The Semiconductor Mission and the Honest Limits of What Industrial Policy Can Buy

When the Union Cabinet approved the India Semiconductor Mission in December 2021 with an outlay of roughly seventy-six thousand crore rupees, the announcement was met with a familiar mix of enthusiasm and skepticism. Enthusiasm, because India has watched the electronics trade deficit widen for two decades while countries like Taiwan and South Korea built the most consequential manufacturing capability of the digital age. Skepticism, because India had tried and abandoned semiconductor fabrication ambitions before, most notably after the Jaypee-IBM and other proposed fabs of the early 2010s collapsed for want of financing and follow-through. Four years on, the picture is more concrete than either the boosters or the doubters initially expected, and it is worth examining precisely what has been built and what remains structurally difficult regardless of how much capital is deployed.

What has actually broken ground

The mission's most visible outcome is the Tata Electronics fabrication plant coming up in Dholera, Gujarat, in partnership with Taiwan's Powerchip Semiconductor Manufacturing Corporation, targeting mature-node chips used in automotive and industrial applications rather than cutting-edge logic chips. Alongside it, Tata's assembly and test facility in Assam and Micron's packaging plant in Sanand represent the less glamorous but arguably more immediately useful end of the value chain: Assembly, Testing, Marking and Packaging, or ATMP, work that does not require the extreme ultraviolet lithography machines that only Dutch firm ASML manufactures, but does create a genuine base of skilled technicians and process engineers. The government's decision to weight incentives toward fifty percent of project cost, among the most generous globally, reflects an understanding that India was entering a capital-intensive race decades late and needed to compensate with money for the years it lacked.

Why chips are not like other manufacturing bets

Semiconductor fabrication is unusual among industries in that its economics are governed almost entirely by yield curves and cumulative learning, not by labour cost or land availability, the traditional levers of Indian industrial policy. A fab that opens today does not compete on the technology it starts with; it competes on how quickly its engineers can push defect rates down and output up, a process that took TSMC, now the dominant contract manufacturer globally, several decades and multiple generations of engineers who stayed within the same firm and the same ecosystem long enough to accumulate tacit knowledge that no manual or consultant transfer can replicate. This is the uncomfortable truth industrial policy in this sector runs up against: capital can build a cleanroom in eighteen months, but it cannot compress the learning curve that determines whether that cleanroom becomes profitable.

The talent and supply chain gap

India's engineering talent pipeline is often cited as a natural advantage, and it is real; Indian-origin engineers occupy leadership roles across the global chip industry, from AMD's Lisa Su's contemporaries to senior architects at Nvidia and Qualcomm. But there is a difference between producing software and systems engineers, where India has genuine depth, and producing the specific discipline of process engineers who understand photolithography, ion implantation and chemical vapour deposition at the level required to run a fab profitably. That talent pool barely exists domestically and is being built essentially from scratch through partnerships with Purdue University and other institutions, a sensible but slow fix. Equally constraining is India's near-total absence from the specialty gas, photoresist and ultra-pure chemical supply chains that a fab consumes daily, all currently imported, which means the mission's early output will carry import dependencies embedded within it even as it reduces India's finished-chip imports.

The right comparison is not Taiwan, it is Vietnam

A more useful comparator than Taiwan's rise in the 1980s is Vietnam's more recent success in electronics assembly, where Samsung's investment created an ecosystem of component suppliers and a skilled workforce over roughly fifteen years, turning Vietnam into a major electronics exporter without Vietnam ever building an advanced fab of its own. India's semiconductor strategy, whatever its packaging suggests, is closer in practice to this ATMP-and-mature-node model than to a bid for leading-edge fabrication, and that is a defensible sequencing choice rather than a failure of ambition. The mistake would be to measure the mission's success against the yardstick of matching TSMC's three-nanometre capability, a goal no serious planner inside the mission has actually set for this decade.

What would count as success

A fair five-year assessment of the India Semiconductor Mission should look for specific, measurable outcomes: whether the Dholera and Sanand facilities reach commercial yield rates comparable to established mature-node fabs elsewhere, whether a second and third wave of private investment follows without matching subsidy intensity, and whether India develops even a modest domestic base in the ancillary chemicals and equipment that fabs require. None of these outcomes are guaranteed, and the mission could still stall the way earlier attempts did if financing partners waver or if global chip demand cycles turn down before Indian facilities reach scale, a real risk given the industry's history of boom-bust investment cycles.

What should not happen is a repeat of the pattern where each new government scheme is graded purely on the press conference that announced it rather than on the harder, less telegenic metrics of yield, retention of trained engineers, and supply chain localisation a decade out. Industrial policy of this kind is a multi-decade commitment, and its architects inside the Ministry of Electronics and Information Technology appear to understand this better than the surrounding political commentary often does. The measure of whether India's semiconductor bet succeeds will not be visible in 2026; it will be visible in whether, by 2035, an Indian-fabricated chip appears in an ordinary Indian-made car without anyone finding that fact remarkable.