Trade war and Moore’s law combine to raise the heat for manufacturers

William Pesek

Good news for Toyota Motor may be a bad omen for the global economy in 2019.

Taiwan Semiconductor Manufacturing Co. told its executives around the globe in early Dec. to do without the luxury Mercedes-Benz sedans that long transported them to work. Toyota Camrys would have to do instead. Plus less business travel in general.

Corporate belt-tightening is hardly unique. Yet sometimes an anecdote encapsulates an ominous turning point in a volatile industry that is key for the world economy. Here, TSMC seems a microcosm not just of economic risks, but also the limits of physics bedeviling the sector.

It is the world’s biggest contract chipmaker, a major Apple supplier and a proud innovator. Coming just amid Donald Trump’s trade war and slowing Chinese growth, TSMC faces a wave of corporate austerity not seen in a decade.

Data provider World Semiconductor Trade Statistics recently slashed its 2019 semiconductor growth forecast to 2.6% from 4.4%. Investment broker CLSA is far more bearish, predicting a drop of up to 5%. TSMC’s sales growth outlook: now roughly 50% below what was expected at the start of the year.

Capital expenditure among semiconductor makers is seen retreating 12% in 2019, says IC Insights. Not promising for an industry with wares powering everything from smartphones to video games, spreadsheets, weather forecasts, cryptocurrencies and data processing.

How bad might things get in 2019? Bad, as China retaliates against U.S. President Donald Trump’s escalating tariffs offensive. Yet there’s a bigger worry dogging the Silicon Valley set: are the chip market’s woes merely cyclical or something darker?

There are really two variables for investors scratching around for answers — one economic, one scientific.

The economic risks are all over front pages as Trump tackles Xi Jinping’s “Made in China 2025” dream. The Chinese president is plowing hundreds of billions of dollars in a bid to become a world leader in computing, self-driving vehicles, biochemicals, pharmaceuticals, renewable energy and robots. The Trump White House is saying “not so fast.”

Trump’s tariffs on steel (25%), aluminum (10%) and $250 billion of Chinese imports (25%) are already slamming Asia’s biggest economy. Data on exports, fixed-asset investment and purchasing managers’ orders all point to a rough 2019.

The White House also seems to be getting creative in efforts to stop China. Case in point: the arrest of Meng Wanzhou, chief financial officer of smartphone giant Huawei Technologies in Canada at the behest of U.S. authorities. Reports suggest Huawei is accused of violating U.S. sanctions against Iran. Either way, Trump seems to be taking off the gloves and going after a key Chinese innovator — adding a tech war to the trade conflict.

What is clear is that Trump’s tariffs are biting. The signals from TSMC, the sole supplier of iPhone core processors, illustrate the trend. TSMC’s chips are the nucleus of the digital economy.

Bitcoin’s travails are another wrinkle. In October, when TSMC cut its 2018 revenue target for a third time since January, Chief Financial Officer Lora Ho blamed “weakness in cryptocurrency mining demand” along with shifts in “inventory management.” Cryptocurrencies are down by up to 90% this year.

TSMC is hardly alone in battening down the hatches. Taiwan’s Foxconn, major iPhone assembler, and its subsidiaries are cutting 100,000 jobs by the end of 2018, Nikkei Asian Review reported in November.

Largan Precision, an iPhone camera provider, reported a 29% drop in revenue in November year-on-year. Catcher Technology, a key metal frame supplier, suffered a 13% year-on-year sales decline in the same month. Lackluster demand for smartphones has companies from Qorvo to Lumentum to AMS corroborating the darkening outlook that has TSMC pivoting from Mercedes-Benz to Toyota.

And yet this might pale in comparison to the chip industry’s longer-term problem: the breakdown of Moore’s Law, long the idea driving Silicon Valley forward. Named after Intel co-founder Gordon Moore, it holds that processor performance doubles roughly every two years. Now, the tech world’s governing principle is having a sort of midlife crisis.

In September 2017, Nvidia Chief Executive Officer Jen-Hsun Huang engaged in a bit of heresy when he all but declared the 53-year old dynamic dead. As 2019 approaches, the view is gaining wider credence as technology struggles meet trade headwinds.

One problem: geometric scaling has slowed down. In other words, tech companies’ ability to make better, faster processors requires shrinking them to fit into devices. They also must be able to withstand heat given off by ever-faster CPU speeds. That means higher costs for packaging and cooling.

For this work, multinationals have become highly reliant on Chinese assemblers, many in the Yangzi delta near Shanghai. With these companies facing the trade war just when they are reaching close to what looks like a technological limit, the outlook is as muddy as the Yangzi’s waters.

This smacks of opportunity, of course. You just need a great new idea. “The problem of device physics is a solvable one, and that is the silver lining to the death of Moore’s Law,” Radoslav Danilak, chief executive of San Francisco developer Tachyum, wrote in a March Forbes op-ed. “The industry’s performance plateau creates a market space and opportunity for new ways of thinking, new designs and new inventions.”

The trouble for investors, though, is that supply chains do not turn on a dime. As designers do their thing, short-to-medium term profits may come under increasing pressure. Also, even when eureka moments emerge, recalibrating a sprawling and geopolitically-sensitive industry involving tens of thousands of suppliers takes time. So does turning a profit.

Here again, TSMC is a microcosm. Its latest 7-nanometer chips proved to be an instant success. In the third quarter, they generated 11% of revenues. The problem, though, is reduced profit margins. It is a hint of how the high costs of designing, producing and introducing more advanced technologies might continue to pressure profits. It is worse if fewer consumers are buying top iPhones — the end product.

This does not mean investors should give up on chipmakers. But it does mean the sight more Toyotas, and fewer Mercedes-Benz cars, augurs poorly for profits in 2019 — and even beyond. By 2020, tech executives may traveling by Uber.

William Pesek is an award-winning Tokyo-based journalist and author of “Japanization: What the World Can Learn from Japan’s Lost Decades.” He was given the 2018 prize for excellence in opinion writing by the Society of Publishers in Asia, for his work for the Nikkei Asian Review.

What's your take on this?