Tata's Dholera Chip Fab to Launch Mostly on 90nm Process

India's flagship chip project is not going to start where a lot of people were assuming it would. According to Bloomberg's reporting, Tata Group's first large-scale fab, in Dholera in western Gujarat, will begin production mostly on 90-nanometer process technology, described as older than what had originally been expected from India's semiconductor debut.

The setup itself is not small. The facility is a roughly $11 billion project (about ₹91,000 crore in the Tata announcement), built with Taiwan's PSMC providing technology and manufacturing support, with planned capacity of up to 50,000 wafers per month of analog and logic ICs. The PSMC partnership does cover a portfolio that spans 28nm, 40nm, 55nm, 90nm and 110nm, so Tata has a technical path toward finer geometries on paper. What the reporting says today is that the launch mix is weighted to the mature end of that range, not the leading edge.

That is a meaningful reset of expectations. India's semiconductor mission has been pitched as a step toward strategic self-reliance, and 90nm is a very different product from what the phrase "first commercial fab" usually conjures in that debate. It is still real, useful silicon, the kind of node that anchors power management, automotive and industrial parts, and there is genuine demand for that capacity currently served mostly out of Taiwan and China. A domestic option matters for buyers who care about supply resilience even if it does not match the marketing.

The honest caveat is that Bloomberg is describing a launch mix, not a permanent ceiling. Whether Tata actually climbs the node ladder, on what timeline, and with which customers is not something today's reporting settles. What the story also does not give you is the yield or cost economics, so it is hard to judge from the outside how competitive a Dholera-made 90nm wafer will be against incumbents that have been running the same node for twenty years.

For customers who were quietly hoping for a nearer-shore 28nm option in the next couple of years, the sensible read is to keep planning as if that option is not there yet. For everyone else, the interesting question is which OSAT, materials and design partners now anchor themselves around Dholera, because a real fab with a real customer pipeline is the seed a domestic ecosystem actually needs.