India's Semiconductor Ambition: The CHIPS Act Race and Why Manufacturing Is Harder Than Policy

In 2021, India launched the India Semiconductor Mission — a $10 billion production-linked incentive scheme designed to plant fabrication and assembly facilities on Indian soil. By mid-2024, the first real projects had broken ground. By 2026, the early returns are in, and they tell a mixed story.

The headline: India is building semiconductor infrastructure, and it's real. The fine print: what India is building isn't what the semiconductor shortage conversation implied it would be.

**The design advantage India already has**

India's semiconductor story usually gets told backwards. The narrative starts with fabrication — TSMC, fabs, wafers — and treats everything India currently does as insufficient.

That's wrong. India already has a significant and growing position in chip *design*. Qualcomm, Intel, AMD, Nvidia, and most of the major US chipmakers have large engineering centres in Hyderabad, Bengaluru, and Pune. Roughly 20% of the world's chip designers work in India. This isn't outsourced implementation work — it's core architecture, RTL design, and verification on leading-edge products.

India doesn't manufacture chips. But Indians design a large fraction of the chips that get manufactured elsewhere.

**What ISM is actually building**

The India Semiconductor Mission's approved projects as of 2025 include:
- **Tata Electronics–Powerchip fab** in Dholera, Gujarat (28nm legacy node). A full fabrication facility — India's first — targeting $11 billion in investment.
- **Micron ATMP facility** in Sanand, Gujarat (assembly, testing, marking, and packaging). Opened 2024.
- **CG Power ATMP** in Sanand, in partnership with Japan's Renesas.

The common thread: ATMP facilities and legacy-node fabrication. Nobody is building a 3nm plant in India. The Tata-Powerchip fab is targeting 28nm — the same node TSMC was running in 2011.

**Why that's not the failure it sounds like**

The semiconductor supply chain isn't just about leading-edge nodes. About 70% of the chips in any given car, home appliance, or industrial device are made on mature nodes — 28nm, 40nm, 130nm. These are also the chips that ran short during COVID-19, not primarily the bleeding-edge logic chips.

India entering legacy-node fabrication is a legitimate strategic play for automotive, defense, and consumer electronics. It reduces import dependency in categories where India currently produces zero chips domestically. The geopolitical logic is sound even if the technology isn't frontier.

**The harder problems**

Where India faces genuine structural challenges:

*Water.* Semiconductor fabrication is extraordinarily water-intensive — a fab at scale can consume millions of gallons per day for ultrapure water processes. India's water infrastructure in the areas targeted for fab development isn't built for that. Dholera is in Gujarat's coastal region, and water logistics is a real constraint that doesn't get enough attention in policy discussions.

*Supply chain depth.* Running a fab requires specialty chemicals, high-purity gases, photomasks, advanced metrology equipment — almost none of which India currently produces. The US and European CHIPS Acts are building these ecosystems over decades. India is starting from near zero.

*Workforce pipeline.* India has chip designers. It doesn't have a large pool of fab process engineers, equipment technicians, or cleanroom operators. These skills are different from design, and they're developed through hands-on experience. There's no shortcut.

**What 2030 looks like realistically**

India's semiconductor ambition for 2030 is a $100 billion ecosystem. That number includes design revenues, ATMP revenues, fab production, and equipment manufacturing — categories at very different stages of readiness.

The design piece is already close to that scale by some measures. The ATMP facilities will be operational and generating real output. The Dholera fab will be running, probably at below capacity.

What India won't have by 2030 is leading-edge fabrication. TSMC's 3nm process in Arizona took years longer and cost more than projected. India's first fab, starting later and targeting legacy nodes, won't close that gap in four years.

The honest version of India's semiconductor story is: it's building real infrastructure in the right places for the right strategic reasons, just not the infrastructure that makes headlines at TSMC investor days.

India's Semiconductor Ambition: The CHIPS Act Race and Why Manufacturing Is Harder Than Policy

// COMMENTS

ON THIS PAGE