economy does. The question is how to connect the supply to the demand."
"Consultancy services," Karan said.
"Computing services," JRD said. "More specifically. The specific skill set that TCS was built to deploy is the ability to use computing systems — to programme them, to analyse the problems they can solve, to manage the process of applying computing power to business problems. The Indian supply of this skill set was — and still is — excellent. The IIT graduates, the regional engineering colleges, the specific mathematical training that Indian education produces at its best." He paused. "TCS has been growing. Slowly, because until the license raj was addressed the domestic market was constrained and the international market was limited by the specific challenge of managing offshore service delivery in an era before reliable telecommunications." He paused. "The license raj is gone. The telecommunications are improving. And the computing industry worldwide is entering a period of dramatic growth that creates demand for TCS's services faster than TCS can grow its capacity to supply them."
"Tell me the specific problem," Karan said.
"Computing power," JRD said. "The constraint on TCS's growth is not the people. We have people. The constraint is the computing infrastructure. A computing services company requires computing equipment — mainframes, terminals, the specific infrastructure through which the services are delivered. The computing equipment available to us is American or European equipment, purchased at import prices that include tariffs, maintenance costs that include sending equipment overseas for service, and the specific dependency on foreign suppliers that means if IBM decides to prioritise another customer we are waiting."
He looked at Karan.
"The Ganesha mainframe," he said.
"The Ganesh-1," Karan said. "Yes."
"The system that processed the Pokhran test data," JRD said.
"The commercial version," Karan said. "The Ganesh-class. Not the specific system from Pokhran, which is at BARC. The commercial architecture on the same chipset."
"Tell me what the commercial version does," JRD said.
"47 million operations per second on the base configuration," Karan said. "Expandable to 94 million with the dual-processor configuration. 8 megabytes of core memory, expandable to 16. The 3-micron ISMC chips give it a processing density that is ahead of any commercially available system in the world."
JRD was quiet.
"Ahead of IBM's current mainframe line," he said.
"The IBM System/370 is the industry reference," Karan said. "The Model 168, which is IBM's current top-end system, runs at approximately 15 million operations per second. The Ganesh-1 base configuration is three times faster."
"Three times faster than IBM's best system," JRD said. He said it flatly. Without drama. The way he received technical information — accepting it, calibrating it, filing it.
"The chip process is the difference," Karan said. "The 3-micron ISMC process gives us transistor density that IBM won't reach commercially for four to five years."
JRD looked at the Siddharth-1 on the desk.
"The same chips that are in that machine," he said.
"The same chip family," Karan said. "Different implementation. The Siddharth's processor is optimised for interactive computing — single user, immediate response. The Ganesh is optimised for batch processing and multi-user scientific computation."
"TCS's computing workload," JRD said, "is primarily data processing — payroll calculation, inventory management, financial modelling, the batch programmes that run overnight and produce reports in the morning. A system that is three times faster than IBM's best processes three times as many batches in the same time, or processes the same batches in a third of the time, which means our operators can do more client work per hour."
"Yes," Karan said.
"The price," JRD said.
"In the range of IBM," Karan said. "We're not subsidising the equipment to build market share. The ISMC cost structure at 3-micron gives us margin at IBM-equivalent pricing."
"You're not cheaper," JRD said.
"We're faster," Karan said. "At the same price. The value proposition is performance, not price."
JRD was quiet for a moment.
"Maintenance," he said.
"In-house service team," Karan said. "The Ganesh-class systems come with maintenance contracts serviced by ISMC's field engineers. The service centres are in Bombay, Delhi, Calcutta, Madras. Parts inventory in each location. Maximum field response time: 24 hours."
"IBM's maintenance turnaround on their mainframes," JRD said.
"IBM's Indian operations service turnaround is typically three to five days because parts are staged in Singapore," Karan said. "If a component needs to go to the regional repair facility, a week or more."
"24 hours versus a week," JRD said.
"Yes," Karan said.
JRD looked at the ceiling.
"TCS currently operates two IBM 360 mainframes," he said. "Leased. The lease cost plus the maintenance plus the downtime cost — when the machines are unavailable, which is more often than IBM would like to admit — the total cost is substantial." He paused. "I want to replace both with Ganesh-class systems and I want to purchase them rather than lease them."
"Why purchase rather than lease?" Karan said.
"Because leasing means dependency," JRD said. "We are leasing IBM's equipment, which means IBM has a relationship with us that allows them to set the terms of the renewal. I want to own my computing infrastructure the way I own my factories."
Karan looked at him.
"You understand capital asset strategy very well," Karan said.
"I should," JRD said. "I have been doing it for forty years."
"The purchase price for two Ganesh-class systems," Karan said, "and the five-year maintenance contract — Aditya will prepare the specific proposal. I can tell you the range now, or we can do it properly with documentation."
"Do it properly," JRD said. "I came here to discuss the larger question, not to negotiate line items. Aditya can handle line items."
He paused.
"I want to talk about TCS," he said. "And what I think TCS can become."
"Tell me," Karan said.
JRD Tata leaned back in the chair.
He had a habit, when he was going to say something he had been thinking about for a long time, of looking at a specific point in the middle distance — not at the person he was talking to and not at the room, but at the specific point where the thought was located, the way people looked when they were reading from memory rather than composing in real time.
He looked at that point now.
"I want to tell you something about what I believe is going to happen in the next twenty years," he said. "Not in the world. In India specifically. In Indian business."
"Tell me," Karan said.
"India is going to become a services economy," JRD said. "Not entirely — we will always have manufacturing, agriculture, energy. But the growth sector — the sector that will employ the most educated people, generate the most foreign exchange, create the most wealth — will be services." He paused. "Specifically, computing services. Information technology. The specific application of computing power to the problems of business and government and everything else that large organisations do."
He looked at Karan.
"The Indian competitive advantage in this sector," he said, "is the same as it has been in every sector where we have competed internationally: an excess supply of technically excellent people at a cost that is substantially lower than comparable people in Western markets." He paused. "An IIT computer science graduate is, by any objective measure, comparable in capability to an MIT graduate. He earns, in India, perhaps fifteen to twenty percent of what his MIT counterpart earns in America. This differential is the engine that makes TCS's business possible."
"The arbitrage," Karan said.
"The arbitrage," JRD confirmed. "We take Indian talent and we apply it to problems that Western companies have and we charge Western companies a price that is below what they would pay for Western talent and above what the Indian talent costs us. The margin is the business."
