-- A worldwide shortage of semiconductors has literally wrecked entire parts of the global supply chain, but somehow, contrary to the automotive sector, the aerospace & defense industry is so far only marginally impacted. But if it does not end soon, something no one seriously expects now, it may disrupt certain productions.
-- However, maybe more importantly, the supply chain crisis also revealed a serious sovereignty issue for Western powers, in terms of security of supply for sensitive components.
PARIS --- The world’s dependence on semiconductors came into sharp public focus in 2021, in particular when entire parts of the automotive manufacturing ground to a halt because of massive computer chip shortages. However, such a disruption was not observed for the Aerospace industry. There are two main reasons for that.
Number one: About one thousand electronic chips are needed to manufacture a modern, mid-range thermal ground vehicle today, and this figure goes up for hybrid/electric vehicles, and exponentially up for self-driving cars, due to the massive computing power needed to process the data coming from multiple sensors in real time.
However, this number falls to a hundred for most products of the aerospace & defense sector...
Another issue is that – believe it or not – chips aboard most commercial airliners were designed some 20 years ago, if not before and most of them will not be changed until the aircraft retires, whereas the chips in your car are regularly updated, whenever a new model, with new embedded electronics, hits the market. You don’t upgrade airliners or even military aircraft that often! And right now, if upgrades are much more frequent than in the past, most of them are only software-driven and rarely require brand new hardware. In addition, safety and reliability guarantees required by civil and military certification processes are much more stringent in the Aerospace business, and any hardware change is inherently long and costly to be implemented…
Number two: The production rates are very different. Automotive OEMs count in millions of units per year, usually in a just-in-time production mode, while most A&D players count in hundreds or a few thousands, with monthly rates of Airbus and Boeing airliners stand between 30 to 50 and up to 60 aircraft/month, at full speed when everything goes right…. So, both the nature and numbers of required chips – as well as stock management – are totally different…
Sovereignty becomes a major issue
Another issue is sovereignty. Following massive offshoring over the last decades, 70% of all chips globally are fabricated in Taiwan and South-Korea, and that figure rises to about 90% for the most sophisticated chips, such as those based on gallium nitride and gallium arsenide, which are used in the defense and aerospace sector. That is one of the reasons why Taiwan Semiconductor Manufacturing Company (TSMC)- a global leading chip manufacturer - has emerged as a flash point in the tensions between the US and China. The company has even been pressured to produce its military-grade chips in the United States. As the CEO of Northrop Grumman, Kathy Warden puts it “having the ability to control a supply chain in semiconductor and chip production, in my view, could become a national imperative”. And, indeed, amid the current crisis, semiconductors and their associated design and production units are now considered a globally-critical industry.
How did the shortage come about?
The reasons for the supply shortage are as multifaceted as the nearly $500 billion semiconductor business itself. First, the Covid-19 crisis slowed down the production of foundries while the demand for electronic products (laptops, televisions, game consoles) and other connected objects soared. At the same time, the demand for electronic chips was also boosted by the deployment of 5G technology, which involves the production of new antennas and new smartphones. In addition, the rapid recovery in Chinese domestic consumption after the first epidemic wave did not help stabilize the market.
Shortages have been exacerbated by several catastrophes that include a fire at a Renesas Electronics chip factory in Japan, a drought in Taiwan and a cold snap in Texas that temporarily shut down factories operated by Samsung Electronics, NXP Semiconductors and Infineon.
Just-in-time production, which helps to keep costs low, also provides little insurance against such disruptions, especially because making a semiconductor is a highly complex manufacturing process. According to the Semiconductor Industry Association (SIA), the lead time - from when a customer places an order to when he receives the final product - can take “up to a total of 26 weeks”. In the automotive sector, average delivery lead times rose from 6 weeks, to 20 then 40 in December 2021.
As a result, suppliers couldn’t keep up with demand for microelectronics that drives everything from smartphones, gaming devices and domestic appliances to industry robotics new vehicle components, computers; This in turn has an impact on data centers that power cloud computing, artificial intelligence, and Internet of Things (IoT) advancements supposed to bring us into a connected future, not to mention the… metaverse, whatever that means or will mean in the future)!
According to figures published last November by the World Semiconductor Trade Statistics (WSTS), the global semiconductor market growth rose from 6.8% in 2020 to 25.6 % in 2021, reaching a market size of $553 billion. This will be the biggest step-up, since a 31.8% increase in 2010, eleven years ago.
