Commentary: Embrace chips, India, and skip those tech dreams

NEW DELHI: It is a truth universally acknowledged that a nation of reasonable size and prosperity must be in want of a semiconductor industry. India is no exception to the global rush to subsidise chip fabrication; the government set aside US$10 billion for new facilities in 2022 and it has plans to double that amount.

There are expectations that money will provide a return soon. Last week, Prime Minister Narendra Modi promised that the country would soon have a “significant share” in a market worth over a trillion dollars: “Our journey began late but nothing can stop us now. India is now moving from the back-end to becoming a full-stack semiconductor nation.”

There have certainly been some successes. Modi was presented with the first “made in India” chips after his speech. But the government’s ambition should be tempered with reality – particularly when it’s spending billions that the harassed finance ministry can barely spare from other priorities.

The chips shown to Modi give us a clue as to what might work swiftly. They were developed by a special laboratory that’s part of India’s space programme, and designed for the harsh conditions of space — or, not incidentally, for missiles.

They’re cheap, functional and robust. But they aren’t cutting-edge: They’re based on 180-nanometre technologies that are decades old, the same sort that were used in Intel’s Pentium-IIIs at the turn of the century. (For context, TSMC is expected to go into mass production for 2nm chips later this year.)

That isn’t a criticism. If these get the job done efficiently, then it’s safer and more effective to use a mature, proven technology.

Nor does it mean India lacks the capability to produce more advanced microprocessors. Last month, the minister for information technology said that 28nm chips would begin to be produced next year at CG Power & Industrial Solutions’s factory in Gujarat. Nine other new plants, with Japanese, Taiwanese, American or Israeli partners, are in various stages of completion.

But as the government pulls together more money for the next stage of its semiconductor mission, it should be cautious. The 28nm chips aren’t exactly new either: TSMC built its first dedicated foundry using that technology in 2011. But they’re more than modern enough to meet most requirements.

PROFIT ISN’T THE PURPOSE HERE

Advanced processors – which US export controls described in 2022 as being those “smaller” than 16nm – are used in AI data centres, mobile phones and a few other such applications. They’re where the big profits are for most global chipmakers.

But profit isn’t the purpose of India’s semiconductor mission. It’s to ensure that other parts of the country’s manufacturing base – automobiles, household electronics – don’t suffer if the supply of semiconductors is suddenly cut off. 

The government is spending billions to bolster economic security, not to earn dividends from cutting-edge products that, so far, only East Asian companies seem capable of producing at scale.

MASTERING LEGACY CHIPS

It’s worrying, therefore, that officials are claiming that “we already have a clear road map to enable companies to manufacture 7nm and above chips within the next five to seven years.” Even if this timeline is possible, it may not be the right choice. 

There is a space on the global supply chain that India can and will fill: Mature designs that provide security and support a wider manufacturing ecosystem. The government shouldn’t let high-end dreams divert its attention or resources from what’s actually needed.

The private sector has already spoken on this subject. Attempts to lure the most advanced chipmakers – TSMC or Intel – to India have not worked. 

Partly that’s because the government has less cash to hand out, and its subsidy system is designed to share risk with companies and not minimise it. But it’s also because there are a limited number of top-of-the-line plants that can be built at any time, and C-suites think they’re better located elsewhere.

If India wants to move up the value chain, it needs to show it has mastered the basics. There’s hard work to be done first: Demonstrating the scale at which legacy chips can be made here, building up an experienced and reliable workforce, developing the innovation, design, and testing hubs that will attract the pickiest manufacturers. That’s honest work, and if it succeeds, it will be enough to be proud of. 

India doesn’t need to stop dreaming about high-end semiconductors. It just shouldn’t chase them.