Despite India’s poor track record, Tata takes a gamble on semiconductors
Last week, Tata group Chairman N Chandrasekaran created a stir when he announced that the salt-to-automobiles conglomerate may enter the semiconductor business.
This is certainly a courageous step given India’s unsuccessful trysts with the industry (see table). But when a conglomerate of the size and stature of the Tata group makes an announcement to this effect, it needs to be taken seriously. It is the first big Indian corporate house to enter this business and has taken preliminary steps to get going. “The group is roping in the best talent in the semiconductor business and setting up a plant to manufacture sophisticated components for mobile phones,” said Satya Gupta, former chairman of India Electronics and Semiconductor Association (IESA).
Randhir Thakur, president of Intel’s foundry division and a key aide to Intel CEO Pat Gelsinger, has come on board as independent director in Tata Electronics, the company through which the semiconductor foray is being put into operation.
This could well be part of a broader plan. Gelsinger is putting in billions of dollars to build foundries across the world to make chips for third-party players and not only for Intel — a clear shift in strategy. This, it has been suggested, could be an opportunity for the Tata group to be part of Intel’s global supply chain, or later consider an alliance to set up a fab plant in the country. The Tata group declined to discuss the issue.
The group has also hired Raja Manickam, co-founder of Tessolve, a Bengaluru-based semiconductor engineering solutions provider that he sold to Hero MotoCorp. He is, sources said, an expert in building an Outsourced Semiconductor Assembly and Testing (OSAT) business, a stage just below the setting up of a foundry plant.
ALSO READ: Tata Motors looking at measures to offset impact of semiconductor shortage
The group is also investing Rs 5,000 crore to set up a manufacturing plant for electro-mechanical components for the mobile phone, industry leveraging know-how from Titan Industries. Talks are on with Apple Inc for a contract.
The timing is opportune. Globally, every country, including India, is looking at reducing its dependence on imports for chips and moving towards self-sufficiency. The huge pandemic-induced global shortage in chips, which is expected to last another year, has impacted key industries from mobile companies to auto makers. The US and Germany are offering generous incentives to support their domestic semiconductor industries to promote self-sufficiency.
So what are the Tata group’s choices? “I think the Tatas will not go for a vertical take-off straight to a fab plant. They will move phase by phase,” said the CEO of a leading global chip design company.
Setting up an OSAT facility could be the right start. The global market for OSAT is estimated at $32 billion and is expected to grow six-fold to $180 billion by 2028. There are already large independent global players in the space — for instance, Taiwan-based ASE Technology Holdings (revenues: $11.7 billion). “Being an OSAT player is one option as it requires $0.5 billion to $1 billion of investment,” the executive said.
The second option is to acquire a second-hand foundry or build a plant making high nanometre chips of 28 nm, a low-tech, low-margin but large-volume market, which would require investments of $2-5 billion. However, with the market shifting to smaller nanometre chips much faster than anticipated as companies demand more processing power, efficiency and less power consumption, the volume chip market could shrink after five or seven years.
But investing in high-tech fab technology (12 nm and below) means a huge bill — $10-15 billion — and a similar amount to upgrade technology in the next five to 10 years with break-evens that could take over a decade.
To get into that game, the Tata group has to follow the global leaders. TSMC, the world’s largest chip-maker, has announced an investment of $30 billion in three years, doubling what it did in the last three years. Intel has committed investments of $20 billion for two new fab plants in the US, while Samsung has a plan to put in $110 billion by 2030.
Yet whatever technology the Tatas choose, no fab plant can survive serving only domestic demand — even though it is estimated at $21 billion. Said Gupta, “The ideal ratio should be 80 per cent global and 20 per cent local. After all, governments can’t force companies to break their existing supply chain and buy from an Indian chip maker because it is made in India. In any case, there are so many different kinds of chips required, which one player can never manufacture for everyone in India. That would be wishful thinking.”
A third, more cautious route is to set up smaller “speciality fabs”, which are wafers, made of gallium nitride or silicon carbide and have multiple applications in chargers, electric vehicles, base stations and data centres. The investment required is just about $250 million to $500 million and the break-evens are much faster.
The last part of the puzzle is how much the government wants to subsidise the industry. Without that any grandiose plans will fail. In one of its presentations, the Ministry of Electronics and Information Technology, or MeitY, had said that it had earmarked $1 billion as incentive for setting up fab plants.
That’s nothing compared to what competing governments are offering, experts point out. But the government has also set up an empowered committee under the minister of commerce to clear high-tech manufacturing projects and support them. With Chandrasekaran one of its key members, he will surely be able to make his voice heard.
