Commentary: Why even Toyota, the world’s supply chain maestro, isn’t surviving this crazy chip shortage
Toyota has been making cars for decades and new units roll out of its factories like clockwork. But now a reversal of longstanding supply chain trends may be on the cards, says IMD Business School’s Howard Yu.
LAUSANNE: Last week, various reports said car giant Toyota was suspending production in Japan.
The outbreak of the Delta variant in Thailand and in Malaysia has forced the closure of factories responsible for car parts normally shipped elsewhere for final assembly.
Added on top of the ongoing crisis is the chronic shortage of semiconductors.
This was devastating news for the auto industry.
Toyota, which has thus far been a true master of supply-chain management, had stayed well ahead of its peers. Thanks to the anticipation of a global shortage, Toyota had stockpiled chipsets at the beginning of the pandemic.
Unlike General Motors or Volkswagen, Toyota possessed a formidable online system that had warned its management team. The system contains a comprehensive database that stores supply chain information for around 6,800 parts. Every day, every week, every month, Toyota communicates with thousands of suppliers at all levels.
But despite its preparation and its accurate forecast, Toyota’s inventory of semiconductors eventually ran out.
And so it is set to be the latest company joining the woes of other automakers in missing orders and earnings.
That global shortage of chipsets hasn’t escaped the notice of governments. US vice-president Kamala Harris’s trip to Singapore included a meeting with executives from GlobalFoundries and 3M. She said: "When we look at the disruption to the supply chain, this is an issue that requires all nations ... (to) work together to coordinate."
In other words, shortage of chipsets is no longer just a commercial concern. It’s becoming a national priority.
Further price increases in cars and consumer electronics could threaten the US with unabated inflation, at a time when consumer spending is rising at the fastest pace in 30 years.
HOW PREDICTION BECAME USELESS IN A SEA OF DISRUPTION
Toyota has gotten car manufacturing right - by having its ducks all lined up.
But what Toyota faces is the convergence of several deep problems outside of its core business competencies – geopolitical and pandemic-related.
Governments in Thailand and Malaysia have no choice but to shut down local factories because they have to stem the tide of the deadly Delta variant – something completely random.
China, home to many big chipmakers, has yet to become self-sufficient in semiconductor production. Without access to the latest American technologies because of US-imposed trade restrictions, China simply can’t ramp up production of chipsets as if it were making sneakers.
It’s simply impossible to map out the final outcome when these problems interact. In this case, prediction is useless. And that’s only half of the story.
RIPPLE EFFECT IN THE GLOBAL SUPPLY SYSTEM
The problem is also one on the demand side. When people started working at home last year, they bought more electronic gadgets. Makers of phones, laptops, game consoles thus ramped up production. They ordered more semiconductors.
That created a sudden surge in demand for suppliers like Qualcomm and Nvidia, which design and sell the chips found in everything from Nintendo Switches to iPhones.
Then in the second half of the year, the economy started to recover, and people started to buy cars again.
Long before the Internet economy, researchers at Stanford in the 1990s observed that a small change in consumer demand for baby diapers at Walmart could trigger much larger changes in Walmart’s wholesale orders to Proctor and Gamble (P&G), the company making diapers.
And that set off even bigger swings in P&G’s demand for the input materials like wood pulp and polyester fabric. The small shifts in the demand for an end good always sends ripples in the supply chain like a cowboy cracking a bullwhip.
The hand may move only by inches, but the tip of the whip snaps several feet through the air: What we call the “bullwhip effect” of the supply chain.
In the case of semiconductors, this “bullwhip effect” came from demand for those same chips.
CONCENTRATED IN THE HANDS OF THE FEW
All these suppliers then put in extra orders with upstream suppliers — manufacturers like TSMC in Taiwan or Samsung in South Korea.
But here is the big problem: TSMC alone commands about 50 per cent of the market share of all semiconductors in the world. But semiconductor production is not something you can ramp up quickly.
Setting up a semiconductor factory requires an upfront investment of as much as US$12 billion, according to chipmaker GlobalFoundries.
And like what TSMC has revealed in the past, a semiconductor factory takes three years to become production-ready from construction to groundbreaking.
Yet, just like in a famine, everyone is panic-buying and ordering even more since supply is limited. Every company is trying to stockpile chips to last them through the crisis.
Even highly experienced tech companies such as Nvidia, Microsoft and Apple are struggling to receive a steady supply.
The only winners are, predictably, TSMC and Samsung, who have seen their share prices rise by 190 per cent and 61 per cent respectively in the past 12 months.
HOW TO ESCAPE THE BOOM- AND-BUST CYCLE
Tesla is one company which has found its own solutions. Elon Musk doesn’t expect much from his industry partners. He is rewriting the car firmware himself, so that he can source new type of chipsets altogether.
More specifically, he works directly with TSMC and Samsung, thereby pushing out the old middlemen of all kinds within Tesla’s supply chain.
In other words, Tesla is using its software muscle to take over more functionalities that used to be located in the more purpose-built hardware.
Tesla isn’t reliant on that many more partners to get its car out of the factory. That hardware has historically been put together by auto parts maker Bosch, chip maker Infineon, and many more in a long chain.
In a sense, Tesla is shortening the supply chain but that’s an approach that obviously requires some deep understanding of software programming. And among carmakers, despite everyone talking about autonomous driving and connectivity, only Tesla has the real capability to make it work.
The auto industry illustrates that being reliant on a handful of trusted suppliers for critical parts and functions can yield a system too vulnerable to withstand external shocks.
This conclusion should not be surprising. If everything is optimised to the extreme and running close to 100 per cent utilisation rate, it might be very cost effective, but it will also teeter on the verge of implosion.
This is why Google and Amazon have redundancies in their computer servers around the world. Their customer data is duplicated and stored in multiple locations.
It’s a more expensive approach, but this is how tech giants avoid catastrophic outcomes.
Executives will never yield to Wall Street pressure to eliminate redundancy in order to save those tens of millions in expenses.
In contrast, when it comes to semiconductors, traditional carmakers have been closing their eyes to activities that lie outside their own companies.
They thought that as long as they had a contract in place, they were safe. Except of course, there are times when perfectly drawn contracts are not even enforceable no matter how loud one screams.
All this wasn’t new to the auto industry. The only difference is Toyota’s exhaustive stockpiling was just a few months ago heralded as a key reason for its exceptional escape of this global chip shortage.
That even the mighty have fallen suggests that approach is no guaranteed protection against disruption.
After decades of outsourcing and disintermediation in the pursuit of profit and efficiency, Big Business may now be seeing the start of a reversal of longstanding trends in supply chain management.
Howard Yu is the LEGO Chair Professor of Management and Innovation, and the Director of the Center for Future Readiness at IMD Business School.