"Yes," Karan said.
"This model," JRD said, "can scale enormously. The demand from Western companies for computing services is growing faster than the supply of Western computing talent. The gap grows every year. The Indian supply of computing talent, properly trained and organised, can fill that gap for a decade or more." He paused. "TCS, organised correctly, can become the world's largest computing services company within fifteen years."
Karan was quiet.
He was looking at JRD with the specific attention he brought to things that were mostly right.
"You believe this," he said.
"I know this," JRD said. "The calculations are not difficult. The demand numbers from the American market alone are enough. Add Europe. Add the domestic Indian public sector, which is beginning to computerise." He paused. "TCS is currently doing approximately eight crore rupees in revenue. In fifteen years, properly scaled, it can do a thousand crore rupees."
"And this requires the Ganesha mainframe," Karan said.
"The Ganesha mainframe is the infrastructure," JRD said. "Yes."
Karan was quiet.
He was sitting in the chair with his notebook closed on the desk beside him and his hands folded in front of him, and he was looking at JRD Tata with the expression he used when he was deciding how to say something that needed to be said carefully.
"JRD," he said.
"Yes," JRD said.
"You are describing something that I believe will happen," Karan said. "The services economy, TCS's growth, the demand from Western companies — I agree with your analysis. India's computing services sector will be one of the largest in the world within two decades."
"Then you agree," JRD said.
"I agree with the prediction," Karan said. "I want to ask you about the model."
JRD looked at him.
"The arbitrage model," Karan said. "The model you described: Indian talent, Western problems, price below the Western cost, margin is the business." He paused. "Tell me what TCS owns in this model."
JRD was quiet.
"What do you mean?" he said.
"In the model," Karan said, "TCS provides services. The client company — the American or European company that pays for the services — retains what? The output. The programme. The system that was built. The intellectual product of the computing work."
"Yes," JRD said.
"And TCS retains what?" Karan said.
JRD was quiet for a longer moment.
"The relationship," he said. "The client relationship. The track record of having delivered the work."
"And the intellectual property?" Karan said.
"The intellectual property—" JRD started.
"Belongs to the client," Karan said. "In the standard services contract, the intellectual property created during the engagement is the client's. TCS writes a payroll system for an American company. The payroll system belongs to the American company. TCS retains the knowledge of how to write a payroll system, which it can apply to the next client's payroll system."
JRD was quiet.
"Yes," he said.
"The knowledge is TCS's asset," Karan said. "The knowledge, and the people who hold the knowledge, and the relationship with the client." He paused. "But the specific product of the work — the programme, the system, the software — is the client's."
"That is the standard commercial arrangement for services," JRD said.
"Yes," Karan said. "It is. And it is a valid commercial arrangement. I am not saying TCS should not do this. I am saying something else." He paused. "The arbitrage model, at scale, looks like this: TCS employs 50,000 Indian computing professionals. Those 50,000 professionals write software and build systems for American and European clients. The software and systems are owned by the clients. TCS receives fees for the work. TCS's revenue is 1,000 crore rupees. TCS's people are employed. TCS's shareholders receive returns."
He paused.
"Now tell me," he said, "what India has accumulated, after twenty years of this, that it did not have before."
JRD looked at him.
"Wealth," he said. "Skilled employment. Foreign exchange. The international reputation of having delivered computing services at world-class quality."
"Yes," Karan said. "Those things. And what else?"
JRD was quiet.
"What do you mean by what else?" he said.
"Does India own anything?" Karan said. "In the specific sense of owning: does India have any software, any system, any intellectual property that it has built and retains? That it can licence, or sell, or build products from, or defend against competitors?"
JRD was very still.
"The client's software is the client's," he said.
"Yes," Karan said.
"The process knowledge is TCS's," he said.
"The process knowledge allows TCS to do the work again," Karan said. "It doesn't allow TCS to sell the output independently. It doesn't allow TCS to stop working and still collect. It doesn't allow TCS to say: we have built something that generates value without continuous labour." He paused. "Process knowledge is the asset of a craftsman. The craftsman is skilled. The craftsman is valuable. The craftsman's work is excellent. But the craftsman must be present and working for value to be generated. The moment the craftsman stops working, the value generation stops."
He looked at JRD.
"Twenty years of computing services at the arbitrage model," he said, "produces 50,000 excellent craftsmen and an organisation that must continually employ 50,000 excellent craftsmen to generate its revenue. The moment you stop employing them — the moment the demand drops, or a competitor offers lower prices, or the wage arbitrage narrows as India's incomes rise — the revenue stops." He paused. "You have not built an asset. You have built a capability. Capabilities are valuable. They are not the same as assets."
The room was quiet.
JRD looked at the Siddharth-1 on the desk.
He looked at it for a long time.
"The Siddharth," he said.
"Yes," Karan said.
"ISMC has built a product," JRD said. "A machine. The machine is manufactured and sold. Once sold, it continues to generate value for the buyer — calculating, programming, storing information — without ISMC doing anything further."
"And ISMC owns the design," Karan said. "The chips. The BASIC interpreter. The manufacturing process. If tomorrow every Siddharth-1 factory stopped running and no new machines were built, the machines that exist continue to work. The design continues to be owned by ISMC. The patents continue to generate licensing revenue. The capability doesn't stop when the labour stops."
"You have built an asset," JRD said.
"Yes," Karan said.
"The LED," JRD said.
"The LED patent generates royalties from Sony, Toshiba, GE, and Philips regardless of whether ISMC runs a single reactor tomorrow," Karan said. "The knowledge is crystallised in the patent. The patent is the asset. The asset generates value independently of continued labour."
JRD was quiet.
He was quiet with the specific quality of a man who had built significant things and who was, in this moment, receiving a framework for understanding something he had noticed but not fully articulated.
"You are saying," he said carefully, "that TCS's model, as I have described it, is the model of a craftsman rather than the model of an inventor."
"Yes," Karan said.
"And craftsmen are valuable," JRD said.
"Yes," Karan said. "Craftsmen are extremely valuable. India needs craftsmen. The 50,000 computing professionals that TCS will employ are doing important and necessary work and they should be respected and compensated for it. I am not saying the services model is wrong."
"But?" JRD said.
