The World’s Most Important Product and the Only Two Companies That Can Make It

All of the chips that go into the world’s most advanced computers are made by a single company – Taiwan Semiconductor Manufacturing Company. And TSMC cannot produce these advanced chips without a stunningly complex machine that only one other company, ASML of the Netherlands, can make.

The uber importance of these two companies creates critical vulnerabilities in a critical supply chain. Moreover, both companies lie in countries allied to the West, an uncomfortable and untenable situation for China. Of course, China does not consider Taiwan to be a separate country and reunifying with it is one of China’s central long-term goals. It’s pretty easy to see the potential for conflict.

The story of how we got to this point is told in a readable and eye-opening new book called Chip War: The Battle For The World’s Most Important Technology. I interviewed the book’s author, Chris Miller, on this Ideas Lab podcast series. The conversation is fascinating – to get a flavor here are some of the book’s key themes:

1. Stealing and copying technology is a poor strategy.

The Soviet Union was good at stealing technology but that never led to an ability to innovate. By the time it had copied and built a chip, the US had moved on, ensuring Soviet technology was always years behind the West. To this day Russia’s military is hampered by an inability of the country to produce the advanced chips needed for things like smart missiles. China is also good at stealing and copying, but this has not helped it create the software or manufacturing capability to produce advanced chips – all of which it must buy from the West.

2. Specialization creates fragility, but also huge benefits.

The risks that come from hyper-specialization are a big problem and it’s easy to criticize the loss of the ability to manufacture advanced chips in the US. But outsourcing manufacturing allowed US companies to specialize themselves – particularly in the software used to design chips. And specialization drove down costs, making technology available to far more people than it otherwise would.

3. The ‘free market’ in chips has always relied on governments.

The chip industry is stunningly innovative. The capabilities of ASML’s machines feel almost godlike – a true triumph of human capability. This creativity has been driven by free market forces, but the industry has always need government support.

The US military was the first customer for US chip firms. Governments in Japan, Korea, Taiwan and China have all funded research, been strategic customers and subsidized their domestic chip companies.

We need the innovation that free market competition engenders but the resulting specialization has created companies so important that their future is not going to be left to the market.

4. Japan and China ALMOST followed the same path.

Japan developed economically by being a low-cost producer of basic items and was encouraged by the US to integrate into the global trading system. It benefited from government subsidies and wasn’t afraid to steal intellectual property. Later its firms moved up the value chain, displacing established western companies and eventually causing panic that it was “overtaking” the US. Along the way it was helped by a financial bubble that created cheap capital and a huge build up of domestic debts.

Sound familiar?

China took a similar path several decades later. The difference of course is political. Japan became a democratic society. For a while it looked like China was moving (slowly) in that direction. Had it continued I believe the US would have been happy to allow globalization and integration to continue apace. Instead, China veered back to authoritarianism and it feels like we are destined for de-coupling and the construction of two separate and more expensive technology ecosystems.

Keep Your Friends Close, But Your Geopolitical Competitors Closer?

Could the two countries get past their different political systems? Instead of building separate, competing supply chains why not deepen integration so that both the US and China would have too much as stake to go to war? This is a technological version of James Fok’s idea of financial mutually assured destruction, which he proposes in his book Financial Cold War.

I asked Miller’s about this during our conversation – he was skeptical:

“Historically it hasn’t worked. Britain and Germany were major trading partners right up until the outbreak of WWI and immediately they flipped and Britain launched a very aggressive blockade to try and starve the German people.”

“(more recently) …the failure of Germany’s strategy to integrate deeply with Russia in economic terms has not provided the results that were hoped for”.

“I hope we are in a state of mutually assured destruction, but I worry that it’s not true. And that worry is what inspired me to write Chip War.”

The New World Is Already Being Built

The Chips and Science Act passed by the US this year is largest ever public investment in R&D. Last month I wrote about how economist Richard Duncan thinks it could be a template for a much larger effort to develop next generation technologies. And just last week TSMC founder Morris Chang, a central figure in the book, was in Arizona with President Biden announcing plans for his company to invest up to $40 billion in two new plants. A step…but just a first one. Miller thinks we’ll still be relying on Taiwan for a lot of chipmaking capacity in five years time.

Meanwhile, Japan and the Netherlands have agreed in principle to join US efforts to ban the export of advance chip-making equipment to China. But it won’t be without struggle. ASML’s CEO is clearly frustrated at US pressure to cut off technology exports to China saying his company has “already sacrificed”. I suspect he’ll lose this battle as the Chip War begins.