The lesser-known prospects in the Asian chipmaking supply chain

FIVE years ago, I argued that electrification, connectivity and changing patterns of vehicle utilisation would steadily erase the boundary between the car and semiconductor industries.

That convergence is now well under way. Of those three forces, the electrification of vehicles has moved furthest by a wide margin.

It has reshaped manufacturing footprints, rewritten the automotive semiconductor bill of materials and transformed upstream supply chains.

At the same time, artificial intelligence is creating another powerful source of demand for the same physical infrastructure: chips, power, manufacturing capacity and the equipment that connects them.

The more interesting investment question today, therefore, is no longer whether these industries intersect. It is where the economics of that convergence ultimately accrue.

The obvious answers are familiar.

Taiwan remains indispensable to leading-edge chips; South Korea to memory; and Japan to semiconductor equipment, materials and power devices. China has built extraordinary scale in electric vehicles, batteries and increasingly, the semiconductor supply chain itself.

These remain formidable positions.

But investors looking at only the largest and most visible semiconductor companies risk missing another part of the story: businesses sitting elsewhere in the value chain, with products, engineering capabilities or manufacturing processes that are increasingly difficult to bypass.

The opportunity may be less about picking one national champion, than identifying the layers the broader technology build-out cannot easily do without.

From assembly to precision manufacturing

India provides one of the clearest examples.

On a visit to Bengaluru recently, I walked the floors of manufacturers serving the automotive industry.

What struck me was not simply that more floor space was being devoted to components destined for EV and semiconductor customers, both domestic and global. It was also the sophistication of what was sitting on those floors.

Equipment that only a few years ago might have been considered too expensive for Indian manufacturers was becoming increasingly commonplace. What I also did not see was equally striking: people. The extent of automation surprised me.

That matters, because India’s manufacturing story is beginning to move beyond labour-cost arbitrage and basic assembly.

Electronics manufacturers are expanding into precision tooling, mechanical enclosures and other higher-value parts of the supply chain, while automotive tier-1 suppliers, who work directly with vehicle companies, are building capabilities in areas such as high-pressure aluminium die-casting and EV power-train components.

This also connects back to the original car-semiconductor thesis.

India’s EV trajectory has developed differently from China’s, led more heavily by two and three-wheelers, while its semiconductor ambitions are initially focused on mature nodes and packaging, rather than the technological frontier.

That combination may prove pragmatic: The workhorse chips used in vehicles, power systems and industrial applications do not all require the most advanced process technology.

But the opportunity today is already broader than a simple “China plus one” diversification story. India brings its own design, engineering and production skills in established and newer companies.

As Indian manufacturers adopt more sophisticated tech, and vehicle companies tighten their customer-qualification process, the sector’s barriers to entry are raised.

This can create greater switching costs for companies, entrenching the position of Indian manufacturers in the supply chain.

Moving up the value chain

Malaysia is a different story. The country has been part of the semiconductor supply chain for decades, particularly in assembly, testing and packaging.

The more interesting development is that selected local businesses are moving beyond traditional contract manufacturing into specialised automated test equipment, advanced-packaging handlers and related engineering capabilities.

That distinction is important. Assembly capacity can move. Intellectual property, accumulated process knowledge and highly specialised customer relationships are harder to transplant.

If Malaysian companies can continue moving into these areas, the economics begin to look rather different from that of a conventional outsourced manufacturing base.

This is precisely the type of layer investors should be looking for: not simply exposure to higher semiconductor volumes, but also businesses with a role that becomes more valuable, as chips become more complex, and packaging and testing requirements increase.

The economic moat comes partly from lengthy customer qualification cycles and high costs to swap suppliers, particularly where tolerance for failure is extremely low.

Proprietary engineering and deep process integration can reinforce that advantage, making established suppliers difficult to displace once they are embedded in a customer’s manufacturing process.

The infrastructure behind the intelligence

Thailand offers yet another facet of the same opportunity.

Its traditional strength in automotive manufacturing remains important, particularly as Chinese EV makers establish production in the country.

But the AI build-out is also creating a different set of opportunities around the physical infrastructure required to make computing work.

AI does not run on software alone. It requires electricity, power management, data-centre infrastructure and increasingly sophisticated ways of moving enormous quantities of data between chips.

Thailand already hosts manufacturing capabilities in several of these areas. These may not be the headline products most investors associate with AI, but they sit inside the physical architecture every large-scale AI system depends upon.

For now, power and data-centre infrastructure may offer the clearest illustration of Thailand’s advantage in the supply chain.

These are mission-critical parts of the AI build-out, where technical certification, customer validation and the costs of changing suppliers can create more durable economics than commodity manufacturing.

Not every layer is equally attractive

There is an important caveat.

A rising tide of semiconductor, EV or AI investment does not automatically produce attractive shareholder returns. Capacity can be commoditised. Government incentives can attract too much capital. Rapid expansion can turn into price competition.

China offers a useful reminder. Its extraordinary scale has created global leaders, but it has also produced intense competitive pressure in parts of the EV and battery supply chain.

For investors, the discipline is therefore to distinguish between growth and economic value.

The more durable layers tend to have something that cannot easily be recreated with capital alone: process knowledge, intellectual property, customer qualification, switching costs, specialised engineering or a manufacturing capability built over many years.

Five years ago, the important insight was that cars and semiconductors were converging. Today, that intersection is no longer the insight.

The question now is where sustainable returns of the combined system can be found – and which companies have built advantages that can withstand the next round of capital spending, localisation and competition.

The answer is unlikely to be one market or one national champion. It will be found, layer by layer, in Asia.

The writer is portfolio manager, Asian equities, at M&G Investments