Audio & Story | India’s semi-plan for semiconductors
The Union Cabinet has cleared a subsidy of Rs 76,000 crore for indigenous production of semiconductors. The Production-Linked Incentive (PLI) Scheme for Large Scale Electronics Manufacturing, PLI for IT Hardware, SPECS Scheme and Modified Electronics Manufacturing Clusters (EMC 2.0) Scheme and other ladles of alphabet soup together commit some Rs 230,000 crore. Sounds great, raises visions of India exporting semiconductors instead of importing these things in volumes that top the country’s oil import bill, and, therefore, constitutes a major policy success, right? Wrong.
This entire scheme lacks a core strategic vision. It would be a waste of scarce resources without catalysing self-reliance where it matters.
But isn’t self-reliance a relic of the Nehruvian past, that we are best rid of? Isn’t it more important to create enough value to be able to pay for anything we need to import, specialise in the production of things in which we have comparative advantage and maximise exports, jobs and income? Unfortunately, in the real world all such subsidy magic and policy enchantments do not produce happy endings in which the main protagonists live happily ever after.
Take just one example. Bill Clinton took the decision to make access to US Armed Forces’ geo-positioning system free for all. The world at large had access to GPS. Yet, the Russians built their own Glonass, the Chinese their Beidou satellite navigation system with 35 satellites, the Europeans their own Galileo system and India, its Navik facility. Why?
There could be contingencies in which the Americans suddenly shut down access to their GPS. Suppose that happens when a Chinese missile is vending its way to a distant target using GPS for guidance; that missile would lose its way more thoroughly than Hansel and Gretel had in the forest.
Also read: The Glasgow climate opportunity for India
India is right now in America’s good books and has no reasons to think that GPS access would not be forthcoming for any Indian use. But things could change, for unforeseen reasons. India cannot afford to be at the mercy of any other power for satellite navigation. Hence, Navik.
This is even more true for advanced silicon. Semiconductors underpin strategic weapons systems, communications, the financial infrastructure, even the opening and shutting of the sluice gates of dams. Suppose complex microchips have built-in backdoors that allow a foreign agency to gain control or snoop. Suppose spares are withheld when they are desperately needed. A country like India that aspires for strategic autonomy cannot afford not to be self-reliant for vital bits in microelectronics.
But isn’t this paranoia of the comic-book variety? Well, paranoia can look funny, admittedly. In the face of persistent concern that Huawei’s telecom gear that came cheap, and was displacing supplies from the likes of Ericsson and Nokia-Siemens, could be compromised, it was decided to test Huawei’s kit for any discreet snooping capabilities. This was during the UPA’s second term. The task of examining Huawei’s telecom kit for vulnerabilities was entrusted to the Indian Institute of Science. IISc, it turned out, did not have the in-house capacity to carry out such inspection — telecom equipment comprises complex bits of software and hardware in a tight embrace, too intricate to disentangle, leave alone yield their separate secrets. So IISc invited external assistance to help with their task. Guess, who turned out to be the most technically qualified and least expensive vendor for the technological assistance asked for? That’s right, it was Huawei itself. The government intervened to quash that deal.
To get all serious, one just needs to look at Huawei’s own fate. Till President Trump came along and barred Huawei as a customer for advanced American-made silicon, most people did not think that microchips would be a source of strategic control. Now we know better.
If India wants to pursue a truly independent path in foreign policy, it cannot depend on external suppliers for sophisticated microchips. It has to build its own chips. It must replicate the chip architecture design business of London-based ARM and Intel, design, using that architecture and add-on elements, the actual processor, like AMD, Nvidia, Qualcom, Apple or Samsung and then manufacture the chip based on the design, as Taiwan Semiconductor Manufacturing Corporation (TSMC), Intel, Global Foundries and Samsung do. In the process, it would need to build sophisticated machines that can engrave nanometre-thin grooves on silicon chips and flow vaporised metal into those grooves to create the tiny circuits that go into the chips.
This level of integration is indeed a tall order, achieved by no country among its own firms, although the US could do it and dragoon some European firms into its service as well. China is on a path to develop its own microprocessor ecosystem from scratch, thanks to the technology sanctions initiated by President Trump that still continue.
Is this impossible for India, with its none-too-scintillating performance on hardware, original software or manufacturing competence? Not at all.
India has had its own brief moments of innovation glory. One is the creation of the Rural Automated Exchange developed by C-DoT, when it was headed by Sam Pitroda, during the Rajiv Gandhi government. These RAXes could withstand extreme heat and cold and the dust and the damp of rural India. They thrived in the Australian outback, as well.
During the subsequent VP Singh tenure, C-DoT was destroyed by two Malayalis, KP Unnikrishnan as the minister in charge and KPP Nambiar as Secretary to the Government. They foisted a collaboration with Alcatel on C-DoT, killing indigenous research and development and taking out what was promising to be a formidable third world developer of telecom kit.
India’s space and atomic programmes continue as shining examples of technological excellence under trying circumstances, of prolonged technology denial by foreigners till the Indo-US nuclear deal of 2008, salaries that are meagre by industry standards and the omnipresent, omniscient bureaucracy that stands between them and the political leadership.
The goal should not, primarily, be to induce extant chip designers and fabricators to set up shop in India, luring them with subsidies. These companies are vulnerable to US sanctions and could cut India off, in none-too-farfetched situations such as a Congressional majority imposing, guided solely by a partisan desire to spite the incumbent president, sanctions on all countries which buy arms from Russia or, indeed, whose name begins with the letter I. The goal should be to promote indigenous capability in all aspects of chip design and manufacture.
It need not happen all at once. ARM’s core can be licensed and modified, to begin with, as Samsung does. Partners like IBM can be roped in to design chips. After all, a whole lot of design work in any company outside Korea and Japan would involve a great many engineers of Indian origin. Some of them can be roped in to work on Project Indigenous Semiconductor Manufacture or Prism for short.
Some of them can be contained in fully funded public sector units. Others can be in startups that are given seed capital by the government.
For this to happen, the government will have to rethink its policy of giving fat subsidies to companies that already have the know-how and the capital to invest in facilities in India.
If India’s increasingly disliked and derided first prime minister had simply offered to fund foreign makers of satellite launch vehicles and nuclear reactors if they would kindly set up shop in India, instead of creating our own agencies to develop them on their own, India would today have neither rockets, except on Diwali, nor nuclear weapons.