The announcement of TATA Electronics’ ₹91,000-crore semiconductor fabrication facility at Dholera, Gujarat, marks a decisive moment in India’s industrial and technological trajectory.

On 16 May 2026, Prime Minister Narendra Modi witnessed the signing of a strategic partnership between TATA Electronics and ASML at the latter’s headquarters in Veldhoven, Netherlands.

ASML, the sole manufacturer of extreme-ultraviolet lithography machines, committed to supplying lithography systems, technical training and ecosystem support for India’s first commercial fab, which is being built in collaboration with Taiwan’s Powerchip Semiconductor Manufacturing Corporation. The facility is expected to produce its first sell able wafers by late 2026.

The Dholera fab will not chase the cutting edge of 3nm or 2nm technology. Instead, it will focus on 28nm chips, with additional capacity for 40, 55, 90 and 110nm nodes. These mature nodes are the backbone of automotive electronics, power management systems, telecom radios, industrial sensors and embedded silicon.

Roughly 70 per cent of global wafer volume sits at 28nm and above, while only 3 per cent lies at 3nm and below. India’s choice reflects pragmatism: it is building for the chips its economy actually consumes, not the ones that dominate headlines.

The criticism that India should aim directly for 3nm or 7nm overlooks the historical record. Every successful semiconductor nation began at the trailing edge of its era and climbed gradually. TSMC started at 3 microns in 1987 and took 35 years to reach 3nm.

SMIC began at 180nm in 2000 and only reached 7nm-class production in 2023 after massive state subsidies. Conversely, Singapore’s Chartered Semiconductor and Malaysia’s Silterra, which either chased the leading edge prematurely or stagnated at the trailing edge without anchors, failed to sustain competitiveness. India’s earlier attempts — from the SCL fab at Mohali to the SemIndia and Jaypee–IBM ventures — collapsed due to lack of market demand or financial closure.

Today, however, the market exists, driven by smartphones, EVs, defence electronics and geopolitical decoupling from China.

The economic rationale is clear. A 28nm fab costs between $10–15 billion to establish, compared to $20 billion for 3nm and $28 billion for 2nm. ASML’s High-NA EUV scanners, essential for sub-7nm production, cost $380 million each, weigh as much as two Boeing 777s and take six months to install.

Beyond affordability, the question is demand. Renesas launched its flagship automotive microcontroller at 28nm in March 2026. The US Bureau of Industry and Security found that 80 per cent of chips in modern cars sit between 28 and 180nm.

A single passenger vehicle contains between 1,400 and 3,000 chips, and the 2021 global auto crisis was caused by shortages of 28–90nm microcontrollers, not 3nm processors. Maruti Suzuki alone cut production by 51 per cent in September 2021, costing the Indian auto industry around ₹25,000 crore in lost sales.

India’s ecosystem is being assembled in parallel. TATA’s OSAT facility at Jagiroad, Micron’s Sanand plant inaugurated in February 2026, CG Power–Renesas at Sanand and Kaynes Semicon’s operations since March 2026 provide the back-end support for packaging, testing and integration.

By the time Dholera’s wafers roll out, India will have a complete supply chain to absorb them. This marks a sharp departure from earlier failures where fabs were proposed without a domestic market or supporting ecosystem.

Challenges remain. Analysts such as Pranay Kotasthane highlight execution gaps, noting that only 15 per cent of the Design-Linked Incentive scheme’s FY24 budget was disbursed. China’s Big Fund-III is deploying $47 billion at 28nm, with SMIC already slashing wafer prices to flood the market.

Yet India’s position is strengthened by Western trade barriers. The US imposed a 50 per cent tariff on Chinese semiconductors in January 2025, with further increments due in 2027, while Section 5949 of the US NDAA will bar federal contractors from using Chinese chips in critical systems from December 2027. Trusted-supplier mature-node capacity must come from somewhere, and India is emerging as a credible alternative.

The Dholera fab is expected to ramp to full capacity by 2028, creating over 20,000 jobs and anchoring a supplier ecosystem for gases, chemicals and packaging.

It will directly support India’s electronics industry, which imported components worth ₹34,000 crore last year. The project represents India’s ambition to walk before it runs, recognising that semiconductor success is a marathon spanning decades.

By choosing 28nm, India has finally arrived at the starting line of a thirty-year journey. The allure of 3nm can wait; the chips India actually needs are the ones Dholera will produce.

Agencies