"But a country of craftsmen," Karan said, "is a country of labourers. The most skilled labourers in the world. Still labourers. Labour requires presence. Labour stops generating value when the labourer stops. A country that only exports labour — even highly skilled computing labour — is permanently dependent on the demand of the countries that buy the labour." He paused. "The wage arbitrage narrows over time. Indian incomes rise. The cost differential that makes TCS's model work in 1975 is smaller in 1985 and smaller again in 1995. The arbitrage narrows and the model that was generating 1,000 crore rupees begins generating 700 crore and then 500 crore — not because TCS is doing worse work but because the economic conditions that made the model profitable have changed."
"You are saying the model is temporary," JRD said.
"I am saying the model, unmodified, is extractive without being accumulative," Karan said. "TCS extracts value from the arbitrage. It does not accumulate assets that generate value independently of the arbitrage." He paused. "If TCS, over twenty years of services revenue, used a portion of that revenue to build products — to develop software that it owns, to create computing systems that it patents, to build the Indian computing IP that is not in the client's hands but in TCS's hands — then in twenty years TCS is both a services company and a technology company. The services revenue pays the operating costs. The technology IP generates royalties. The products generate licensing revenue. TCS is not dependent on the arbitrage because TCS owns things that have value regardless of the arbitrage."
JRD looked at him.
"You are telling me," he said, "that TCS should build products while providing services."
"I am saying," Karan said, "that a computing services company that generates substantial revenue and spends all of that revenue on operations — on salaries and computers and office space — and retains no accumulated intellectual capital at the end of each year is, in the long run, more fragile than a company that retains some of that revenue as accumulated intellectual capital. The specific form of that capital — software products, patents, proprietary methodologies that are yours to licence — is less important than the fact of its accumulation."
JRD said: "This is the distinction between a firm and a workshop."
Karan looked at him.
"A workshop," JRD said, "employs skilled craftsmen, takes orders, delivers work, is paid for the work. The workshop is entirely dependent on the order flow. No orders, no work, no income. A firm — in the sense I mean — employs skilled people, delivers work, AND develops proprietary methods and products that generate revenue independently of the order flow. The firm can survive a slow order quarter. The workshop cannot."
"Yes," Karan said. "That is the distinction."
"TCS, as I have described it, is a very large and very skilled workshop," JRD said.
"Currently, yes," Karan said. "It can be changed."
"Tell me what you mean by changed," JRD said.
"The revenue from services," Karan said. "The margin from the arbitrage. Instead of spending all of it on operations and returning the remainder to shareholders — take a portion, say fifteen to twenty percent, and invest it in the development of proprietary software and computing tools. Payroll systems that TCS owns and licences to multiple clients rather than building for each client from scratch. Database management tools. The specific software infrastructure that every large computing services engagement requires and that TCS currently builds new for each client."
"Rather than standardising," JRD said. He saw immediately what the direction was.
"If TCS builds a payroll system for one client from scratch," Karan said, "and then builds a payroll system for the next client from scratch, and then for the third client from scratch — TCS is doing the same work three times and getting paid three times. The work is valuable. The payment is fair. But TCS has not accumulated anything." He paused. "If TCS builds a payroll system once, owns it, and licences it to three clients — TCS does the work once and receives payment three times. The third payment is essentially profit with no associated labour cost. The asset — the owned payroll system — generates value independently of further work."
"Software as a product rather than as a service," JRD said.
"Software as a product," Karan said, "is the difference between owning something and doing something. Both are valuable. The company that does both owns the future."
JRD was quiet for a long time.
He looked at the Siddharth-1.
"The BASIC interpreter in that machine," he said.
"Sunita wrote it," Karan said. "Once. It's in the ROM. Every Siddharth-1 we sell has her interpreter. She doesn't rewrite it for each machine. She wrote it once and it replicates."
"The interpreter is the product," JRD said. "The machine is the delivery mechanism."
"Yes," Karan said.
"And the machine," JRD said, "is the product for the factory that manufactures it, but the factory also generates value from the manufacturing itself, which is the services component—" He stopped. He looked at Karan. "You have done both simultaneously."
"Yes," Karan said.
"The LED," JRD said. He was working through the framework. "The research produced the device — the product, owned by ISMC. The licensing of the device design to GE and Philips and Sony and Toshiba — services, in the sense of applying ISMC's expertise to helping those companies build production capability. Both simultaneously."
"The product generates royalties regardless of whether ISMC does anything further," Karan said. "The services — the production knowledge transfer — generate additional revenue. The services are additive to the product revenue. Neither depends on the other."
"You don't need the services because the product royalties are sufficient," JRD said.
"The services strengthen the relationship," Karan said. "They create goodwill. They demonstrate capability. They produce the kind of technical partnership that generates future opportunities. But they are additive, not foundational. The foundation is the product."
JRD nodded.
He was a man who nodded when he had understood something, not when he was agreeing to be polite. The nod meant: received, integrated, placed.
"TCS's services revenue," he said, "is currently the foundational revenue. The services are not additive to a product base. The services are the base. Which is—" He paused. He was completing the logic. "Which is fragile in the long run because the foundation can be undercut by wage convergence or competition or demand shifts in ways that a product base cannot be undercut."
"Yes," Karan said.
"So TCS must build the product base," JRD said.
"While building the services business," Karan said. "Not instead of it. The services business is valuable and should grow. But simultaneously, a portion of the services revenue should be reinvested in building the product base — the owned software, the proprietary methodologies, the computing tools that are TCS's intellectual property rather than the client's."
"What is the specific proportion?" JRD said.
"That is Aditya's question," Karan said. "He will have a number. My instinct is fifteen percent of services revenue, reinvested annually in product development. At TCS's current revenue, that is — perhaps one crore rupees per year. A reasonable R&D programme."
"Growing with the services revenue," JRD said.
"Yes. In five years, if TCS's revenue has grown as you project, the fifteen percent reinvestment is five crore rupees. In ten years, perhaps twenty crore. By which time the product IP being generated is significant."
"And the product revenue begins to supplement the services revenue," JRD said.
"And eventually," Karan said, "the product revenue is comparable to the services revenue, and TCS is no longer a services company that happens to own some products. TCS is a technology company that delivers services."
JRD was quiet.
He was looking at the middle distance again — the specific point where the thought was.
He was doing what he had been doing for forty years: seeing something that he was being told and understanding it not as an abstract principle but as a concrete operational plan, the plan that translated the principle into actions, into budgets, into organisational structures, into the specific decisions that would be made by specific people in specific roles over a specific timeline.
