ATREG’s President and CEO Stephen Rothrock and Yole Group’s President and CEO Jean-Christophe Eloy consider the state of the global semiconductor fab market in a post-COVID world.
The worldwide pandemic initially diverted investment from fab transactions, but the resilient semiconductor industry, struggling with supply chain constraints and skilled labor shortages, has received renewed interest thanks to the potential financial support of U.S. and European CHIPS legislation. Stephen and Jean-Christophe discuss the emerging business models that could enable the global semiconductor industry to to migrate to leading-edge and mature technology with optimal manufacturing capacity.
During COVID, the constrained supply chain has caused a significant rise in prices in all areas of the semiconductor world, except for the memory business. Semiconductor sales growth of 25% in 2021 is expected to fall to 11% this year, driven partly by price increases. The automotive sector in particular grew significantly because of the fast development of electric vehicle technology, but strong demand has also been pushing up prices in the industrial and data center markets. Another factor for lower growth is inflation which has resulted in less orders for smartphones and other consumer products. There have also been blockages in the semiconductor supply chain as a result of the war in Ukraine, coupled with growth in hyperscalers, start-up companies, and artificial intelligence which are all becoming more impactful areas for the industry and therefore pushing the need for more microchips around the world.
There is no doubt that recessionary winds are blowing, compelling chip companies to consider this when looking at ramping their wafer production. However, a lower growth rate does not necessarily mean a decline in the manufacturing needs many companies have for their products. Demand remains strong and both IDMs and foundries need to find ways to increase fab capacity and meet the current and future challenges that lie ahead.
Although the semiconductor industry is able to grow very fast, the need for more capacity and the significant amount of investment associated with it remain difficult when budgets are limited. The resilience of chip makers’ supply chains will be tested when trying to quickly adapt in the current economic and geopolitical environment. Monitoring and identifying weak spots and the industry’s building strategies to address these will be important in the months ahead.
An increasing number of companies are interested in end-to-end design and manufacturing capabilities. Leading-edge technology is going to benefit extensive R&D programs, but the key issue has been the lack of wafer fabs at this level. One important reason for the lack of fabs worldwide is the significant cost required to build them and bring them online. Considering that a leading-edge fab at 5nm costs somewhere in the vicinity of $5 billion+ and a little less than $2 billion at 10nm, the cost differential is huge. Government incentives will undoubtedly help, but while China’s has dedicated some $73 billion in semiconductor subsidies so far (not accounting for government grants, equity investments, and low-interest loans exceeding $50 billion alone), the U.S., Europe, and Japan have not been able to come even close to matching that figure. The impending European CHIPS Act and the recently passed U.S. one mean we are likely to see more leading-edge capabilities developing around the world. Korea and Japan are also adopting their own financial support legislation. Even with all these initiatives, it will probably be three years before we start seeing that manufacturing capacity hit the market.
There has been no major consolidation of investment in the last three quarters, a delay leading everyone to play catch-up and invest. There is a need to further expand production because of the need for more leading-edge devices. Worldwide supply is limited with just a little over 150 300mm fabs around the world, with 42 of them located in Taiwan, 33 in China, 19 in the U.S., and just 12 in Europe and the Middle East. While leading edge is important and attracts a lot of investment, a huge portion of the activity remains at 200mm, with significant demand for different nodes at 90nm up to 180nm processes. There are around 230 200mm fabs worldwide, 51 of which are located in the U.S. and 49 in Europe and the Middle East.
There needs to be a balance between the new fabs being built today at advanced geometries against the long-term need for production at less advanced nodes. Texas Instruments is an example of a company who has stated the goal of building new 300mm fabs to serve their analog production needs at less than cutting-edge nodes. The objective here is to get the full benefit of these rare brownfield 200mm manufacturing sites. This is also an important trend for the industry because it means a major reorganization of production brought about by the difficulty in accessing 200mm equipment.
Many of the largest companies are primarily moving to leading edge as evidenced by Intel’s announcement to build in Ohio, following upgrades to its fabs in Phoenix, Israel, and Ireland. The level of investment announcements has been significant – Micron’s $100 billion plan in Central New York, Samsung’s and Texas Instrument’s greenfields in Texas , or BOSCH, Infineon Technologies and STMicroelectronics’ investments in 300mm in Europe. It is important to note though, that not all of these 300mm fabs are leading edge – many are 14nm and above.
There are geopolitical factors at play too. Taiwan has overseen TSMC’s rise to the point that the company now controls more than 55% of global foundry manufacturing today. This means a lot of smaller companies are now wondering what will happen to them as such large players as Apple and Qualcomm dominate demand within the foundries. This concern leads to them looking for their own fab manufacturing opportunities to become less reliant on others and ensure the security of their supply chain.
Alongside this, a trend for power electronics using compound semiconductors presents an opportunity for the semiconductor industry. We are seeing a big shift in investment into GaA, SiC, and GaN manufacturing facilities. For example, Wolfspeed has built a leading-edge greenfield silicon carbide fab in New York’s Mohawk Valley and recently announced a new materials facility in North Carolina to support the SiC business. OSRAM has also announced an $850 million investment in an eight-inch facility for microLEDs. Compound semiconductor investments such as these will significantly expand the available capacity of these products to serve end-markets such as mobile phones, 5G networks, and electric vehicles.
It typically takes 18 to 24 months to build a fab shell, install the equipment, and then another 18 months to get the tool line operational and qualify product for volume production. This has resulted in unprecedented demand for brownfield sites as companies seek turnkey facilities to meet increasing demand. Some chip makers adopting a fab-light strategy are selling facilities to other users to secure long-term supply instead of considering alternative uses. This is the case for onsemi whom ATREG has helped sell three operational facilities in the last year alone, including its 150mm fab in Oudenaarde, Belgium sold to BelGaN Group BV, a company that will convert the former CMOS facility to GaN production using the eight-inch conversion kits available for 70% of the tools.
