For forty years the semiconductor industry treated assembly, testing and packaging — ATP, in the jargon — as the part of the value chain you did not have to think about. The wafers came out of the fab worth thousands of dollars each; the packaging houses in Taiwan, Malaysia and China cut them up, wired them into plastic and ceramic, tested them, and shipped them on for a few percent of the chip's value. It was logistics with a cleanroom attached.
That era is ending, and the reason is not technology. It is paperwork. When the United States, Japan and the European Union began insisting that 'secure' supply chains include the back end, assembly suddenly required what fabs had always had: host governments, subsidy frameworks and strategic rationales. India's semiconductor mission covers packaging at fifty percent capital support. Vietnam's national strategy names ATP explicitly. What used to be a line item is now a policy instrument.
The economics underneath the politics
The shift is real but it needs qualifying. Lumera's packaging capacity model counts $31 billion of announced ATP investment outside the traditional Taiwan-Malaysia-China corridor for 2024–28, of which about $9 billion has broken ground. India accounts for six committed projects, Vietnam four. But the concentration of skill is harder to move than the capital: a wire-bond line can be bought; the process engineers who keep its yield at 97 percent take a decade to grow.
This is the constraint nobody's subsidy solves. India's Semiconductor Mission documents assume roughly 12,000 packaging engineers and technicians by 2028. Our estimate of the actual pipeline — from the IIT packaging programmes, the OSAT firms' own training academies and experienced hires from Southeast Asia — is about 7,500. The gap will be filled, partially, by expatriate Malaysians and Taiwanese doing two-year rotations, which is precisely how the knowledge transfer is supposed to work.
The last ninety nanometres of a chip's journey have become a geopolitical decision, and nobody trained a workforce for that.
The advanced-packaging story is the second act. Chiplets made ATP the most technically interesting part of the industry: heterogeneous integration, hybrid bonding, 2.5D interposers. The new geography of commodity assembly will be decided by subsidies, but the geography of advanced packaging will be decided by proximity to fabs and design houses. India's OSAT projects are overwhelmingly conventional today. The test of the mission is whether any of them lands an advanced line before 2028.
The companies with quietly excellent positions here are the equipment vendors — the bonders, testers and handlers are agnostic to where the lines are built, and everyone is building. For the rest of the industry the advice is unglamorous: qualify a second ATP supplier now, before the geopolitics forces you to qualify a third.