"The product development team," he said.
"Separate from the services delivery teams," Karan said. "Not integrated. Integrated product development and services delivery is what produces the situation where every client gets their own custom version of the same system, which is exactly the situation you want to avoid. The product development team builds the products. The services teams use the products as the foundation for client engagements."
"A research and development function within TCS," JRD said.
"Yes," Karan said.
"That requires a different kind of person than the people TCS currently employs," JRD said. He was thinking about this carefully. "The services business requires people who are excellent at understanding client requirements and translating them into working systems quickly. That is a specific skill — customer-facing, delivery-focused, responsive to changing requirements. The product development function requires people who are excellent at solving problems that don't have a pre-existing client specification. People who can identify a problem that hasn't been articulated as a client requirement and build a solution for it speculatively, based on their understanding of what computing services companies need."
"Those are different skills," Karan said.
"Very different," JRD said. "The services person is reactive. The product developer is speculative. In my experience in manufacturing, the factory manager and the research engineer are also very different skills. The mistake many companies make is assuming that excellence in one implies excellence in the other."
"That mistake produces companies that are excellent at services and produce mediocre products," Karan said. "Or excellent at products and mediocre at delivering them."
"Or," JRD said, "that hire product developers for the wrong reason — to make the company look innovative — without actually giving them the organisational space to develop real products. Which is worse than not having product developers, because it costs money and produces nothing."
"Yes," Karan said. "The R&D function has to be genuinely separate, genuinely protected from the quarterly delivery pressure, and genuinely funded at the level required to produce real output."
"15 percent," JRD said.
"As a starting point," Karan said.
JRD was quiet.
"Finding Sunita Raos," Karan said. "That is the organisational problem. They are not common. They require the specific combination of technical excellence and the willingness to build things that are not yet specified by a client requirement. In the culture that services businesses produce — the culture of responsiveness, of delivery, of client satisfaction — the Sunita Raos are uncomfortable. They move at a different pace. They think about different problems."
"You are describing my research engineers at Tata Steel," JRD said. He was connecting it to his own experience. "The people who worked on the new alloy grades — the work that eventually produced the high-grade steel for your armour programme. They were not factory people. They were laboratory people. They were uncomfortable in the factory because the factory operated on timelines that their work couldn't accommodate. When I gave them a separate facility and protected them from the factory's operational pressure, they produced the results. When I tried to integrate them into the factory operations, they produced less because the factory's culture consumed them."
"Yes," Karan said.
JRD looked at the Siddharth-1.
He reached out and touched the keyboard.
"May I?" he said.
"Please," Karan said.
JRD typed, slowly — he was not a touch typist, he was a hunt-and-press typist, the typing of a man who had learned to type in the 1930s when typing was a skill that senior executives were not expected to have:
PRINT 1 + 1
The machine said: 2
He looked at it.
"The machine answered immediately," he said.
"Yes," Karan said.
JRD looked at the screen.
"The services model," he said, "is: I bring my expertise, I answer the client's question, I depart. The client has the answer. I have the fee."
"Yes," Karan said.
"The product model is: I build a machine that answers questions without my having to be present for each answer. The machine answers. I collect the machine's deployment revenue."
"Yes," Karan said.
"The machine is not a replacement for expertise," JRD said. "Expertise built the machine. Expertise maintains and improves the machine. But the machine's operation does not require expertise to be present at every answer."
"That is the distinction," Karan said. "The product crystallises the expertise. The product operates independently of the expert's continuous presence. The expert is freed to develop more expertise and to crystallise more of it into more products."
JRD looked at the blinking cursor.
He looked at it for a long time.
Then he said: "I have been running companies for forty years. In all of that time — in steel, in chemicals, in hotels, in aviation, in automobiles — the principle has been: build the asset. The factory is the asset. The hotel is the asset. The aircraft is the asset. The asset generates value when properly managed."
He paused.
"When I conceived of TCS, I thought of it as a service business, which meant I thought of it differently from how I think about a factory. A factory is an asset that I own and operate. A service business is an expertise that I apply. I did not think of TCS's knowledge and methods as assets in the same way I think of a factory as an asset." He paused. "You are telling me that the knowledge and methods can be crystallised into products — into software — that behave as assets in exactly the same way."
"Yes," Karan said.
"The software is the factory," JRD said.
Karan looked at him.
"The software is the factory," he said. "The software runs without the engineer present. The software produces output without the engineer's continuous application of expertise. The software is the crystallisation of the engineer's expertise into an asset that operates independently." He paused. "The patent is the factory. The LED patent generates royalty income without my scientists continuously working on LED devices. The work was done once, crystallised into a patent, and the patent operates independently."
"Yes," Karan said.
JRD leaned back in the chair.
He was quiet for a moment.
"I am going to make one change to your framework," he said.
Karan waited.
"You said fifteen percent of services revenue reinvested in product development," JRD said.
"As a starting point," Karan said.
"I will do twenty-five percent," JRD said. "In the first five years. While TCS is building the product capability and the culture of product development. After five years, when the product revenue begins to be meaningful, I'll reduce the reinvestment percentage because the product revenue itself funds the next generation of product development."
Karan looked at him.
"That is more aggressive than I suggested," he said.
"I am seventy-one years old," JRD said. "I do not have the luxury of being modest with my timeline." He paused. "Twenty-five percent for five years establishes the product function with sufficient resources to produce real output. At fifteen percent, you spend three of the five years building the capability and two years beginning to produce. At twenty-five percent, you spend two years building and three years producing. The output in year five is dramatically different."
"The services teams will notice the margin compression," Karan said.
"Yes," JRD said. "And I will explain why the margin is being compressed and what the reinvestment is expected to produce." He paused. "In my experience, people who understand the reason for a difficult decision accept it more readily than people who are told the decision without the reason." He paused. "If I tell TCS's leadership that twenty-five percent of margin is being reinvested in product development because TCS's services revenue is structurally fragile without a product base, and I explain what I mean by structurally fragile in precisely the terms you have used with me this afternoon, they will understand. They may not like it. They will understand."
"And if they don't like it?" Karan said.
"Then they will find a company whose leadership is content to build a very large and very skilled workshop," JRD said. "And TCS will find people who want to build a technology company."
He said this with the quiet finality of a man who had been making decisions like this for forty years and who had learned that the quality of the decision was more important than the comfort of the people it affected.
"You are going to do this," Karan said.
