Semiconductors in India

The arena of confrontation between the US and China that is most visible is Taiwan. Another arena is semiconductor chips and America’s increasing its steps to prevent chips getting to their foes. It started with Russia over the Ukraine war, and then the US decided to ban certain specialized chips to China, and now the US banned chips key to AI from the Middle East.

 

The other aspect of the chip wars is that the US is trying to de-risk itself from its dependency on chip manufacturing being in China, or too close to China for comfort (aka Taiwan). Hence the urgency to move some of that chip manufacturing to the US itself. It is in this context that countries like India have seen an opportunity – after all, some fraction of the manufacturing moving out of China could come to India. If India plays its cards right.

~~

In that context, Pranay Kotashane cautions that the role of the government in semiconductors historically has been at most as an enabler rather than the one running the show. The Indian government, he says, should follow the same course – facilitate things (taxes, land, labour etc) but not pick specific companies as the beneficiaries.

 

Why? First, he says, semiconductor manufacturing requires huge money initially. If a government puts in that money, it gets tied to the success of the companies it invested in. It will then inevitably try to prop up those companies by pouring in more money or changing laws to benefit them. That is a terrible idea (remember PSU’s?). Instead, let the private sector put in the money and take the risk and reward, he says.

 

Second, the semiconductor industry can change suddenly. It can create new winners suddenly and decimate the old guard swiftly. Such outcomes are best left to the market to cope with, and governments hate the political fallout of investments gone bad. Best then they stay out of the industry. Other than facilitating things in general for the entire industry, not for specific companies.

 

The third reason comes in a different article where Kotashane explains why Moore’s law (the famous prediction, not “law”, that the number of transistors in an IC would double every two years) keeps working even after so many decades – Gordon Moore made his prediction in the 1960’s!

“That's the reason that the retail prices of electronic products fall rapidly even as newer products become faster and better.”

 

How come Moore’s law keeps on holding true? Advances in semiconductor manufacturing often require huge money to be invested in upgrading facilities. Shouldn’t that have nullified the benefit of doubling-of-transistors-in-an-IC? And how come people kept finding newer ways to miniaturize further and further?

 

The answer?

“That Moore's prediction became a law is a testimony to human ingenuity and decentralised innovation.”

The key point is that it is not only due to technological improvements but also due to new business models. Like decoupling chip design from chip manufacturing. Specialists in manufacturing alone could focus on one area (manufacturing, not design) and also get the benefits of scale (if they manufactured for more and more companies, they could optimize practices).

 

Periodic changes in business models are very disruptive – they create new winners and wipe out the existing ones. Yet another reason why the government shouldn’t be picking companies to bet on.

~~

If the Indian government gets it right, the semiconductor manufacturing industry could turn out to be a major benefit for the country. It will create a lot of low end jobs, and it will lead to learning the basics. Over time, Indians will get to higher end activities, including chip design itself. And everything in the modern world runs on those chips. So one hopes the government gets it right.