An attractive price for a potential defensive AI play

The sprawl of the semiconductor sector sometimes feels as complex as the chips themselves. There are chip designers such as Nvidia (US:NVDA), and manufacturers such as TSMC (TW:2990). Then there are those making equipment for manufacturers, ASML (NL:ASML) being the most prominent example. In the US, there are three main equipment players, namely Applied Materials (US:AMAT), Lam Research (US:LAM) and KLA Corporation (US:KLAC).

KLA can be distinguished further: it prioritises process control and yield management tools. In short, it provides testing and inspection products that ensure semiconductor manufacturers can run at full speed. 

Process control makes up nine-tenths of revenues and gross profits (the remainder is other component manufacturing and the testing of other chip equipment). KLA’s leading position – the company says its market share is five times bigger than global competitors’ – has helped produce a highly profitable business. High switching costs, as well as KLA's generous R&D expenditure, give it a good competitive moat. Returns on invested capital are above 30 per cent, while its sector-leading operating margin sits at close to 40 per cent.

While KLA’s ‘patterning’ segment, which accounts for a fifth of total sales, is facing some competitive pressures from businesses in Japan and China, its wafer inspection business (half of all revenues) is growing rapidly thanks to the artificial intelligence (AI) race, with sales up 36 per cent year on year in the third quarter.

The company also has other growth opportunities such as ‘advanced packaging’, as chipmakers now reaching the limits of miniaturisation start to stack chips closer together. Greater complexity will mean a greater need for testing.

Behind all this, though, is a nagging question. Can you be a quality company if you are so dependent on the macroeconomic cycle, and if you are a US business selling into China?

KLA currently expects China’s share of its total sales to fall from 42 per cent this year to around 30 per cent next year, as US restrictions placed on the sale of semiconductor technology reverberate. The slack will be taken up by the US; North America sales doubled in the third quarter as businesses rush to build domestic semiconductor capacity, and now make up 18 per cent of the total.

As to the cycle itself: KLA has a good track record at growing its business through different market conditions, as the chart below highlights. It also has a useful knack of regularly beating analyst estimates come earnings time.

But the fact remains that the company’s shares, like peers', tend to catch a cold when the climate gets frostier elsewhere. Warnings from the likes of Intel (US:INTL) and ASML have hurt in recent weeks; mixed guidance from Applied Materials last Thursday didn’t help either. KLA's shares are now 30 per cent below their July highs.

The question of where exactly we are in the semiconductor cycle is an open one, complicated by an AI boom that has arguably created a two-speed market. With the memory chip market still struggling, analysts at Citi think the cycle is in ‘phase three’, which calls for defensiveness. Even so, KLA remains “the best place to hide” in the equipment sector, in their words. The analysts think KLA’s guidance has de-risked the China issue more effectively than is the case at Lam or Applied Materials, and in particular, cite the company’s focus on the ‘leading edge’ chips that are driving AI investment.

This focus contains risks of its own: a single customer, TSMC, accounts for an ever-larger share of leading edge manufacturing, and on top of this there are clearly questions over where AI appetite goes from here. But KLA’s strong track record and relatively attractive valuation provide a more reassuring way to tap the semiconductor sector.

Investors' Chronicle's Quality US Shares 2024

This falling valuation offers opportunity

This century-old toolmaker has ambitious targets

A new leader is getting this rail stock back on track

Four quality US stocks – and how we found them