Blitz Bureau
NEW DELHI: India has approved ₹1,65,685 crore of semiconductor projects across 12 commercial facilities in six States. Three are operational. The interesting number is not the outlay — it is the gap between twelve and three, and what has to happen inside that gap.
Semiconductor manufacturing is the hardest thing a country can decide to learn, and it is worth being honest with readers about why. A fabrication plant is not a factory that can be commissioned and then run. It is a chemistry laboratory operating at the scale of a steel mill, requiring ultrapure water by the million litres a day, power that cannot flicker, a supplier base for gases and photoresists that mostly sits in three countries, and — the binding constraint everywhere — engineers who have done it before. Countries that have tried to buy their way in have generally discovered that the equipment is the cheap part.
First silicon is the milestone that counts: the Tata-PSMC fabrication plant at Dholera is targeting its first wafers off the line in December 2026.
Assembly teaches a country to work. Fabrication teaches it to invent. India is deliberately doing the first before the second.
At a Glance
• Approved investment: ₹1,65,685 crore under the India Semiconductor Mission
• Facilities approved: 12 commercial sites across six States
• Operational: three
• Semicon 2.0 outlay: ₹1,27,500 crore, cleared by the Union Cabinet
• First operational unit of this cycle: Micron’s ATMP plant at Sanand, inaugurated 28 February 2026
• First fabrication plant: Tata-PSMC at Dholera, under construction
• First silicon target: December 2026
• Recent Cabinet approvals: Crystal Matrix (compound-semiconductor fab and ATMP, Dholera) and Suchi Semicon (OSAT, Surat)
Which is why the sequencing matters more than the headline outlay. India began with assembly, testing, marking and packaging — the back end of the chip industry, where the yields are forgiving and the learning is real. Micron’s Sanand plant, inaugurated in February, is an ATMP facility. Suchi Semicon at Surat is an OSAT. These plants will not design a processor, but they will produce something more useful in the medium term: several thousand Indian engineers and technicians who have run a semiconductor line to specification, and a domestic supply chain of chemicals, substrates and clean-room contractors that did not exist five years ago. The fabrication plant at Dholera, targeting first silicon in December, is the harder step, and it is being attempted with a partner — Taiwan’s PSMC — that has run these lines for decades.
The structural point for a reader thinking beyond this year is that semiconductors are the test case for whether India can build a deep industry rather than an assembly economy. The measures that will tell us are not ribbon-cuttings. They are yield rates after the first year, whether a second fab is financed with less state support than the first, whether Indian firms start supplying gases and masks rather than importing them, and whether the engineers trained at Sanand stay in the country. Semicon 2.0’s ₹1,27,500 crore buys India time and a second attempt if the first stumbles — which, in an industry where nearly every entrant has stumbled, is the sensible thing to buy. The honest way to report this is neither triumph nor scepticism. It is a decade-long project, three years in, with the hardest milestone four months away.
