Routinised technologies, such as those characteristic of the 'older' steel and chemical industries, were embodied in large continuous process plants requiring 'lumpy' investments in innovation, commercialisation, capacity creation and market acquisition. This made access to a critical size of capital crucial for entry, queering the pitch for the big players from the developed industrial countries. Moreover, dramatic innovations in these sectors were few and far between, with much technological development consisting of marginal changes that were cumulatively, rather than instantaneously , significant. These marginal changes were, in turn, very often stimulated by knowledge gained in the act of production, which led up to expensive R&D exercises that created new commercially-usable knowledge. Thus not only were these industries dominated by big firms created with lumpy investments, but much of the technological change that occurred in the industry originated within or as a result of the activitiy of these big firms themselves. Entry into the industry by new, especially small new, players was rare. Competition was restricted to that between the dominant oligopolistic firms straddling domestic and world markets. To the extent that technological change could trigger a competitive challenge from outside the industry, this was largely the result of the emergence of substitute products or wholly new industries that rendered the older industries less significant or even irrelevant.
 
As compared with this the IT sector is seen as characterised by low costs of entry and an easily accessed and almost universally available knowledge-base for innovation. What is of special significance is that the sources of this knowledge in a significant segment of the hardware and almost all of the software segments of the industry are conventional routes such as journals, conferences, seminars and publicly or privately financed training programmes. This makes it easy for wholly new entrants to acquire the knowledge base required for cutting edge technological contributions to the industry, as was and is true of at least some of the myriad start-ups in Silicon Valley. It is also true, that the production of a range of products varying from components like printed circuit boards to peripherals like modems, on the one hand, and the assembly of personal computers, on the other, do not require large investments for entry. These, unlike other IT products, are heterogenous products that can be put together on the basis of a combination of internally produced and externally sourced intermediates and components. The real 'technology' here is a system architecture that maximises the benefit derived from an appropriate combination of sub-systems, components and peripherals. That is, the heterogenous nature of the product allows a producer to restrict his own production activity to a sub-set of the total elements that enter the product and/or just the design and assembly of the final product. This requires that the technology in the sense of the knowledge for system design is freely available and is easily appropriated. But once such technology is available, the investment required for entry can be kept to the minimum, with labour intensive assembly often involving negligible investment in capital equipment. This allows for the emergence of international brands from the developing world, as is true of Acer from Taiwan.
 
The logic of easy entry is even more true of the software sector, which is skilled-labour intensive and requires little by way of capital investment. Not surprisingly, over the years the software segment of the IT sector has come to dominate the industry in India. As Chart 3, which details the structure of the Indian IT industry in 1999-00 shows, software production accounts for two-thirds of the turnover of the industry. However, with software exports accounting for 45 per cent of the turnover, it should be clear that hardware sales marginally dominate the domestic market. But with software exports growing at a much faster rate (57 per cent in 1999-00) than domestic IT expenditure (30 per cent), it should be obvious that the industry would soon be almost overwhelmed by software production. And, software is the segment where the advantages of easy technology access and low entry-level investments are most prevalent.
Chart 3 >>
 
As mentioned earlier, in the past, these special characteristics of the IT sector, which substantially reduce technological and financial barriers to entry by small players, were seen as underlying the success of small Silicon Valley start-ups. Those start-ups not only challenged traditional giants like IBM, but have since grown to become major players in their own right and have changed the structure of the industry. The growth of the IT industry in India and elsewhere in the developing world and India's success as a software exporter, seem to suggest that these characteristics hold for firms in developing countries as well. Not surprisingly, it is now being argued that what was true for the Silicon Valley start-ups in terms of their ability to break through barriers to entry should be true for the developing countries. It is this perception that underlies the optimism that information technology heralds a new era of reduced international inequality.
 
It should be obvious that this argument cannot be extended beyond a point, especially in the hardware sector. The core of the computing business is dominated by a capital-intensive and oligpolised product like the microprocessor, the market for which is dominated by a few producers like Intel, Motorola and AMD. And the technologies driving a range of peripherals like printers and networking products, for example, are proprietory and are not replicated without a licence. Further, while capital investment requirements for production may be small, production for geographically and quantitatively large national and world markets require high sunk costs. This takes the form of initial expenditures on marketing, retailing and the creation of an after-sales service network. Deep pockets and/or access to large sums of capital are therefore a prerequisite for entry into this segment of the hardware sector.