DETROIT—STMicroelectronics, a major producer of semiconductors, is benefiting from booming demand for automotive chips. Its revenue this year will grow to $12 billion, a level the company didn't expect to reach until 2023. Like other chip makers, ST, based in Geneva, has been challenged by unexpectedly high demand for semiconductors, particularly from auto makers and Tier 1 suppliers, which account for one-third of its sales.
STMicroelectronics CEO Jean-Marc Chery warns that the chip shortage will last until 2022, although auto makers should get some relief in the second half of this year. Chery, 60, explained why in an interview with Automotive News Europe Associate Publisher and Editor Luca Ciferri and Correspondent Andrea Malan.
Q: A lot is being said about auto makers not having enough microchips, which has affected their vehicle production plans and their bottom lines. But STMicroelectronics and other chip makers are not in crisis. Your company recently reported very good financial results for 2020 and the first quarter of 2021. Why is there such a big difference?
A: I wouldn't classify this situation as a shortage or a crisis. Yes, there is clearly a gap between the demand and the production capacity. However, we have to put the situation in perspective. Last year, the semiconductor industry was a market worth $430 billion, which was a 7 percent increase on 2019. You have an industry which, despite the pandemic, grew by 7 percent and prepared for growth of more than 10 percent in 2021. The problem is that demand has risen by 20 to 25 percent above the planned capacity increase.
Why did demand rise so much more than expected?
The chip industry has been preparing for years to support automotive megatrends connected to the shift toward smart mobility, such as the electrification and digitalization of light vehicles and heavy transport. The sector is also supporting the move toward manufacturing with more energy-efficient systems as well as demand for smart devices, connectivity and the Internet of Things.
One effect of the pandemic was that it pulled forward these megatrends by two years. That means a market that was supposed to reach $580 billion in 2023 or 2024 is already there now. A big reason for this was the shift to remote working, which took place when companies had to send their employees home to protect them from the virus. That boosted demand for PCs, smartphones, tablets and the overall infrastructure needed to support this—cloud computing solutions, larger data servers and so on.
The second reason why demand grew so fast was the new carbon footprint regulations for industrial and high-end consumer products such as white goods and small power tools. Reducing their carbon footprint requires more chips to improve their efficiency and energy management.
Last but not least came the automotive trends toward the electrification and digitalization of the vehicles. Therefore, three planets aligning at the same time, which was not very well anticipated by everybody.
Were there any other factors?
Yes, there are two more. First, the trade war between the U.S. and China has caused some inventory buildup. This isn't for short-term production needs; it is to protect some industries from a potential de facto embargo. The second is the part of the automotive value chain that relies on just-in-time supply of parts. When vehicle demand surged last summer, semiconductor makers were already very busy. The alignment of all these factors created this huge gap between short-term demand and production capacity.
What kind of impact have you felt at STMicroelectronics?
We are all facing the same situation. I can attest that ST reacted very fast. For example, when I made a presentation to the financial community in late December 2020, I told them that given the data available at the time, I believed ST would reach $12 billion in revenue by 2023. That was a big disappointment for them. Now we are on track to reach $12 billion in revenue this year.
How did you react?
After seeing how the different markets we serve moved in December, January and February, we pulled forward a lot of capital expenditures. We raised our investment target for 2021 from $1.5 billion to $1.6 billion to more than $2 billion to build the additional capacity.
But there is a physical limitation. We cannot go beyond that because of equipment lead times and product qualification times.
Given these limitations, how long will there be a gap between supply and demand?
At least another six months.
Is it true the chip industry works with an eight-month lead time? Also, are orders on a "take-or-pay" basis, which means if my company orders 100 chips, we have to pay for 80 upon delivery regardless of whether we take 80?
Let's first try to depict what is a normal situation. The chip industry's average annual compound growth rate is 6 to 7 percent. We have our business plans, we know our projected cash flows and what investments we need to make. With customers, there are normal negotiations. When we are asked for some very special product, we might ask for a take-or-pay contract. Usually, though, that's not the case.
ST's policy is to say: "If you give us one year to 18 months of visibility, we will be better able to manage our capex and manage our risk." In the last few months, because of the surge in demand, we had to try and adapt ourselves in terms of capacity and capital expenditures. We are investing well above our normal run rate. It is normal that we have some price increase in the supply chain.
Is it the case for everybody in the industry?
This was not the case at some semiconductor makers and foundries. Usually foundries—semiconductor contract manufacturers—operated exactly like an integrated manufacturer: They invest on the basis of a normal growth rate and the market is well balanced. But we know that in some cases some of them said: "If you want us to invest more, you have to commit, because we have taken the risk of the investment. You have to pay or to mitigate the risk with take-or-pay contracts."
That was because of the huge surge in demand, which outstripped the compounded average growth rate.
Is it fair to say that when auto makers canceled their orders last spring because of the pandemic, they lost their place in line, leaving chip makers no choice but to say, "Sorry, we can't help," when the auto makers came back asking for chips again?
For the overall chip industry and for foundries such as TSMC, automotive represents roughly 10 percent of their total revenue. Meanwhile, 80 percent goes toward personal electronics and the related infrastructure, and the other 10 percent toward industrial equipment, IoT and health care. When the automotive industry woke up and said, "I want twice as many chips," there was simply no capacity left because so many other industries had already loaded the foundries with work.
How will auto makers get them when it's not possible to increase capacity?
Starting in the second half of this year they might cut supplies to other industries. That means that the chip shortage might start to impact the availability of white goods and power tools.
So if the auto industry's pain is only going to shift to other industries, when will the shortage really end?
Our forecast is that we should start seeing an improvement to the overall situation in the first quarter of 2022. However, it's important—and we are working on this with auto makers and Tier 2 suppliers—to avoid building inventory at the wrong end.
What does that mean?
We know that the automotive supply chain is very complex and fragmented. This makes it necessary to have stockpiles of parts here and there, even for an industry that works according to just-in-time principles.
We have to avoid building excessive inventory and better align the forecasts for car production and component production. The current period is already difficult enough because of the overall geopolitical and economic behavior.
Could you elaborate?
There are so many companies saying they want to have their own "fabs" to manufacture their own semiconductor designs. There is constant background noise related to trade conflicts. Europe, the U.S., China all say they want to be more independent when it comes to semiconductors. All this turmoil is not helping us operate smoothly.
Is ST ready to stop working with automakers if they ask the company to build up too much inventory?
If we have to choose between building up inventory and losing our customer, we will not lose our customer. We will try to find a middle ground that works for both parties.
Is the current gap between supply and demand partially because the auto industry uses less technologically advanced chips?
Ten years ago, the auto industry was not among the early adopters of new technologies. That is no longer the case. Auto makers have become pathfinders of new technology because of the accelerated transformation toward digitalization and electrification. A complex visual processor for ADAS (advanced driver-assistance systems) has chips that are at the same level of those used in the flagship smartphones of the latest generations. The inverter in an electric car is the most advanced technology found in any power device. Therefore, the automotive industry is at the same level as the world's most advanced industries in terms of the complexity of technology.
The auto industry accounts for 10 percent of global chip production. Will this grow significantly with the move to electric and autonomous vehicles?
The global market of semiconductors will double between now and 2030, rising from $520 to $530 billion to $1 trillion by that time. While the semiconductor content in an electric car and one with Level 2 automation will double, the relative weight of the automotive industry as a whole will remain more or less the same, at around 10 to 12 percent.