Demand explodes, but EU and US are marginal players
While global demand for semiconductors has exploded, the role of the EU and US in designing and manufacturing them has decreased. For example, the US share of global semiconductor fabrication has dropped to 12% today, compared to 37% in 1990, according to the US-based Semiconductor industry association (SIA). American chip companies now almost exclusively rely on Asian contractors for advanced processes, and the US share of global capacity is forecast to drop to 10% by 2030, while Asia’s will climb to 83%. The situation is worse in Europe, where 90% of the semiconductor chips come from Taiwan and Korea since European companies such as German Infineon Technologies, Franco Italian Microelectronics and Dutch NXP moved their fabs across the Pacific.
As Franz-Stefan Gady, a fellow with the International Institute for Strategic Studies, explains: European defense contractors have preferred to establish business relationships with overseas chip suppliers such as Samsung or TSMC rather than keeping suppliers in business on their own continent. As a result, “European advanced semiconductor manufacturers produce low-volume niche products with limited military applicability” Gady says.
Although the military & aerospace semiconductor market is expected to grow by $3.89 billion between 2020 to 2025 according to Market forecast consultancy Tecnavio, it’s only a drop in the bucket compared to the global commercial semiconductor market, which is a hundred times bigger. Due to safety/security requirements, certification processes, small production batches and long life-cycles, chips used inside commercial airliners as well as in most military platforms are usually venerable and well-proven. Apart from a few exquisite systems, such as the new adjunct processor incorporated in the Boeing’s Super Hornet Block III, which allows the onboard computer to process and fuse more data in far less time, increasing a pilot’s situational awareness.
In the A&D sector, the demand for semiconductors will be driven by aircraft manufacturers, and more specifically by aircraft electronics systems.
A&D sector begins to feel the pinch
Manufacturers will have to upgrade and modernize aircraft, increase fuel savings and lower emissions. Actually, major OEMs like Boeing and Airbus already boosted their demand to avoid any supply shortage. As A&D procurement cycles are much longer than consumer electronics or automotive, the sector avoided the scale of supply woes faced by auto makers and machinery companies. Also, aircraft manufacturers were producing fewer jets than before the pandemic. But by the end of 2021, as Airbus CEO Guillaume Faury confided there was “mounting pressure on commercial aerospace's supply chain”. As Eric Bernardini, global co-head of aerospace, defense and aviation at consultants AlixPartners detailed: “supply chain strains are increasingly becoming visible for the production of narrow-body jets, which have seen a pick-up in demand due to a recovery in short-haul trips”. In the long-run, the shortages could delay and increase the costs of programs such as next-generation fighters, satellites, unmanned drones, and other sophisticated sensors, all interconnected with each other.
At the same time, acquiring advanced semiconductor devices is being undermined by the systemic changes in the semiconductor industry, which is now driven by commercial demands for consumer products. The car industry, as well as telecommunications, is increasingly buying semiconductors that used to be produced for the aerospace and defense sector. For example, gallium arsenide and gallium nitride-based chips used in radiofrequency integrated and monolithic microwave integrated in military communications, space capabilities or active electronically scanned antenna (AESA) are now needed to produce 5G electronics.
Industry raises concerns
These challenges have raised concerns among the industry with 59 CEOs, including BAE Systems Inc., Infineon Technologies Americas Corp, GlobalFoundries, Inc. or Stellantis writing to the US Congress last December to urge for political action to support, design, research, and manufacturing of semiconductors on US soil.
One risk stems from the fact that the A&D sector has been increasingly incorporating commercial semiconductors in its supply chain. For example, Field-Programmable Gate Arrays (FPGA) are used in military systems due to their low cost and high modularity. Nevertheless, some A&D chips need to be more durable, with a 100% reliability and have a higher heat or radiation tolerance. Those are used in demanding electromagnetic operations, signals intelligence, military communications, space sensors or radar applications. They belong to the compound semiconductors category which includes gallium arsenide and gallium nitride. Most of those are designed by US or European chipmakers but fabricated in overseas foundries, primarily by TSMC in Taiwan and Samsung in South Korea. When it comes to gallium arsenide foundry market, Taiwanese firms hold over 90% of the market. They are now also in high demand due to 5G infrastructure deployment currently happening across the world, creating potential bottlenecks and rival needs…
In 2004, after the Senate Armed Services Committee investigated counterfeit electronic parts in the US military and found 1 million bogus parts, including sensitive components aboard several types of combat platforms, the Department of Defense (DoD) established the “Trusted Foundry Program”. Since then, as part of the Defense Micro-Electronics Activity (DMEA), a DoD laboratory identifies companies deemed secure and trustworthy enough to produce chips exclusively for the military. They assess the integrity of people and processes involved in the production of microelectronics based on four standards. Trusted sources are asked to: (1) provide an assured "chain of custody" for both classified and unclassified integrated circuits (ICs); (2) ensure that there will not be any reasonable threats related to disruption in supply; (3) prevent intentional or unintentional modification or tampering of the ICs; and (4) protect the ICs from unauthorized attempts at reverse engineering, exposure of functionality, or evaluation of their possible vulnerabilities.
