In the 1970s, Hsinchu and Mohali stood at much the same fork.

Neither country could yet compete with the United States or Japan in chips. Both knew foreign technology would have to be acquired and then absorbed at home, and both put a state institution in between. Taiwan established the Industrial Technology Research Institute, or ITRI, in 1973, and India’s Cabinet approved Semiconductor Complex Ltd. at Mohali in 1976.

The difference came afterwards.

K.T. Li, remembered as the father of Taiwan’s economic miracle, had served as economic affairs and finance minister. From 1976 he oversaw science and technology and steered Taiwan towards precision manufacturing over design. ITRI licensed process technology from RCA and trained a generation of engineers. ITRI then moved both into private companies, United Microelectronics Corporation in 1980 and Taiwan Semiconductor Manufacturing Company in 1987.

Mohali’s capability stayed inside the state.

Both countries acquired foreign technology, but only Taiwan built companies able to absorb, improve and keep what was learned.

Morris Chang, educated at MIT and Stanford and a Texas Instruments veteran, founded TSMC on ITRI’s technology and engineers. TSMC would make chips for firms such as Qualcomm and Nvidia, never competing with its customers. With each process generation, manufacturing knowledge accumulated inside the company. TSMC came to own the manufacturing technology without owning the chip designs.

A government development fund supplied 48.3% of TSMC’s starting capital, and Philips contributed about 27.5% in cash and technology. TSMC listed in Taipei in the early 1990s and in New York in 1997. In 2025 the company’s Taiwan tax bill was estimated at $6.3 billion, the highest on the island. The state’s early bet had come home.

Mohali’s road turned in February 1989, when a fire destroyed SCL’s production line. SCL, now a laboratory, sits under the IT ministry. Capability survived, but ownership never followed.

That is the puzzle.
Why did two similar starts end so differently?

Taiwan’s own story shows a licence is a starting point, not a dead end. RCA gave ITRI a recipe. Absorption needed engineers who knew the recipe well enough to find its faults and fix them. Improvement needed companies able to keep those engineers and the yield data they generated. Retention needed ownership, because knowledge created inside a foreign-owned plant belongs to the foreign owner.

Research spending is a flow, written off each year. Technology capital is a stock, built through learning accumulated inside a company.

India’s private companies spent ₹1.18 trillion, about $14.3 billion, on R&D in 2023-24, while TSMC alone spent $7.9 billion in 2025. In 2022-23 India paid $10.6 billion to use IPR owned abroad and received $1.3 billion.

Technology compounds when firms build the capacity to learn, improve and develop the technology over successive generations. A licence fee alone buys none of those. Where learning does not accumulate, dependence persists.

India never lacked engineers. Much of the knowledge they built migrated overseas or into foreign-owned capability centres. One road carried knowledge away, and the other turned knowledge into capital.

Technology capital built inside companies adds up to national strength. Chip exports helped Taiwan, an energy importer like India, run a $181 billion current-account surplus in 2025. Its chip industry is widely called a “silicon shield”.

India’s Comeback
Better late than never, India is back on the semiconductor road.

Semicon 1.0 in 2021 committed ₹760 billion, about $10 billion, and Semicon 2.0 in 2026 added ₹1.28 trillion, about $13.3 billion.

Tata Electronics is building a foundry with Taiwan’s PSMC, CG Power and Kaynes are packaging chips, and L&T is entering design. ASML, the dominant maker of lithography machines, began Indian operations in 2026. The road is reopening, with Indian private companies at its centre.

Chip startups are rising beside the fabs. About 620 have been founded in five years, raising just over $1 billion in all, roughly 13% of TSMC’s research spending in 2025. Some have moved beyond design services to build IP of their own, and foreign buyers have noticed. Japan’s Renesas bought Steradian, a Bengaluru radar-chip designer, in 2022. Germany’s Infineon agreed in 2026 to buy C2i, a Bengaluru power-chip specialist.

A sale can reward founders and investors, fund further development and put an Indian team on a global product platform. The engineering need not leave India. Infineon plans to make C2i a centre of excellence in Bengaluru. Ownership is a different matter. The IP now belongs to a foreign parent, which decides what gets developed next and where. Foreign acquisitions are welcome. The question is what Indian companies keep.

Factories, funding and acquisitions are milestones, not the destination.

The real measure comes later.
Can Indian engineers improve the process without the foreign partner?
Can Indian companies patent the result and raise equity on its strength?
Can a promising startup list in India rather than sell to a foreign buyer?

Subsidies will be spent and factories will open. The test is not whether India makes chips. The test is whether Indian companies own the recipe, and can write the next one.