The foundries used to supply chips to a major player. These days there is a change of strategy. Foundries are being built across the world to supply chips for third-party players. Gelsinger follows this strategy.

There is a business of Outsourced Semiconductor Assembly and Testing (OSAT). This is a stage below the setting up of a foundry plant. There is a good global market for OSAT. The players in this field are ASE, Taiwan and some independent global players. OSAT requires an investment of $0.5 billion — $1billion.

There is a business of making high nano-metre chips of 28 nm. This is a low-tech, low margin business, but it has large volume market. It requires an investment of $2 — 5 billion.

The market is shifting fast to smaller nanometre chips since companies demand more processing power and less power consumption. High-tech fab technology ( 12 nm and below) requires huge investment — $10 –15 billion. Technology upgradation requires another $10-15 billion in next 5-10 years. The break-evens could take more than a decade.

International chip makers such as TSMC, Intel and Samsung have committed $20 to $30 billion over the next 2-3 years.

A fab plant has to sell 80 per cent of its production globally and 20 per cent to domestic users to survive.

There is a route of specialty fabs which are waters of gallium nitride or silicon carbide. These have multiple applications in chargers, EVs, base stations and data centres.

Tata proposes to be a part of semiconductor industry. They will have to consider the above options carefully. They can also buy a second hand foundry.