On the horizon: A semiconductor war

Topics semiconductor | USA | China

The technology war between the US and China is largely a battle over semiconductors. Semiconductors or chips/integrated circuits(ICs) are the lifeblood of all digital products and digitalisation. They are embedded in every product possible and their penetration is only rising. Just to give context, a simple MacBook today will have over 16 billion transistors compared to tens of thousands in the Apollo lunar module, which landed on the moon in 1969. With the advent of 5G, expect chips to be embedded into every product as connectivity becomes all pervasive. An electric vehicle will have 3-5 times the number of chips as the most advanced internal combustion engine cars of today. This number goes up another five times with the advent of autonomous cars.

From a strategic perspective, chips are now more important than oil. In the case of China, they import 50 per cent (by value) more IC’s than oil, and it is their single biggest import. China is the world’s largest consumer of chips and accounts for between 25 and 33 per cent of the global semiconductor demand.

The US dominates and leads in most of the key technologies in the semiconductor value chain, and in the areas it does not lead, its allies dominate. This dominance in semiconductors is the basis for much of its global technological and geopolitical leadership. In the Trump administration, this technology lead was effectively weaponised, as the US government was able to cripple any Chinese company that it saw as a strategic challenger. Huawei, at one time the largest smartphone company in the world, has had to cede market share and product leadership as US sanctions bite.

China recognised its strategic vulnerability as far back as 2014, and decided to build a large and globally competitive semiconductor industry. It pursued this aim with the most well-funded state-sponsored industrial policy programme in history, spending hundreds of billions of dollars in grants, direct funding, tax breaks and low-cost capital support. It also encouraged global M&A to acquire technology. Despite being incredibly successful in most other areas of industrial policy focus, be it high-speed rail, renewables, electric vehicles and batteries or Artificial Intelligence, China has found mastering semiconductors hard.

Even today, if you look at the semiconductor value chain, China has a limited presence in the higher value parts of the chain. There is virtually no Chinese presence in the semiconductor capital equipment, software or materials like specialised chemicals ,wafers and gases, used to make chips.

While China has a large chip manufacturing capacity, almost 20 per cent of global capacity, the majority of this is controlled by foreign firms like TSMC or Samsung. Foreign-owned firms account for more than 50 per cent of production volumes and more than 75 per cent of value. China-owned capacity is concentrated almost entirely in manufacturing older, less sophisticated chips. High volumes, but low value.

If we break down the value chain for semiconductors, the two main pieces are IC design and IC fabrication (manufacturing). Other parts of the chain are assembly test and packaging, software, materials and semiconductor capital equipment.

Only those firms have had success in semiconductor fabrication which have been able to invest the billions of dollars needed to keep pace with the relentless march of Moore’s law and the continued miniaturisation of chips. The top three chip manufacturers, Intel, Samsung and TSMC each spend more than $20 billion annually in semiconductor capex. A single Fab will cost billions. There is also the relentless trend towards outsourcing of manufacturing, with the vast majority of chip companies getting their semiconductors manufactured by specialised outsourcers called foundries. These fabless chip companies focus on design and let the foundry do the highly capital-intensive and complex job of actually making the chip.

Despite pumping billions into its national champions like Semiconductor Manufacturing International Corporation (SMIC), China has not been able to create a global leader in semiconductor manufacturing. SMIC is only one-tenth the size of the largest foundry (TSMC) and, at a minimum, five years behind in manufacturing technology. It has not been able to narrow the technology gap, despite the large state support. China’s share of global foundry revenue remains under 10 per cent. Despite being a capital-intensive industry entirely focused on manufacturing capability and excellence, historically its source of competitive advantage, China has not been able to make much of a dent. It shows the barriers to entry and the difficulty in keeping up with a continuously declining cost curve based on yield improvement. TSMC is the giant in the foundry space with current revenues of $45 billion, technology leadership and a three-year capex plan of $100 billion.

Illustration by Binay Sinha
In chip design, China has had more success. While even today almost 50 per cent of IC design work by value is done by fabless US firms, China has been successful in creating world class firms like HiSilicon, focused on designing chips for mobile and telecom use. There are numerous other start-ups in IC design in China with a very robust eco-system. China has about a 15 per cent market share in fabless chip design revenues globally.

In the other niches of the semiconductor supply chain, China has a minimal presence. For the foreseeable future, China will remain dependent on the US. Despite more than a decade of effort and hundreds of billions of dollars of capital, China is even today totally reliant on the US and its allies for the equipment, software and materials used to make chips. It will be decades before this will change.

The implications of the above are clear for India. As we also aspire to be a manufacturing hub for electronics and mobile phones, it is only a matter of time till semiconductor imports surge and cross that of petroleum for India. Given the China experience, it is unlikely that we have the resources to even attempt to manufacture semiconductors. It would make sense for India to focus on chip design and leave manufacturing for others. In 1990, 100 per cent of all chip manufacturing was in the US, Europe and Japan. Today, only about 30 per cent is in these three blocs, with the balance being in China, Korea and Taiwan. Given the strategic importance assigned to controlling your chip supply chain, every major economic power wants to have a chip fab in the country. There will be a surge in new capacity globally. India will have more than enough choices of where to import from. It makes sense to use our scarce capital to incentivise and promote chip design and niche areas of manufacture in older generation chips. We have absolutely no chance to even contemplate manufacturing contemporary technology semiconductors. />
The writer is with Amansa Capital



Business Standard is now on Telegram.
For insightful reports and views on business, markets, politics and other issues, subscribe to our official Telegram channel