The headline number, and the number underneath it

India's Semiconductor Mission (ISM) has now approved 12 projects worth a combined $20 billion, or roughly Rs 1.65 lakh crore 1. That figure has become shorthand for India's arrival as an Asian chip hub, cited in every pitch deck about diversifying away from Taiwan and Korea. It is also misleading in a specific, investable way: of those 12 projects, three are operational, and every one of them is an assembly, packaging or test line, not a wafer fabrication plant 1 6. The distinction matters more than the headline number, because it tells investors which layer of the supply chain India has actually captured, and which layer is still a 2028 promise.

What is actually shipping

Micron Technology's $2.75 billion ATMP facility in Sanand, Gujarat, opened on February 28, 2026, and is now shipping DRAM and NAND modules for phones, data centres and automotive customers 6. Kaynes Semicon's roughly $348 million OSAT plant in the same Sanand cluster reached commercial operation on March 31, 2026, packaging and testing around 6 million chips a day 6. CG Power's joint venture with Japan's Renesas and Thailand's Stars Microelectronics began commercial production on July 4, 2026, also in Sanand 6. All three take chips fabricated elsewhere and finish them locally. None etches a transistor on Indian soil. That is a legitimate, fast-payback business, and it is why Sanand has become India's most credible electronics cluster. But it is downstream value capture, not the upstream fabrication capacity that determines whether India becomes a genuine alternative source for the chips themselves.

The fab that was supposed to change that just slipped

The one project meant to answer that question is the Tata Electronics-PSMC 300mm fab in Dholera, Gujarat, roughly a Rs 91,000 crore build. On July 17, 2026, Tata Electronics' own spokesperson confirmed to reporters that commercial production, previously guided for the end of 2026, is now expected in mid-2028, an 18-month slip 2 8. More striking than the delay is the technology downgrade: the plant will start at the 90-nanometer node, two full generations behind the 28-nanometer node that Tata Sons chairman N. Chandrasekaran described in the company's own FY25 annual report as the deliberate starting point 2 3. PSMC has framed the sequencing as normal industry practice, ramping mature nodes before advanced ones, but the gap between the public commitment and the operating plan is now a matter of record, not spin 2. Construction is roughly halfway complete, with first silicon still the near-term milestone to watch rather than commercial volume 2.

Two more approvals, same pattern

On May 5, 2026, the Cabinet approved two additional projects in Gujarat worth about Rs 3,900 crore combined, bringing the ISM total to 12 5. Crystal Matrix will build an integrated compound-semiconductor fab and packaging line in Dholera for GaN-based Mini/Micro-LED displays; Suchi Semicon will build an OSAT line in Surat for lead-frame and wirebond packaging used in air conditioners, EV battery-management systems and solar inverters 5. Both are real capacity and real jobs. Neither is leading-edge logic fabrication, reinforcing that India's near-term chip buildout is concentrated in packaging, testing and specialty compound semiconductors rather than the advanced-node wafers that anchor AI and premium-device supply chains.

Delhi's own diagnosis: Semicon 2.0

The government's response is the more interesting signal for investors. On July 15, 2026, the Cabinet approved Semicon 2.0, a second-phase outlay of about Rs 1.275 lakh crore (roughly $13.4 billion), explicitly targeting the layer beneath the fabs 4 7. The stated problem: every chip plant India has built or is building runs on semiconductor-grade chemicals sourced almost entirely from Japan and Germany, and on fabrication equipment sourced almost entirely from the United States and the Netherlands 4. Semicon 2.0's six pillars, spanning design, fabrication, packaging, research, talent and equipment and materials manufacturing, amount to an admission that assembly-stage self-sufficiency was the easy part 4.

The investable read

For Asian supply-chain investors, the near-term opportunity in India remains where the capital already proved out: backend assembly and test, where Sanand's three operators have working revenue lines and Crystal Matrix and Suchi Semicon are building toward the same model. The wafer-fabrication thesis, the one that would let India compete with Taiwan or Korea on leading-edge output, is now explicitly a mid-2028 event at best, and its credibility took a real hit when Tata's own node commitment did not survive contact with construction. The segment worth tracking next is the one Delhi just funded: the chemical suppliers, gas producers and toolmakers that would need Indian footprints to service Semicon 2.0's mandate, a less crowded trade than the fabs themselves and one where global materials and equipment names could find a first-mover advantage before local substitutes emerge. Watch first silicon at Dholera in 2026 as the next real checkpoint; a further slip there would say more about India's fab ambitions than any new approval announcement will.