Illustration picture of semiconductor chips on a circuit board
Credit: Reuters Photo
Across the world, countries are spending wagonloads of taxpayer money to develop or support their semiconductor industries. This pipeline of cash has yielded some successes and some failures; the difference is usually down to how strategically the money has been spent. The Indian government, which has also loosened its purse strings, needs to figure out a strategy — and soon.
Under Prime Minister Narendra Modi, federal budgets have generally been pretty stingy with subsidies. But $11 billion has already been set aside, and largely allocated, to various chip-related projects. News emerged last week that another $5 billion to $10 billion might follow.
These aren’t small numbers by Indian standards. The government might spend five times more on this sector than it does on support for solar panels or autos — and 50 times more than it plans for labor-intensive sectors such as leather goods. According to officials, when various funds from state and federal sources are added up, three-fourths of the capital cost of a new semiconductor manufacturing facility will be borne by the government.
Yet it’s not clear officials know exactly what they want from their semiconductor mission. They seem to have multiple objectives, and that always leads to a bit of mess when it comes to designing policy.
Is the goal to increase the amount of value added in India? Or to create jobs? Is the government’s primary concern reducing India’s dependence on China, or substituting imports overall? Do officials think domestic production is essential to support the creation of a broader electronics sector, or to provide supply-chain resilience to the growing automobile sector?
What they say in public is troubling. The emphasis seems to be on onshoring production and promoting what Modi calls “aatma-nirbharta,” or self-reliance.
That is hardly a sensible objective. Not even the US can really aspire to true semiconductor self-sufficiency, given the complexity of chip supply chains. There’s no reason to think India can, especially starting from scratch.
Nor should the nation worry too much about “strategic dependence” on chip imports, as long as a significant proportion come from friendly sources such as South Korea, Taiwan, or the West. Reducing China’s share in Indian imports might make sense, but semiconductors aren’t a sector where this should be a concern. A large proportion of China’s future production will go to satisfy its own demand for chips.
Instead, the focus should be on identifying the spots in the supply chain where India can enter quickly and scale up effectively. Fabrication is glamorous, but as multiple countries across the world have discovered, the startup costs can be ruinous.
Requirements are stringent as well. Some companies interested in manufacturing in India may have given up because it seemed too tough to secure a reliable supply of ultra-pure water. The private sector needs to be given a freer hand in identifying the right locations for its investments. Most new plants have been steered toward states ruled by Modi’s party, including his native Gujarat.
Officials should focus on addressing some of these questions first. They also need to work harder on an unexpected problem: getting the right workers. India may be the world’s most populous country now and claims to have thousands of engineering colleges. But quality control in higher education is so poor that new plants will struggle to recruit a workforce.
Officials are at least aware that one of the biggest bottlenecks is the availability of engineers. If their plans are to come to fruition, they estimate they need between 10,000 and 13,000 trained workers by 2027. Some private forecasts of the talent shortfall are much higher.
Restructuring academia to focus on research and training will pay dividends. India should aim to staff not just its own plants, but also to fill manpower requirements in the rest of the world.
Fabrication isn’t the only place where money is made in the chip supply chain. Design, assembly and packaging together represent about half of the sector’s revenue. India already has a fifth of the world’s chip designers, but only 7% of actual design facilities.
New assembly plants are coming up and government support for design has been increased. But more could be done if New Delhi spends its money rationally. The real bang for its buck will come through investing in those sections of the supply chain where India may have a comparative advantage.
Fabs can follow once an ecosystem develops. There’s no point obsessing over self-reliance when even a giant like the US is only one cog in a global machine.