Start building in India

Traffic lights, phones, washing machines and supercomputers all run using microprocessors. Microprocessors are electronic chips that are at the heart of most electronic systems found today. They perform the computations needed to enable smart traffic signals, WhatsApp calls, wash-rinse-repeat cycles and weather prediction.

By Rakesh Kumar
  • Published 1.02.18
  •  

Traffic lights, phones, washing machines and supercomputers all run using microprocessors. Microprocessors are electronic chips that are at the heart of most electronic systems found today. They perform the computations needed to enable smart traffic signals, WhatsApp calls, wash-rinse-repeat cycles and weather prediction.

Considering how critical microprocessors are to today's applications, one would imagine that India builds most, if not all, of its microprocessors. A casual observer, therefore, is often surprised to learn that there are no microprocessor chips being designed or fabricated in India today. India, directly or indirectly, imports all the microprocessors it needs.

This can be attributed to several factors. First, building microprocessors, especially high-performance microprocessors, is a prohibitively capital-intensive business. A single high-end microprocessor today can easily cost hundreds of millions of dollars to design (more than $300 million for a recent Intel processor, for example). It also requires fabrication facilities that can cost over $10 billion (TSMC, the world's largest semiconductor fabrication company, estimates the cost of a 3nm manufacturing plant to be $20 billion). Second, the expenses are not one time, but recurring, because of frequent technology updates (TSMC invests roughly $10 billion every year to keep its manufacturing up-to-date). This requires extreme doggedness and discourages new participation. Third, microprocessor intellectual property is heavily export-controlled. Developing non-infringing competing technologies is not an easy task.

In spite of this, India must start investing in building microprocessors. These, especially high-performance ones, increasingly drive some of the nation's most strategic investments in defence, industry and the sciences. Importing microprocessors to support these strategic investments makes them susceptible to export control, trade embargoes and security trojans. This should be of concern to India. The United States of America had famously restricted India's access to critical high-performance computing technologies in the late 1980s and early 1990s because of nuclear proliferation concerns. The US and the European Union had imposed similar sanctions after the Pokhran test. As India's strategic and military ambitions intersect with those of other world powers, it is not difficult to foresee scenarios where microprocessor producing nations attempt to harm Indian interests by using access to microprocessors as leverage. In 2015, the Barack Obama administration had debarred the sale of high-end microprocessors to the Chinese government arguably to slow down China's march. India must guard against such possibilities through indigenous development of microprocessors.

Imported microprocessors also come with security risks engineered by State actors. The Edward Snowden leaks showed how American and British intelligence agencies colluded with microprocessor companies to weaken the security of their processors. India's national security as well as its citizens' privacy are at risk when foreign processors are used to drive the nation's critical applications.

Even from a purely economic angle, there is a strong argument for reducing dependence on foreign processors. India's chip imports are second only to oil. The country pays firstworld prices on imported microprocessors since there is no alternative. An indigenous microprocessor may, at the least, bring down prices of the imports owing to increased bargaining power. In an optimistic scenario, it may serve the needs of the Indian market.

How should India approach building microprocessors? First, the effort must be driven by the government, not private industry. Not unlike India's investments in space, atomic energy or light combat aircrafts, indigenous development of microprocessors would require the level of persistence and capital that only the government can provide. Second, political consensus needs to be built around the effort; even a single microprocessor's timeline may span multiple governments. Third, India should prioritize microprocessor design over fabrication. Microprocessor fabrication is at least one order of magnitude more expensive than design. Also, India already has a relatively large hardware design talent pool that can be leveraged. Furthermore, careful design can be used to obviate security and privacy concerns even if manufacturing is untrustworthy - so, prioritizing design allows security risks to be managed even if the designs are manufactured elsewhere.

Care must be taken to reduce the financial burden of the microprocessor building effort. Public-private partnerships should be used once a project's risks have been managed to an acceptable degree. Microprocessor customizations should be used to lure large private players (for example, a Flipkart). A simpler microprocessor should be used at scale (through preferential access), both inside India as well as in regions in need for cheaper information technology equipment (for example, Africa), to reduce the new capital needed for the next, more complex, microprocessor. Market size should be leveraged to extract favourable technology transfer deals to accelerate the microprocessor development process.