India's IT Tale: Part I
An overview of Indian aspirations in technology.
2026-08-15(Sat) 14:00 AST
India is known for a lot of things; it is, after all, a large country with very many people. But one association people make with the country is the association of India with software, and especially outsourced software.
This association is not without truth: Roughly 25% of India’s total exports come entirely from the software industry. The sector employs many millions of people, and employment in the sector allowed India’s large technical labour base to be put to good use. More striking is the nature: Indian software exports aren’t merely outsourced call centres, much in the way of the Philippines; nor are they the high value, foreign multinational-dominated fields that Ireland and Israel (before Israel’s post-2008 growth in startups and indigenous software) seem to specialise in. The field is dominated by unglamorous middle work that, nevertheless, is often done by local conglomerates, something rare in the new software industries of countries.
This article explores how India developed its software and wider tech ambitions. Part II will focus on why India chose software (the seemingly random choices India made in development until recently are not truly random.)
First, a history
India is not a new entrant into the field of computing; in 1956, the country was the first in Asia to acquire a digital computer, with the acquisition of the British HEC-2M which India modified and built the programming for the computer itself. In 1960, the country commissioned its first computer, the TIFRAC, inspired by the ILLIAC-1, itself based on von Neumann’s IAS machine. While already outdated (the TIFRAC took 5 years to be built; in that time, the vacuum tube basis of computers, like the ILLIAC-1, was already replaced with transistorised computing), it provided important programming exposure for many of India’s sophisticated computer users later down the line.
But Indian computing would be most defined in its early years by the involvement of IBM. The computing behemoth had entered the country in 1951, and 8 years later was starting to manufacture IBM computers inside India. This manufacturing mostly took the form of refurbishing (or, as IBM would call it, “rebuilding”) old computers from the US; after URMs were replaced there with the 1401, they ended up in India, with 1401s themselves meeting the same fate when replaced with the System/360. IBM in some years made up around 80% of the computer market share in the country, and this position allowed for some wonderful extortion; renting a computer was far more expensive, with some computers going for 1,200 dollars abroad while costing 20 thousand in annual rental in the country (Indian computing was mostly on a rental basis as the upfront costs of purchases were too much for many companies.)
Eventually, it got to a point. Legislation was amended in the 70s to force foreign companies to divest equity to domestic actors, and by 1977 the government demanded IBM do as the law said on the threat of being expelled from the country. Upon IBM’s refusal, it was promptly expelled from the country. The expulsion of IBM from India marked a turning point in the ambitions of the country’s sector, with the few domestic actors in the scene taking advantage of it, as will be discussed below.
IBM v. Everyone Else
The shape and nature of the Indian IT industry would begin to take place a few years before the expulsion of IBM. In 1972, the newly created Department of Electronics, or DoE, was tasked with increasing the foreign exchange reserves in India. They settled on the Software Export Scheme that year; the department would allow the import of computers, an import tightly controlled, by a business, in exchange for the business promising to export software using the computer that would be worth the computer, later on changed to worth double the computer.
After a three year delay in actually importing a computer due to licensing issues (never change, India…), a subsidiary of the Tata Group, Tata Consultancy Services, would become the first company to partake in the SES. TCS’s leader, F.C. Kohli, used his extensive computing connections in the US to land a contract with computing company Burroughs.
TCS did not own a Burroughs computer, something the company was unaware of. In spite of this, TCS wrote the code and impressed Burroughs enough for the company to be able to land an independent IT contract with the Detroit Police Department, a rather unusual first choice.
While TCS was starting to shape the IT industry, India had an IBM problem; in 1974, the country had only 224 computers, and more than half came from IBM. IBM reached this position through less than desirable means; for example, the company increased the price of its computers in the country when rupee devaluations occurred, because of leases and rentals being denominated in dollars. There was also significant upmarking of products, and a parliamentary report in 1976 accused IBM of transfer pricing, exaggerating expense claims, underpaying excise, and misusing the import entitlement to import “junk” computers that were far less valuable elsewhere, while offering much higher prices.
In 1973, under the FERA legislation, IBM was ordered to reduce its holdings in its own India operations; IBM deferred for two years after winning an extension conditional on eventually doing so, and then won another 2 year extension. But by 1977, with a new government in charge, India’s patience ran thin and the company was plainly told by Industry Minister George Fernandes to “get out”, which it did.
The expulsion of IBM left a vacuum in the Indian computing scene, with around a thousand employees laid off. The vacuum would be quickly filled by companies: some founded privately by ex-IBM employees, such as Computer Maintenance Services, International Data Management, and other small companies, as well as other companies like TCS which wanted ex-IBM talent for their emerging mode of operation, and the state-owned Computer Maintenance Corporation (different from CMS) which the government forced IBM to hand over its maintenance operations to. IBM’s departure also forced the rise of minicomputers and microcomputers, which often ran UNIX. Before UNIX use became more widespread in the 90s, the Indian banking sector was already being digitised by the OS, and UNIX knowledge became an important selling point for Indian programmers wanting to outsource work, which we will get to right now.