"I am going to do this," JRD confirmed.
They were quiet for a moment.
Outside, the Gorakhpur complex was doing what it did on a Republic Day: reduced activity, a kind of settled quality, the factory breathing at lower intensity. Through the study window, the flag was visible — the tricolour on the factory's main building, which flew every day but which on Republic Day flew with the specific significance of twenty-five years of independence, the silver anniversary of a country that had been trying to become what it said it was.
JRD looked at the flag.
"Twenty-five years," he said.
"Yes," Karan said.
"When independence came," JRD said, "I was forty-four years old. I had been running the Tata group for six years. The specific hope I had in 1947 — the specific hope that every industrialist who was serious had — was that independence would produce a government that understood the relationship between industrial capability and national power, that would make decisions in support of industrial development rather than in restriction of it." He paused. "The License Raj was the opposite of that. For twenty years — from roughly 1950 to 1971 — every decision about what to produce, how much to produce, where to produce it, required a licence from the central government. The government that was supposed to enable industrial development was strangling it instead."
"1971," Karan said.
"The year the stranglehold began to loosen," JRD said. "And since then—" He looked at the factory complex through the window. "Since then, Tata Steel has done in four years what it could not do in twenty."
"Tell me," Karan said.
"The capacity expansion," JRD said. "We had been requesting permission to expand the Jamshedpur facility since 1955. Permission refused, or delayed, or granted in partial form that made the expansion uneconomical. In 1971, when the industrial licensing regime was rationalised, we applied for expansion again. Approved in three months." He paused. "We built in eighteen months what we had been trying to build for sixteen years. The steel production capacity at Jamshedpur is now three times what it was in 1971."
"The pipeline steel," Karan said.
"The Barmer-Jamnagar pipeline," JRD said. "And the Bombay High infrastructure. The specific grades of pipe steel that your petroleum programme required — we produced them. We could produce them because we had the capacity. We had the capacity because we were finally allowed to build it."
"The economy," Karan said.
"The specific relationship between industrial capacity and economic output is not subtle," JRD said. "When companies can produce what the market demands without government permission for each production decision, the market grows, the companies grow, the employment grows, the tax revenue grows, the government can fund more — everything expands. The License Raj made everything contract. Its removal made everything expand." He paused. "Four years of growth since 1971 has exceeded twenty years of constrained growth before it."
Karan said: "Tata Motors."
"TELCO," JRD said. "The commercial vehicle programme. We were producing trucks and buses under licence from Mercedes-Benz. The licence relationship meant that every design decision, every specification, every new model required approval from Stuttgart. When the licensing restrictions were rationalised, we could make our own decisions. In 1972, we decided to indigenise — to replace the Mercedes components with components designed and manufactured in India." He paused. "The Tata truck that is now in production is substantially Indian in design. Not entirely — we still use foreign components for certain specifications we haven't yet mastered. But the trend is clear. Every year, the Tata truck becomes more Indian."
"What is the timeline to full indigenous design?" Karan said.
"Three to five years for the standard commercial vehicle range," JRD said. "Longer for the higher-specification vehicles — the heavy construction vehicles, the specialised industrial trucks. Those require engineering capability that takes longer to develop."
"SPEI could help," Karan said.
JRD looked at him.
"The specific process engineering challenges of the Tata truck's indigenous component programme — the precision machining of the engine components, the gearbox assembly, the suspension system fabrication — these are scale-up problems in the same sense that the LED scale-up was a scale-up problem. The SPEI methodology may be applicable." He paused. "Bring it to Pillai. He's the right person to assess the applicability."
JRD wrote something in the small notebook he had produced from his jacket pocket.
"Pillai at BHU," he said.
"SPEI's educational director," Karan said. "He'll be in Gorakhpur at the end of the month. I'll make the introduction."
JRD wrote.
"The Taj," JRD said.
"Hotels?" Karan said.
"The hotel business is — booming is the word everyone uses," JRD said. "The specific mechanism is the same as the rest: the License Raj removal allowed us to build where and what the market demanded rather than what the government permitted. We have opened four new Taj properties since 1971. We have four more in planning." He paused. "The specific demand driver is foreign business travel to India. As India's industrial and commercial activity expands — as your petroleum programme and your defence exports and now the computing business bring foreign commercial visitors — those visitors need hotels of international standard. Taj provides that."
"The Gorakhpur hotel," Karan said.
JRD looked at him.
"You need one," JRD said.
"The Japanese delegation," Karan said. "The Sony and Toshiba engineers who visited in March. We put them in the factory guest facility, which is functional but is not what international visitors expect."
"The Sony and Toshiba visit was public knowledge?" JRD said.
"The LED technology transfer visit," Karan said. "It generated press."
"Yes," JRD said. "I read about it." He was quiet for a moment. "A Taj property in Gorakhpur. The specific location — not in the city centre, adjacent to the industrial complex. A hotel for business visitors rather than for tourists."
"Is that viable?" Karan said.
"Let me look at the numbers," JRD said. "The specific visitor volume — international business visitors to the Shergill complex — it would require a minimum threshold to support the fixed costs of a quality property." He paused. "But the trend is clear. As your programmes grow and your international commercial relationships multiply, the visitor volume grows with them." He made another note. "I'll have our hotels development team look at it."
Karan said: "Tata Power."
"The SEZ grids," JRD said. "The specific contractual arrangement — Shergill's special economic zones require reliable industrial-grade power that the national grid cannot guarantee. Tata Power provides the backup and stabilisation capacity. It is a straightforward commercial arrangement."
"It's more than that," Karan said.
JRD looked at him.
"The arrangement created a model," Karan said. "For other SEZs. For other industrial zones that are being established as part of the license raj reform's consequence. The specific model — private power company providing industrial-grade power under long-term contract to industrial clients, separate from the national grid — that model is being replicated."
"Yes," JRD said. "We have four similar arrangements under discussion. The specific demand — reliable industrial power — exists everywhere that serious manufacturing is being established."
"Infrastructure as a business," Karan said. "Not infrastructure as a government function."
"In the right political environment," JRD said. "Which we now have." He paused. "Partially. The government still owns the national grid. The private arrangement works at the margin of the national grid, not as a replacement for it. The full private power generation model — where a private company builds generation capacity and sells power competitively — is not yet politically possible."
"It will be," Karan said.
JRD looked at him.
"You are confident about this," he said.