Other companies choose to adapt existing premises, converting them to include fab space or adding a cleanroom as an extension to an existing building. Minnesota-based foundry SkyWater Technology expanded its Bloomington 200mm fab with a Department of Defense grant that provided for most of the build cost. There are multiple factors compelling these and other large, mid-sized, and smaller chip makers to look into refurbishing existing facilities or building new ones. In the U.S., the typical cleanroom building cost used to be $5,500 a ft2 three years ago, but today it is closer to $8,500 a ft2. A cleanroom usually spans one third of the entire fab footprint and needs to be efficient to see a return on investment within five years.
In Albany, New York, a nanocenter was built with some 75,000 ft2 of cleanroom space made available to smaller users for R&D and small-volume production. While some companies are very concerned about control and intellectual property in this nanocenter model, others may find this could be the alternative they need to building smaller field facilities themselves. In today’s context, it will be crucial to continue building on existing semiconductor ecosystems and clusters to provide a more robust supply chain and ensure additional support infrastructure so critical to fab operations.
Both U.S. and European CHIPS legislations are meant to also offer subsidies for new firms and small players, not just large manufacturers. Considering the allocation of funds over a four-year period is the equivalent of two years of China’s spend, it remains to be seen how many smaller projects will get funding. With other schemes such as the Infrastructure Act and the Science Act in the U.S. competing for funding, there is going to be a lot of competition for construction engineering companies and skilled trade workers. These human capital issues represent a significant additional challenge for the semiconductor industry.
The current situation is compelling chip makers to work smarter and join forces to maximize their limited resources to lower manufacturing costs. The recent announcement made by ST Microelectronics and GlobalFoundries is an interesting business model. The partnership to create a jointly operated 300mm fab next to ST’s facility in Crolles, France shows that companies are getting more comfortable about how they deal with – and protect – their intellectual property and that they can create modules within fabs that run in parallel.
We are likely to see more of these partnerships in the not-too-distant future as a way of optimising fabs in terms of loading, efficiency, time-to-market, and cost, especially in the compound semiconductor space where two or three companies could jointly operate a single facility.
Return on investment
As important as the design of wafer fabs is, it is equally important for a company to keep manufacturing costs in check, balanced with a sound strategic R&D investment plan. In addition, product cycles can be very short – a process may take two years to set up, but be made redundant when a competitor takes the lead in the market. This explains why the semiconductor industry has moved cautiously in building fabs and has done so with a focus on clusters.
Wafer fabs need to operate at 75% or above capacity to be cost-effective. Given the high cost of fab operations, chip makers need to be able to manage the process efficiently which has led to so much consolidation over the past decade. Companies can make single products in a streamlined process. Texas Instruments is a good example of this, focusing on analog and 300mm and now dominating the market. The same consolidation strategy has been successful for Intel, Qualcomm, and Nvidia.
The financial markets’ keenness for leading-edge companies to have gross margins of 40 to 60% will need to be addressed because this is not necessarily in sync with the need to carry inventory. As we have learned from COVID, lean inventories can be problematic for the supply chain. Investment in production is one solution to improving margins, so that fabs can process more devices and lower their costs.
The nature of the global semiconductor industry means that plans in place today – and their outcomes – will only be evident in four- or five-years’ time. Both IDMs and foundries are evolving their technologies, manufacturing expectations, and ideas for how to address the increasing demand for more and more microchips. We anticipate they will find new sources of capital, new partnership models, and new ways of adapting their business practices to adjust to the new global semiconductor manufacturing landscape.
Yole Group is curently working on a new technology & market report, Overview of the Semiconductor Industry. This analysis will be available in 2023. For more information, please contact firstname.lastname@example.org.
About the authors
Stephen Rothrock is Founder, President, and CEO of ATREG, Inc. Stephen founded ATREG in 2000 to help global advanced technology companies divest and acquire infrastructure-rich manufacturing assets, including wafer fabs (front- and back-end) as well as MEMS, solar, display, and R&D facilities. Over the last 20 years, his firm has completed 40% of all global operational wafer fab sales in the semiconductor industry, a total of 100 transactions representing a value of over $6 billion. Recent global acquisitions and dispositions have involved Texas Instruments, onsemi, Wolfspeed, Elmos, Allegro MicroSystems, VIS, GlobalFoundries, Infineon, imec, Micron, NXP, Sony, Qualcomm, Renesas, and Fujitsu to name just a few.
Prior to founding ATREG, Stephen established Colliers International’s Global Corporate Services initiative and headed the company’s U.S. division based in Seattle, Wash. Before that, he worked as Director for Savills International commercial real estate brokerage in London, UK, also serving on the UK-listed property company’s international board. He also spent four years near Paris, France working for an international NGO. Stephen holds an MA degree in Political Theology from the University of Hull, UK and a BA degree in Business Commerce from the University of Washington in Seattle, USA.
Jean-Christophe Eloy is Yole Group’s President and CEO. Created in 1998, the market research and strategy consulting company has grown to become a group of companies providing marketing, technology and strategy consulting, media in addition to corporate finance services. His mission is to oversee the strategic direction of Yole Group of companies, including Yole Intelligence, Yole SystemPlus and PISÉO.
All year long, Jean-Christophe builds deep relationships with leading semiconductor companies, discussing and sharing information across his global network. His aim is to get a comprehensive understanding of their strengths and guide their success.
Jean-Christophe is a graduate from EM Lyon Business School (Lyon, France).