Coolies to Cables
Software, especially the outsourcing of it, is much more labour intensive than it is capital intensive. What Indian software companies were starting to sell to the outside world was a large and cheap pool of good-enough labour.
In the early days of the sector, most export contracts were lower-skilled “time and materials” contracts; Indian programmers, often paid ten times less than an American programmer of comparable quality, flew to the site of their clients often for a few months to work on the computers there.
This contract work is relatively low stakes; the earliest and least skill intensive parts of coding, which itself as a whole only made up 15% of a software project’s revenue, was the first to be outsourced in this way. This was low status work, known as “body shopping”, and the programmers engaged in it were called “cyber coolies”; accusations of slave labour were levelled against HP for its hiring of contract TCS workers in the 90s.
Nevertheless, it allowed Indian programmers and their contracting companies to get a positive reputation, a thing that can make or break software contracts, and which began the stereotype of the Indian IT guy. It would also allow them to potentially increase the work they could do in status, which leads us to the establishment of satellite and phone links, that could allow work to be done more remotely.
As early as the 70s, near the start of the IT industry’s growth in the country, feasibility studies showed that satellite links could allow US clients to offload work in even cheaper ways without needing to do expensive flying back and forth; experimental links with Italy were already established, and in 1984 Indian policy was changed to permit software exports to happen via satellite and phone links, which would have the side effect of reducing the incidence of body shopping and prove especially important near the Y2K bug.
The first big company to establish a satellite link was Texas Instruments, which built a satellite station in Bangalore. Fresh from an internal crisis after its failed entrance into home computers, and wanting to cut costs and invest in what it saw as a growing market, TI built a link to its headquarters so that its software could be built without the need for bodyshopping, a practice later followed by other companies.
It was not solely satellite links either; phone and data links played an important role. But while Indian satellite links were nonexistent, Indian phone links were…just bad.
Some outsourcing companies were simulating how to get documents and information from the US to India in the apartments of their executives, having to rely entirely on written communication to get information from one room to the other, as a representation of just how poor the phone system was. Less anecdotally, it was described as “perhaps the world’s most archaic, inadequate, overloaded and inefficient phone system”, with no faxes, or networking, or e-mailing. The waiting list for a landline number was half a million, and it took up to five years of waiting to get one.
Of course, this had to change. India had only one international telecommunications provider, VSNL, owned by the state through the Department of Telecommunications. For a 64 kbps half circuit, VSNL charged 180 thousand dollars a year, more than twice the global average for a data link of the same speed; this, alone, led to many software contracting companies that couldn’t do bodyshopping due to lack of resources deciding to close shop. But when the state, through the DoT, began demanding payment, it was surprised to find out that…the state, again, was offering to pay for the links itself for the software firms.
The Department of Electronics, the same people behind the SES, found that India had only 2 weeks of foreign exchange reserves left in the early 90s, and tasked with increasing that number through the software industry, decided to begin stonewalling the DoT. First, the DoE began paying for the software links itself, then the DoE escalated further by illegally importing CGR equipment from France and setting up its own data links at far cheaper prices. This forced the DoT’s VSNL to cut their own prices to be roughly in line with the global average. In 2 years, high speed data and phone links went from only three, to two hundred.
This reminds us of the unusual role the state played in developing the IT industry, leading me to..
Not a beauty contest
Pramod Mahajan, who served as the IT minister of India, once said the following: India leads "in IT and beauty contests, the two areas the government has stayed out of."
Mahajan's quip is representative of what many Indians themselves think is the explanation for why India specifically succeeded at IT specifically, and not any other technology intensive field like much of hardware. I would like to argue against this, and give my own reasons for why India succeeded.
To begin with, India invested heavily into technical and engineering education starting in the 1950s, contemporarily seen as "overinvesting"; this investment allowed even poorer people to attend a good technical school and study topics that would be the foundation for a programming education. This education was also done in English over any Indian language; by the time IBM left the country, India had a large and Anglophone technical base.
Secondly, Indian protective policies meant that, wherever possible, value stayed inside the country. The Indian IT scene is unique in it being led mostly by local ventures rather than foreign MNCs. And the way the protective policies of the IT sector differed from the protective policies of, say, manufacturing, is very telling. Recall, for example, that from the earliest days of the industry, software companies were compelled to export software, meaning that rather than having to face a captive domestic market, they needed to be able to stand on their own legs in a competitive global market. If you are learned in how South Korea developed its manufacturing sector, this should strike you as remarkably similar to the Park strategy of "export discipline"; companies that could survive outside the domestic market were rewarded, and companies that couldn't were culled. Software Technology Parks were also established, allowing the concentration of the IT sector into areas where different laws applied. And, of course, the DoE/DoT spat above.