"The economics are inevitable," Karan said. "The national grid cannot keep pace with industrial demand growth. The demand will force the policy. The policy will open the market. When the market opens, Tata Power is the natural leader — it already has the infrastructure, the expertise, the commercial relationships."
JRD wrote.
"The timeline on this," he said.
"Ten years," Karan said. "The policy reform that opens power generation to private competition. Not before ten years. Possibly fifteen."
"I will be eighty-one or eighty-six," JRD said. "Whether I am still running the group by then—"
"The group will be positioned correctly regardless," Karan said.
JRD looked at him.
"You think about succession," he said.
"You have built something that outlasts any individual," Karan said. "Including you. The Tata group's institutions, values, commercial relationships — these survive the chairman."
"That is the intention," JRD said. "It is also not guaranteed. Institutions outlast their founders only when the successors understand the founders' intentions accurately enough to carry them forward in new contexts." He paused. "The specific intention that has guided the Tata group — that business should serve the nation's development, that profit is the measure of sustainability rather than the purpose of the enterprise — that intention must be understood by successors who have not lived the founding moment."
"You should write it down," Karan said.
"I have been writing it down for thirty years," JRD said. "Letters, memoranda, annual reports. The writing exists." He paused. "The reading of the writing is another matter."
"Make the reading obligatory," Karan said. "Not as a formality. As a condition. The specific understanding of what the group is and what it is for — make that understanding a prerequisite for leadership."
JRD looked at him.
"You are twenty-four years old," he said.
"I know," Karan said.
"You are telling me how to manage succession in a company I have been running for thirty-seven years," he said.
"Yes," Karan smiled and said.
JRD looked at him for a moment.
Then he said: "You are probably right."
He said it with the specific quality of a very experienced man acknowledging something from a very young man that the very experienced man should have already arrived at independently — not grudgingly, with the honest acknowledgement of a man who respected the quality of the thought regardless of the source.
At five in the afternoon, Aditya came in.
He came in the way he entered rooms where significant conversations had been happening — with complete attention to what the room contained, the quality of presence that he had developed across four years of being in the room where things were decided. He looked at JRD and at Karan and at the Siddharth-1 on the desk and at the notebooks that both of them had been writing in.
"The Ganesha configuration for TCS," Aditya said. "Do you want the proposal now or formally tomorrow?"
"Formally tomorrow," JRD said. "Tonight I want chai and then I want to try programming this machine." He indicated the Siddharth-1. "Karan offered one. I will take him up on it this evening if one is available."
"I'll have the demo unit packaged for you before you leave tomorrow morning," Aditya said.
"Good." JRD looked at Aditya. "The TCS product development investment. Twenty-five percent of services revenue in the first five years. After year five, I'll need a financial model for the product revenue ramp and the reinvestment rate adjustment."
Aditya looked at him. Then at Karan.
"I'll build the model," Aditya said.
"Thank you," JRD said.
He looked at Karan.
"I want to ask you something," he said.
"Ask," Karan said.
"The Siddharth-1," JRD said. "Your personal computer. It launched four days ago."
"Yes," Karan said.
"The specific product," JRD said. "The keyboard, the screen, the BASIC interpreter, the cassette storage — this is the machine that anyone can use."
"Yes," Karan said.
"TCS's services business," JRD said, "provides computing services to large clients. Corporations. Government agencies. The specific scale of client that can justify an IBM mainframe or a Ganesh mainframe." He paused. "The Siddharth-1 is a machine that a person can buy for ₹4,800 and use at home or in a small office."
"Yes," Karan said.
"The software for the Siddharth-1," JRD said. "The programmes that people would want to run on it. Accounting software. Inventory management. Payroll for a small business. These are the computing needs of the Indian small business sector, which is orders of magnitude larger than the corporate sector."
He was looking at Karan with the specific expression of someone who has just found the connection between two things he was holding separately.
"TCS's product development," he said. "The specific products — the software that TCS develops and owns — could be software for the Siddharth-1."
Karan looked at him.
"The small business computing market," JRD said. "Shopkeepers. Accountants. Small manufacturers. The people who need calculation and record-keeping and financial management but who are not large enough to need a mainframe. They need the Siddharth-1. And they need software for the Siddharth-1 that solves their specific problems."
"Yes," Karan said.
"TCS builds that software," JRD said. "TCS owns it. TCS licences it. The revenue from the software licences supplements the mainframe services revenue." He paused. "ISMC produces the hardware. TCS produces the software. Two different companies, owned by two different groups — Shergill and Tata — serving the same market from complementary positions."
Karan looked at him.
He thought about what was happening in this moment — what the conversation had produced. The specific connection that JRD had just drawn: the personal computer market and the software market, serving the same small-business computing need, from complementary directions.
He thought about what that connection would become.
He thought about a future he had seen and was building toward — the specific future where the personal computer revolution produced not just hardware but an entire software ecosystem, where the companies that owned the software were the companies that defined what the hardware was used for, where the real value was not in the machine but in what the machine ran.
He thought about TCS, in this conversation's consequence, becoming not only a services company but a software company. A company that owned the programmes that ran on the Siddharth-1 and its successors. A company that was, in the language of the future, both a services company and a software company.
He thought: this is the conversation that changes what TCS is.
He said: "If TCS builds software for the Siddharth-1, ISMC will list that software in the Siddharth-1 software catalogue. We are building a library of software for the Siddharth — programmes written by users and by companies like TCS that run on the machine. The software catalogue ships with every Siddharth-1."
"A distribution channel," JRD said.
"Yes," Karan said.
"For TCS's product software," JRD said.
"Yes," Karan said.
"Every Siddharth-1 buyer," JRD said, "receives a catalogue of software available for their machine. TCS's accounting software is in that catalogue."
"Yes," Karan said.
"And buys TCS's software," JRD said.
"And pays TCS's licence fee," Karan said.
JRD looked at the Siddharth-1.
"You built the distribution channel into the hardware," he said.
"The catalogue ships with the hardware," Karan said. "Yes."
"This is very specific," JRD said.
"Yes," Karan said.
"You planned this," JRD said.
"I built the hardware as a platform," Karan said. "A platform needs software. The software needs a distribution mechanism. The catalogue is the mechanism."
"The software developers," JRD said, "who write for the Siddharth-1 — they sell through the catalogue, which ISMC distributes with every machine."
"Yes," Karan said.
"And ISMC takes a percentage," JRD said.
