What do these cars have in common?
Probably more than you think.
For one, they were all designed by the renowned designer Frank Stephenson, who has also worked for the likes of Ferrari, Ford, BMW, and Maserati.
But maybe more surprisingly, they share quite a bit under the hood too. That’s because, despite belonging to different brands and market segments, all three of these cars are likely facing production delays due to a global shortage semiconductor chips.
Semiconductors are present in almost all electronics and have been used in cars’ internal computers for decades. What is different today is that the chips going into cars are crucial components for advanced features like touch screens, navigation systems, driver-assist features and more. Concurrently, the adoption of electric vehicles is starting to accelerate meaningfully, driving further chip demand in the sector.
The chips necessary for the cars of today are the same ones that power our smartphones and tablets, and other high-end electronics that are ubiquitous in our modern world. And for all of what we do see and interact with regularly, there is even more behind the scenes that powers the infrastructure and enables the software that keeps our digital world running.
All of this is to say is that technology has become an essential feature of products and services across the economic spectrum, and this trend is only accelerating. And semiconductors are central to all the consumer, enterprise and military electronics and services that we increasingly rely on.
Understanding the Semiconductor Industry
While the technology itself is interesting and worth learning about, I will leave the technical aspects to the experts. What is important to note though, is that the production process for semiconductors is exorbitantly expensive and requires exact precision. These constraints are important in understanding how the industry has developed and what are the implications.
The industry is best understood if broken down into a few different segments:
There are the traditional companies that both design and manufacture their own chips, otherwise known as integrated device manufacturers. The most well-known companies like this are Samsung, Intel, and Texas Instruments.
There are also “fabless” firms, companies that focus only on chip design while outsourcing their manufacturing to others. Qualcomm (likely responsible for designing the chip in your phone), Nvidia and AMD (both best known for high-end graphics processing units or GPUs) are some great examples. Apple too has recently made headlines for their newly designed M1 chip that is making a lot of people at Intel and their shareholders nervous.
Next are the “foundries” that focus on the actual manufacturing of semiconductor chips. Samsung is an important player in the foundry business, but by far the most important chip manufacturer is Taiwan Semiconductor Manufacturing Co. (TSMC). This company is responsible for manufacturing all the high-end chips for Apple, Nvidia, Qualcomm and others. TSMC has been extremely successful as a pure play foundry on the guarantee that they will never compete with any of their customers and their technical ability to efficiently mass produce the most cutting-edge designs.
The last pieces of the puzzle are the semiconductor design software and manufacturing equipment necessary to make it all work. The best-known software firms are Synopsys and Cadence, while companies like Applied Materials, KLA, Lam Research and ASML make the manufacturing equipment.
The massive expense, complexity, and scale required to manufacture semiconductors and continue innovation in the industry is truly mind-blowing. Each semiconductor contains billions of transistors that need to be made perfectly so the chips can function properly. It is in this area that TSMC has made a name for itself with its years spent perfecting the manufacturing process and investing in the most advanced machinery. In 2021 projects to spend between $25 and $28 billion on capital expenditures, the majority of which will be allocated towards leading edge capacity. A sizeable portion of that money will most definitely be spent on acquiring the most advanced extreme ultraviolet (EUV) lithography devices from ASML – the only company capable of making them – for around $120 million per machine.
Big Trouble in the Little South China Sea?
Given the strategic importance of semiconductors for the most advanced technologies, the industry is also central to the increasingly contentious relationship between the U.S. and China. As the global semiconductor industry is heavily reliant on U.S. chip design software and manufacturing equipment, the U.S. government has been exerting their leverage over the past few years in attempting to limit China’s access to semiconductors. The Trump administration banned the sale of advanced U.S. chips to Chinese supercomputer labs and Chinese telecom players Huawei and ZTE. It also put pressure on the ASML in early 2020 to prevent the sale of its latest EUV technology to China. These restrictions are likely to continue in some fashion under the new administration.
This is a big problem for China. Trying to catch up in the chip business even without these hurdles would be difficult task. While the money and the will to develop the chip industry is certainly there, the production process is not like baking a cake. The precision manufacturing of companies is a result of years of experience in developing and implementing the necessary techniques. Coercing talented engineers and hacking to steal intellectual property can help, but there are certain parts of the process that virtually cannot be imitated. Especially if you do not have access to the software and equipment that you need.
That brings us back to TSMC. Because TSMC has been so good at manufacturing, chip designers across the world, from start-ups to global heavyweights like Apple, have foregone building their own manufacturing capacity and outsourced to Taiwan instead. Not only is TSMC central to the industry, responsible for nearly 52% of the global foundry business, but it is only 180 km away from Mainland China. Given the heightened tensions, this proximity has raised more than a few concerns.
For one, given the amount of capital expenditure and know-how necessary to set up semiconductor manufacturing capacity, it is not as if the U.S. has any way to quickly ramp up production in case China is able to cut off access to Taiwan. It would take years, if not decades, to replace these production capabilities, especially for the most advanced chips. It also does not help that the world’s second largest semiconductor manufacturer is Samsung in South Korea.
At the same time, the globalized nature of the semiconductor supply chain means that any such move would immediately limit their access to other necessary components. Even if China were able to block access to Taiwan and seize the means of production, they would have trouble doing so.
And it is well understood that by China that if the situation were to devolve then any such move would be met with a resounding response by the U.S. and vice versa. Cutting off access to semiconductors to such an extent would be akin to when the U.S. placed an embargo on oil and gasoline exports to Japan in August 1941, the country’s most important import with 80% of supply coming from the U.S. We know what happened next.
Which is why I do not think that it is likely to happen. Neither China nor the U.S. nor the rest of the world can afford it. Tensions are likely to continue and supply chains reorganized to meet the new realities of global power politics, but secular technological trends of this magnitude cannot be stopped. Just look at how global semiconductor index has performed over the past five years:
Despite the trade war that started in 2018, growth in the semiconductor sector continued unabated. And like the tech sector in general, semiconductor companies saw strong growth in both topline and profitability in 2020, thanks to the pandemic-accelerated trends in data, servers and digitalization. Given the planned capital expenditures by companies like TSMC and Samsung (which plans to double its leading edge foundry capacity in 2021) going forward, its clear that the industry does not expect any potential restrictions to impact the industry over the long term. Financial market participants clearly agree.
The reality is that semiconductors are now the indispensable building blocks to our always-connected world. If previously driven by the 12-18 month demand cycle for PCs and laptops, semiconductors have now entered a secular growth cycle where there is exceeding demand for advanced chips for exciting technologies like 5G, artificial intelligence, high-end computing, big data, cloud computing and more. These applications in turn are helping accelerate advances in biotechnology and medicine, creating novel business models and industries, and completely upending the way the old economy works along the way. Just ask the automakers.