According to Republican Senator Roy Blunt, two facilities currently operate under this program, one in Vermont and one in New York, but they only produce a small percentage of the nearly 2 billion semiconductors the DoD acquires each year. Now, both the Congress and the DoD want to move away from the trusted foundry model and instead devise ways to ensure that the chips themselves, wherever they come from, are certified in tests to be trustworthy and reliable. Victoria Coleman, the Air Force’s chief scientist, recently explained that “the focus of the new approach is less on who makes the product and more on whether it meets defined standards — measures that are still being developed”.
New production plants could ease shortages
In the meantime, several new plants should open in the coming months or years in the United States. Texas Instruments is spending 9 million to open two fabs in Texas and Utah that it bought from Micron Technology in 2021 and TSMC, after being pressured by the DoD last year, is building one in Arizona and Samsung in Texas.
According to Marketwatch expert Brad Slingerlend, these initiatives will not significantly impact the supply chain in the short term. The Arizona foundry will have an annual capacity of about 240,000 wafers, less than 2% of TSMC output of 12 million wafers. In addition, the technology deployed at the fab will be several years behind what TSMC and Samsung will offer in Asia. As the US have decided to relocate the production as much as possible, it is not sure that it will be economically sound.
According to a study jointly published last year by the SIA and the Boston Consulting Group, “the 10-year total cost of ownership of a new front-end fabrication facility in the United States is 30% higher than in Taiwan, South Korea and Singapore, and 37% to 50 % higher than in China. In addition, the 10-year cost of a state-of-the-art fab, including both initial investment and annual operating costs, ranges between $10 billion and $40 billion.
Just last week however, Intel announced an initial $20 billion investment to develop and manufacture semiconductor chips at two plants in Ohio. A first step that could actually grow up to $100 billion according to Intel CEO, and become the largest chipmaking complex in the world if Congress passes stalled legislation to support the industry.
It is not sure that raising production in the EU or the US is the answer in that field, and China knows that already! Although it spent billions in the late 2010s to develop its own indigenous semiconductors industry, at least six major Chinese chip-building projects failed over the past three years, underlining the difficulty and high costs associated with developing high-end chips! Today, it manufactures only about 15% of its annual consumption at home and depends on producers in Japan or Taiwan. In 2020, it imported $350 billion, more than its oil purchases.
To catch up in the field of semiconductors, Beijing is quietly trying to increase acquisitions abroad, hiding behind shell companies. In May 2021, the Chinese investment fund Wise Road Capital filled a takeover offer for the French company Unity Semiconductor SAS. And it was not the first time Wise Road Capital was trying to purchase a European company. According to the South China Morning Post, several research or production centers that fell into its own hands have been transferred to China. Just a month after its takeover, the Singaporean company United Test and Assembly Center announced the construction of a factory in China's northeastern province of Shandong. The German Huba Control, acquired a month earlier from Siemens, suffered the same fate. Part of its production has been relocated to the Sichuan province…
With companies calling for governments support, initiatives are multiplying. The US Senate already passed a $52 billion dollars Chips Act to boost domestic semiconductors R&D and manufacturing, and the Biden administration is considering invoking the Defense Production Act to force companies in the semiconductor supply chain to provide information on inventory and sales of chips.
In October 2021, EU leaders announced their own European Chips Act to support increased research, design and testing capacity as well as ensure national investments are coordinated with those of the broader union. Thierry Breton, the European commissioner for internal markets, detailed that the Act would set a semiconductor research strategy, bring together European chip production efforts, as well as provide a framework for international cooperation and partnership. Under the EU's Covid-19 economic recovery package, it has already pledged $150 billion to a “Digital Compass” initiative that will improve member states digital infrastructure, and digital skills. As part of that, it planned to produce at least 20 percent of the world's semiconductors (by value) by 2030, up from 10 percent last year”.
It is not sure that governments will be able to solve the semiconductor crisis on time as the supply chain shortage is probably temporary, given the amount of investments made by various countries. In addition, there remains little room for maneuver especially when one looks at the other side of the chip: Asia.
South Korea will spend $451 billion on becoming a semiconductor manufacturing giant. In Taiwan, TSMC alone has announced on January 14th that it sets this year's capital expenditure budget at $40 billion to $44 billion, even higher than last year's $30 billion! What’s more, it’s not clear that the strategic imperative that motivates the great return of public policies will have a positive impact on the sector, or if, on the contrary, it will undermine its innovative capacity.
Whether state support for semiconductors industry re-shoring will be fast enough to recover a genuine security of supply as well as a credible innovation power remains open for debate.
What is sure is that Aerospace & Defense players are now watching much more closely this part of their supply chain, on both sides of the Atlantic.