"Yes," Karan said.
"What percentage?" JRD said.
"That is Aditya's number," Karan said.
"Aditya," JRD said, to the young man in the doorway who was still taking notes.
"Twenty percent," Aditya said, without looking up.
JRD looked at Aditya.
"Of every software sale through the catalogue," Aditya said. "The software developer receives eighty percent. ISMC takes twenty percent for the distribution and the catalogue position."
JRD looked at the Siddharth-1.
"So," he said. "TCS builds accounting software for the Siddharth-1. Lists it in the ISMC catalogue. Every Siddharth-1 buyer who wants accounting software finds TCS's product in the catalogue. Pays TCS's licence fee. TCS receives 80 percent. ISMC receives 20 percent."
"Yes," Aditya said.
JRD was quiet for a moment.
"That is," he said, "the specific model that I need TCS to enter."
He looked at Karan.
"You have built the system that I need TCS to plug into," he said.
"Yes," Karan said.
"The distribution channel. The hardware platform. The customer base that grows with every Siddharth-1 sold." He paused. "The software market that TCS can serve is the software market that your hardware creates."
"Yes," Karan said.
"You thought about this before building the hardware," JRD said. It was not a question.
"I thought about what a personal computer needed to be useful," Karan said. "Useful requires software. Software requires developers. Developers need distribution. Distribution requires a mechanism. The mechanism is the catalogue."
JRD looked at him.
"You built the ecosystem before the participants knew they were going to be in it," he said.
"I built the foundation," Karan said. "The ecosystem builds itself once the foundation exists."
JRD was quiet for a long time.
He was looking at the Siddharth-1. At the blinking cursor. At the machine that was waiting for someone to type something.
"I came to Gorakhpur today," he said, "to buy computing equipment for TCS. I am leaving with a different understanding of what TCS should be and a commercial position in the ecosystem you're building." He paused. "I believe this is what always happens when I come to Gorakhpur."
"You usually leave with something you didn't expect," Karan said.
"Yes," JRD said. He stood. He stretched slightly — the stretch of a seventy-one-year-old who had been sitting for several hours and whose body registered this. "I want chai," he said. "And then I want to learn BASIC."
"There is a Quick Reference Guide," Aditya said.
"Good," JRD said. "I have always been a quick study."
At nine in the evening, JRD was still at the Siddharth-1.
Karan had shown him the basics — literally, the BASIC interpreter basics — in the first twenty minutes. Then he had left JRD alone with the machine and the Quick Reference Guide, because JRD was the kind of person who learned things by working at them rather than by being taught, and the presence of a teacher interfered with the working.
He had come back an hour later to find JRD had written a programme that calculated the compound interest on Tata Steel's long-term debt at various interest rate scenarios.
He looked at the programme.
He looked at the output — a table of scenarios, each row a different interest rate, each column a different number of years, the compound interest calculated at the intersection.
He said: "The table formatting."
"I found the TAB function," JRD said. "In the guide. It allows you to specify the column position. The table is formatted using TAB to align the columns."
Karan looked at the programme.
It was seventy-three lines of BASIC. The data entry was handled with INPUT statements that prompted for the parameters. The calculation was in a nested FOR-NEXT loop. The output was formatted with TAB into aligned columns.
"This would have taken you how long with a calculator and paper?" Karan said.
JRD thought about it.
"Two hours," he said. "With errors. The recalculation if I changed a parameter — another hour."
"With the Siddharth-1?" Karan said.
"I wrote the programme in forty minutes," JRD said. "The first run had two mistakes — a misplaced parenthesis and a wrong variable name. I corrected them in five minutes. The programme now runs in three seconds and I can change any parameter instantly."
He looked at the output.
"The specific calculation I have wanted to do for three months," he said. "I have been asking the finance team to run the scenarios. They run them and they send me tables. The tables are correct but they represent a specific set of parameters that they chose. If I want different parameters, I ask again and wait." He paused. "Now I choose my own parameters."
"Now you have the tool," Karan said.
"Now I have the tool," JRD confirmed. He looked at the machine. "The specific feeling of having the answer immediately rather than waiting for it — that is not trivial. The delay is not only a time cost. The delay changes the quality of the thinking. When you wait for an answer, you plan the next question in advance. When you have the answer immediately, you can think about what the answer means and ask the next question based on the meaning."
"The interactive nature of the machine," Karan said.
"The machine is interactive," JRD said. "The mainframe is not interactive. The mainframe runs a batch. You submit the batch. You wait. You receive the output. You plan the next batch based on the output." He paused. "The Siddharth is a conversation. The mainframe is correspondence."
"That is the distinction," Karan said.
"TCS's clients," JRD said, "use the mainframe for the correspondence — the large batch processes, the monthly payroll, the quarterly financial reconciliation. They would benefit from the conversation — the interactive analysis, the scenario modelling, the quick calculation that produces insight rather than record." He paused. "TCS could provide both."
"TCS could provide the software that runs the conversation on the Siddharth-1," Karan said, "while also providing the services that run the correspondence on the Ganesh mainframe."
"Two products," JRD said. "One for the desktop. One for the data centre."
"Both owned by TCS," Karan said.
"Both owned by TCS," JRD confirmed.
He saved the programme to cassette.
He watched the machine write to the tape.
He looked at the screen: SAVED: INTEREST
He said: "I want to show this to Nani Palkhivala."
Nani Palkhivala was one of India's most distinguished lawyers and one of the Tata group's closest legal advisors — a man of formidable intelligence who was also, as it happened, deeply interested in economic policy.
"He'll find uses for it in fifteen minutes," JRD said.
"Send me his address," Karan said. "I'll have a unit sent."
"He'll want to understand how it works," JRD said.
"The Quick Reference Guide is sufficient for most people to get started," Karan said. "If he wants to understand more deeply, Sunita Rao is writing a programming textbook. Available by July."
"A textbook," JRD said.
"The Siddharth-1 Programming Guide," Karan said. "Six chapters. Appropriate for someone with no prior computing background. Sunita wrote the BASIC interpreter. She knows which parts of the language are difficult for new users and which are straightforward."
"She should also teach," JRD said.
Karan looked at him.
"The specific skill of explaining computing to people who have no computing background," JRD said. "It is not the same skill as computing. It is the skill of translation — translating between the machine's logic and the human's logic. Sunita apparently has both skills."
"She's twenty-eight years old," Karan said.
