China and the US reshaping their semiconductor supply chains present both challenges and opportunities for India: Chris Miller

Raw material is not the most complex part of the chip industry. It’s refining them to the requisite level of precision because in the chip industry, manufacturing processes often require 99.9999 percent purity for many materials. That’s actually an underestimate for certain of the materials involved. The challenge isn’t digging them out of the ground or acquiring the materials in their raw form, but processing and purifying them. It happens at specialty chemical firms, largely in Japan, which is the world’s leader in this type of ultra-purified, ultra-refined chemical and material production.

It is a huge issue in every semiconductor company I talk to and in every country that I go to. This is a place where India has plenty of work that it can do to further develop its talent pipeline. A large number of Indian employees are already working in semiconductor companies, doing things like design and verification. This is a positive, as these employees have gained expertise in the industry. The new batch of engineering talent from India should be coupled with business expertise. Especially when you start talking about establishing new companies, partly it’s about technical expertise, but a lot of it is about identifying the business opportunity and understanding marketing, supply chain management and product development. Those are things that you don’t learn in engineering class, but you do learn in business school and in running businesses. This part of the talent problem often gets underestimated in discussions, but it’s actually just as important as a degree in engineering or material science.

Most chips consumed in dollar value would either be in the smartphone or data centre sectors. Both of them consume roughly a quarter of all chips that are produced in dollar value. And that’s because the chips that go inside your phone and in data centres are extremely expensive. Other sectors, like consumer devices and traditional manufacturing, consume cheaper chips. A new car often has 1,000 chips inside, but many of those chips are much cheaper. Sometimes they can cost only $1 or even less, whereas a data centre chip can cost up to $10,000.

The fact that China and the US are trying to see supply chains reshaped presents both challenges and opportunities for India. The challenge is that because China is investing so much and subsidising its firms so heavily, it is difficult for companies from any country to compete with China in an open market. Even Japan, Taiwan and the US find themselves competing head-on with heavily subsidised Chinese competitors. The Chinese firms don’t have the most advanced technology. But in lower-tech chips, they’re just as good as many established firms, and that’s why established firms are feeling a lot of price pressure.

The opportunity is that the US, Japan, Taiwan, Korea and European firms are trying to establish China plus one supply chain. In other words, have a China-focussed supply chain largely intended to sell to Chinese customers, and then have a rest-of-the-world supply chain that involves as little of China as possible. This presents a lot of opportunity for India in chip making, testing and packaging, and device assembly. The majority of this today happens in China, which is going to shift. Apple is a good example of this. Until recently, it assembled almost all of its phones in China. Reportedly, it’s trying to assemble up to a quarter in Tamil Nadu by the end of next year. That’s just one company, emblematic of the shifting supply chains that are underway.

The Chinese approach has not been primarily market-driven. China has invested in firms that are not very profitable. There has been little use of a market mechanism to determine where the Chinese investment funds are going. Measuring return on invested capital, the Chinese chip investment funds are probably among the worst venture capital firms in the world in terms of return. That’s probably not a model that you’d want to emulate. Thinking hard about how you get a good return on capital is key to long-term viability, which means you’ve got to invest in firms that have a viable business model in the long run.

On the bright side, China has done much better where it has attracted device assembly, like the assembly of phones or PCs, which is a really low value-added business in the assembly. The new smartphone costs $1,000; the actual assembly process only costs $5 or $10. Take an iPhone, for example. Of all the materials inside, around a quarter of them by value are sourced from Chinese firms. In the initial iPhones, it was only the assembly that was done a decade ago. China has dramatically increased the value that accrues to Chinese firms in smartphone production by learning to produce the components as well. And we’re in the very early stages of seeing that in India. It’s going to take some time for Indian firms to learn how to produce and competitively produce the types of components that go into electronics. It’s not a good idea to just focus on assembling devices because you’re not capturing a lot of the value. Whereas if you start producing components, suddenly more value starts accruing to your firm.

They’ve got vast investment programs underway to develop new technologies. They’re going to keep shrinking transistors and packing them more densely together. These firms also continue to invest in new design methodologies to ensure they’ve got better designs as well as better manufacturing processes. That’s going to be just as important as the manufacturing improvements that are coming down the line.

AI is going to play a bigger and bigger role in the chip industry and change the way chips are designed and produced. AI will be a larger source of demand for new types of chips. That’s what we see right now, with Nvidia becoming the world’s third-most valuable company. A series of companies and products that emerge will be driven by the need for better chips for AI.