Thirdly is the rise of Unix. Programming used to depend entirely on the hardware; the mainframe OS was tied to the mainframe machine itself. You could not modify it without the machine. This is why the nature of early Indian IT work was body shopping. But by the 80s, open standards, the workstation, the rise of Unix as an operating system common to all computers, and the rise of C as a programming language also common, meant that code would be written anywhere, by any programmer, for any computer. This, along with the previously established increase in data links in India, meant that it was now possible to fully outsource software to the country. India just needed a massive, common bug that would be expensive to fully fix without having to outsource. That leads us to...
Y2K and the Golden Age
The Y2K bug now is known as mostly just a meme; an overly exaggerated issue that resulted in nothing happening much. But the decision of the COBOL developers to use two digits to encode a year was seen at the time as a looming disaster waiting to happen that needed urgent fixing. An urgent fixing costing an estimate of hundreds of billions of dollars. In this environment, there was a large appetite for making that estimate lower.
India entered the fray; the government and the industry as a whole allowed software companies to subcontract each other, as well as teaching programmers COBOL, and then launched a massive advertising spree to US businesses, advertising the fact that Indian labour was far cheaper and still capable of the daunting task of preventing Y2K. The sector prepared for it, with TCS building an entire office dedicated entirely to fixing Y2K code, a thousand programmers churning out millions of lines of code every single day.
This resulted in massive profits for the sector. 2.3 billion dollars were made in the last few years of the 90s, and Indian software companies became globally known in the field. Additionally, it helped reduce the incidence of body shopping, and by the 2000s onsite work would make up only a minority. Software exports exploded from 164 million dollars at the start of the 90s to 6.2 billion by the turn of the millennium, and expanded further to 100 billion dollars by 2013.
Infosys, one of the bigger IT companies, became the first Indian company to get listed on the NASDAQ. The sector as a whole is now responsible for 6 million jobs directly, contributing 8% of India’s entire GDP. At one moment, hard to remember in the current episode of the Sino-Indian enmity, Chinese government ministers were studying how Infosys developed software, visiting the company’s headquarters. Nevertheless, this is not a particularly stable industry, especially in the current situation, which brings us to..
Current
One of the bigger and earlier fears of the scene and of IT development in general was automation. If programming was to be automated, capital intensity would increase and labour intensity would decrease. This would hurt the Indian IT scene the worst due to its founding bargain being its labour intensity. Additionally, Indian IT is extremely bottom heavy; most employees are programmers, not analysts nor project leaders, unlike in more developed sectors where more employees are analysts than they are programmers. No upskilling happened. Even in the 80s, people were remarking that increasing code automation meant the age of the programmer was over; and for once, this contemporary prediction of the future isn’t something that one looks back at with laughter.
Growth rates of the sector began to slow throughout the 2010s and 2020s, and by 2024, the sector grew only by 3.8%, with incremental revenue declining from 19 billion in 2023 to 9 billion in 2024. Job growth slowed too; from 300k jobs added in 2023 to 60k jobs added a year later. Many companies posted declines; Infosys and Tata had their headcounts decline by 25k and 13k in 2024, the first decline on record and the first decline since 2005 respectively. Tata then had to lay off 12k employees in 2025, the largest layoff of the company. Share prices of both companies, and many IT companies in general, have seen massive decline since 2024.
Why is this the case? I will point to two reasons.
The first one is that since 2022, Large Language Models have exploded in popularity, and of the many things they can do, they are the best at code; specifically, the type of basic code that Indian programmers often do. I will remind you again that the Indian IT sector is very bottom heavy, and AI’s gains have been concentrated in that exact field. It doesn’t make any sense to contract someone when an AI can do the job better and cheaper.
In fact, it doesn’t make sense to contract, period. If one is going to hire a human, why not hire them directly? This leads me to my 2nd, and in my opinion bigger, factor in the decline of the Indian IT industry. The value of Indian software companies is as expensive middlemen, coordinating a large and bureaucratically complex workforce and telling them what to do exactly. As software capabilities have grown and as companies invest, many of them decided to just..hire directly. This is done through a Global Capability Centre, of which there are around 2 thousand, with 2.3 million Indians employed in such a manner. This work is often higher skilled, and it directly undercuts the business model of companies like Infosys.
In Conclusion…
India developed an IT industry because it needed a proper and decently high value export sector that did not require expensive, capital heavy investment, and which could rely on the sheer labour pool and its size.
But things change. Demand for software, for doing it cheaply, will always exist. But now Indian IT can’t just compete on its cheapness, it will need to figure out another way of competing. The sector as it is won’t last, but I think it will exist, just likely to be smaller, more top heavy, and upskilled. It’s too big to fade away.
But I would still wish to talk about the nature of how the Indian IT sector became a thing. No single government actor did everything described above; rather it was a sort of mess of institutions fighting each other and getting their own way. And when you read about India you will notice that this pattern isn’t limited to the development of IT. I think this is a kind of…asynchronous developmentalism.
And that term will get its own article.
This article is part I of a two-part series on India’s IT development and asynchronous developmentalism. read Part II here.
This article was originally a script that I wrote, commissioned by Asianometry. I will link the video here when it comes out.