"Ability is not age-dependent," JRD said. He said it with the tone of a man who was making a point that had personal relevance.
Karan said: "She is going to become quite famous."
"Yes," JRD said. "The programmer who wrote India's first personal computer's operating language. The textbook author. The teacher." He paused. "In five years, every engineering student who uses a Siddharth-1 will know who she is."
"She'll find that uncomfortable," Karan said.
"Good people usually do," JRD said.
At ten-thirty, JRD closed the Siddharth-1.
He had been programming for two and a half hours. He had written three programmes: the compound interest table, a currency conversion calculator that he had extended to handle multiple currencies, and a schedule planner that organised his meeting priorities by a scoring system he had developed on the fly.
He sat back.
"Tomorrow morning," he said. "Before I leave. The Ganesha proposal."
"Aditya will have the formal proposal ready at seven," Karan said. "The configuration, the price, the maintenance contract, the delivery timeline."
"Good." JRD looked at the darkened screen of the Siddharth-1. "You know," he said, "the specific quality of the machine — the fact that it waits, that the cursor blinks, that it is ready for the next question — it reminds me of something." He paused. "When I was learning to fly, in 1929, the aircraft I learned on was a Puss Moth. De Havilland. A beautiful aircraft. It was responsive in a way that modern aircraft are not — modern aircraft have so many systems that mediate between the pilot's input and the aircraft's response that the direct connection is gone. The Puss Moth responded directly. You moved the stick, the aircraft moved." He paused. "This machine has that quality. You type, the machine responds. Direct. No mediation."
"That was the design intention," Karan said. "The machine should feel responsive. If the machine feels slow — if there is any perceptible lag between the user's action and the machine's response — the user begins to feel managed rather than served. The machine should feel like it is waiting for you rather than like you are waiting for it."
"The Puss Moth never made me wait," JRD said.
"No aircraft should," Karan said.
JRD looked at him.
"You fly?" he said.
"No," Karan said. "But the principle applies to everything that is supposed to serve the person using it. The aircraft, the machine, the factory, the hospital, the school. The institution should feel like it is waiting for the person, not like the person is waiting for the institution."
JRD was quiet.
"In forty years of building organisations," he said, "I have not heard that stated so precisely."
"It is perhaps more obvious when you are young," Karan said. "When you are young, the institutions that are supposed to serve you — the schools, the government offices, the procurement bureaucracies — they all make you wait. They are not waiting for you. You wait for them." He paused. "Everything I have built has been built so that it waits. The aircraft waits for the pilot. The tank waits for the crew. The machine waits for the programmer. The factory waits for the customer's order."
JRD looked at him.
"That," he said, "is the philosophy."
"Yes," Karan said.
"I have been looking for the word for what I have been trying to do my entire career," JRD said. "The word is: wait." He paused. "The Tata group should wait. The hotel should wait for the guest. The steel mill should wait for the customer's specification. The airline should wait for the passenger." He paused. "Waiting is not passivity. Waiting is readiness. The aircraft that waits for the pilot is an aircraft that has been made fully ready for the pilot's arrival."
JRD was quiet for a moment.
Then he said: "I am going to sleep. I am seventy-one years old and I have been awake since five in the morning and I have learned to programme a computer and I have had one of the most productive conversations of the past decade and I am very tired."
Karan stood.
"The guest room is ready," he said.
"I know," JRD said. "Leela Devi told me at dinner that she had prepared it and that she hoped I did not require special pillows because the pillows were the household's standard pillows and they were perfectly adequate for anyone who was not unnecessarily particular."
Karan looked at him.
"She said that?" he said.
"She said it very directly," JRD said. "I found it entirely charming." He stood. "She also said that if I was cold in the night, there were extra blankets in the cupboard and I should get them myself rather than waking the household."
"She is consistent," Karan said.
"She is excellent," JRD said.
He walked to the door.
At the door he stopped.
He said, without turning: "The services without ownership observation. The workshop versus the firm. I want you to know that what you said this afternoon—" He paused. "I have been in business for forty years. I have had many important conversations. The specific observation — that a service without crystallised ownership is labour, and labour is the most fragile form of value generation — I have known the edges of this idea for years and have not seen it stated plainly." He turned. He looked at Karan. "Thank you for stating it plainly."
He walked to the guest room.
Karan stood in the study.
He looked at the darkened Siddharth-1.
The cursor was not visible — the screen was off. But he thought about the cursor, blinking in the dark, ready, waiting.
He opened his notebook.
He wrote: January 26, 1975. Republic Day. JRD Tata visited. We discussed TCS, the services model, the workshop versus the firm, the software catalogue, the two-product strategy.
He paused.
He added: He learned BASIC in two and a half hours. He wrote three useful programmes. He has been running one of India's largest industrial groups for thirty-seven years and he treated the machine with genuine curiosity rather than with the condescension of a senior person encountering a junior person's tool.
He paused.
He added: This is why the Tata group has lasted. The intellectual honesty. The willingness to receive an idea from any source, evaluate it on its merits, and apply it without needing to have had the idea first.
He looked at what he had written.
He added: The cursor blinks in the dark. Ready. Waiting. This is what everything should be.
He turned off the desk lamp.
He went to bed.
Outside, on the factory's main building, the tricolour flag was dark in the January night.
Twenty-five years of Republic Days.
The twenty-sixth beginning tomorrow.
End of Chapter 181
Meeting Record — 26 January 1975 (Republic Day)Location: Shergill Industries, GorakhpurAttendees: JRD Tata; Karan Shergill; Aditya Shergill (afternoon)
Commercial:
Ganesha mainframe for TCS: 2 units, formal proposal for review January 27 TCS software for Siddharth-1 catalogue: principle agreed, terms to be negotiated Taj Hotel Gorakhpur: feasibility study commissioned SPEI-Tata Steel collaboration: introduction to Pillai to be arranged
Strategic:
TCS product development investment: 25% of services revenue, years 1-5; reduce to 15% as product revenue grows TCS two-product strategy: mainframe software (services) + Siddharth software (desktop) Software distribution through ISMC Siddharth catalogue: 80/20 split (developer/ISMC)
JRD's BASIC session: 2.5 hours. Three programmes: compound interest table, currency converter, meeting schedule planner. Assessment: "The machine is a conversation, not correspondence."
Karan's notebook, January 26:The cursor blinks in the dark. Ready. Waiting. This is what everything should be.