India’s $10 Billion Semiconductor Plan | Neelkanth Mishra Explains

ELI5/TLDR

India had zero semiconductor manufacturing and decided to buy its way in with $10 billion in subsidies, covering 50-70% of the capital cost for anyone willing to build a chip plant on Indian soil. Neelkanth Mishra, a financial advisor to the mission, walks through the messy reality of getting from PowerPoint presentations to actual silicon — the embarrassing early pitches, the Micron breakthrough, and why India is roughly where China was 25 years ago. The punchline: this is a 20-year game, the first budget is already spent across 12-13 projects, and the second round is about to get more ambitious.

The Full Story

India Had a Head Start and Blew It

In the early 1980s, India’s semiconductor facility in Mohali was ahead of Taiwan. Then nothing happened for decades. Two revival attempts — around 2007 and 2016 — both fizzled, cratering India’s credibility with global chip firms. The third attempt, launched in December 2021 with the India Semiconductor Mission, had one crucial difference: the Prime Minister personally committed.

The $10 billion budget (76,000 crores) came with a simple structure. The advisory board initially planned tiered subsidies — 40% for advanced chips, 30% for less advanced. In their first meeting, they scrapped the complexity.

“Right now we have nothing. Let’s keep it very simple. Let’s increase it to 50% capex subsidy.”

State governments topped up another 20-30%. Meaning an entrepreneur could own a whole fab by putting in just 20-30% of the capital.

The Semiconductor Value Chain is Absurdly Concentrated

Mishra maps the industry like a network of tiny oligopolies. 50-70 blocks of expertise, each dominated by two or three companies. ASML makes the only extreme ultraviolet lithography machines. The lenses inside them are single crystals grown over months by Zeiss. Two companies — Teradyne and Advantest — control 90% of testing equipment. In memory, it’s Samsung, SK Hynix, and Micron.

The numbers on precision are staggering. Silicon ingots require 99.9999999% purity — nine nines — because you’re stitching together individual atoms. TSMC’s advanced clean rooms circulate air 90 times per minute. A cutting-edge chip now uses more than 50 elements from the periodic table, up from 15 just two decades ago.

And here’s the uncomfortable financial truth: across decades, only three companies globally — Samsung, TSMC, and (formerly) Intel — have covered their cost of capital in chip manufacturing. Mishra’s initial reaction as an analyst was blunt: why are we wasting money on this? He was told to stop thinking like a financial analyst and start thinking like a country.

The Micron Breakthrough

The early days were grim. Proposals came in from companies with no team, no clear funding, and foreign partners putting in zero capital for a 5% equity stake. One foreign partner lectured the advisory board for 15 minutes about beggars not being choosers.

Then Micron showed interest in a packaging plant. Mishra recalls hearing the news while taking his kid to the dentist.

“We can’t wait for the right time before they change their mind. Let’s get them through the door.”

Micron was the ideal first customer. A solid US-listed company — no bankruptcy risk, no gold-plating risk, built-in demand (70% in-house, 30% outsourced, with outsourcing cut first in downturns). The credibility boost was enormous. But even then, it took 9 months just to set up an Advanced Pricing Agreement because no framework existed for taxing a semiconductor company in India.

Where Things Stand Now

The first budget is fully allocated across 12-13 projects. The big players: Tata (wafer fab and packaging, partnered with Taiwan’s PSMC), CG Power, Kaynes Technology, and Micron. CG Power has already produced India’s first packaged chip. Micron’s test batches are out, with mass production imminent.

India’s one wafer fab (Tata/PSMC) is targeting 40 nanometers by 2028 and 28 nanometers by 2030. For context, the global cutting edge is 2-3 nanometers. India is starting roughly 20 years behind the curve — but you can’t skip the apprenticeship.

The second round of funding (ISM 2) is in the pipeline. Mishra wants it to target the 7-12 nanometer range for projects coming online around 2031-32. The quality of incoming proposals has improved dramatically — bigger balance sheets, better technology partners, more serious teams.

The Geopolitical Tailwind

US-China decoupling is quietly subsidizing India’s ambitions. American EDA and equipment companies, barred from selling to China, have lost billions in revenue. They’re terrified that Chinese competitors will sprout up in their absence. So they’re offering massive licensing discounts to Indian startups.

“If you’re in India, till the time that you’re generating revenue, I’ll give you massive discount on your licensing.”

Equipment companies face the same calculus. The result: what costs $250 million globally in chip design might cost $100 million in India — cheap engineers plus desperate foreign suppliers equals a genuine cost advantage.

The Chip Design Gap

India has 300,000 chip design engineers — one-fifth of the world’s total. But effectively zero chip design companies of any size. Those engineers all work for multinationals like Intel (30,000 employees in India alone). The barrier isn’t talent. It’s the ecosystem: expensive EDA tools, years-long design cycles, and the need to predict what phone companies will want 3-4 years out.

The PLI scheme for mobile assembly is helping here — once you have device manufacturers on the ground, chip designers can see what’s coming. Indian startups are already working on GPU-competitive designs, with tape-outs planned for 2028-29.

The 20-Year Bet

The goal is 12-15 large plants. Once that critical mass exists, the government can step back and let the ecosystem self-sustain — local suppliers for clean water, gases, chemicals, precision equipment. The unexpected upside: a clean room in Kalaburagi cost one-fifth what it costs in Arizona.

Mishra insists India shouldn’t chase self-sufficiency. The semiconductor industry works because it’s distributed across oligopolies in different countries. India’s job is to carve out a few of those 50-60 specialization blocks and become indispensable in them.

Claude’s Take

This is a genuinely informative insider account. Mishra has the rare combination of financial analyst skepticism and policy-maker pragmatism — he openly admits he initially thought the whole thing was a waste of money, which makes his subsequent enthusiasm more credible. The “beggars can’t be choosers” anecdote and the dentist-office Micron story give you the texture that policy discussions usually lack.

The honest framing is refreshing. No one is pretending India will compete with TSMC at 2 nanometers anytime soon. The argument is simpler: you need to be in the game to eventually spot opportunities, the way TSMC spotted the IDM-to-fabless transition. That’s a reasonable bet.

What’s missing: any discussion of the risks. What if the subsidized companies just burn through capital and produce uncompetitive chips? What if the geopolitical tailwinds reverse? What if China’s compound semiconductor push collapses pricing for everyone? Mishra mentions the 20-year timeline but doesn’t wrestle with what happens if political will evaporates after the next election cycle.

The conversation could also use more on the environmental and water costs of fab construction — a real concern in water-stressed India.

claude_score: 7 — Good. A well-informed insider giving a clear-eyed tour of a complex industrial policy story. Points for honesty and specificity. Loses a point for being one-sided (no serious pushback from the interviewer) and another for some repetition in the transcript.

India's $10 Billion Semiconductor Plan | Neelkanth Mishra Explains | SparX