JDI HIGASHIURA FAB SALE TO SONY NOW COMPLETE
ATREG is pleased to announce that the acquisition of Japan Display, Inc. (JDI)’s large operational LCD panel manufacturing fab located in Higashiura, Japan by Sony Semiconductor Manufacturing Corporation (SCK) has officially closed. In May 2022, JDI announced that the company would end production at the Higashiura fab by March 2023 to increase JDI’s cost competitiveness and profitability.
“While concerned about the size of the site and the short timeline to find and secure a buyer, ATREG brought multiple interested parties to the table and through a competitive process allowed JDI to select the best buyer for this well-designed, infrastructure-rich facility,” explains Stephen Rothrock, President & CEO of ATREG. “Congratulations to both parties on a successful outcome.”
Click here to read our full press release. Click here to read JDI’s press release.
EXECUTIVE Q&A WITH STACY RASGON, BERSTEIN RESEARCH
ATREG recently sat down with Stacy Rasgon, Managing Director and Senior Analyst, U.S. Semiconductors and Semiconductor Capital Equipment at Bernstein Research, a global leader in equity research and trading, to discuss the latest investment trends in the global semiconductor market.
You have been following the global semiconductor market for years. If you were to summarize the current state of the market in a few words, what would you say?
I think the sector and market have never been more dynamic than they have been over the last few years, whether cyclically, technically, or geopolitically. On the cyclical side, COVID shortfalls and supply chain whipsaws have driven a return to higher levels of volatility as companies wrestle with demand booms and busts, shortages and gluts, inflationary pressures (and frankly benefits as the industry has had no problems passing along the resulting higher input costs), and significantly higher capital intensity, and yet the industry continues (through the cycle) to grow. On the technical side, new innovations are coming to the forefront as Moore’s Law ebbs (with the advent of at-scale extreme ultraviolet lithography (EUV), new transistor structures, chiplets and 3D packaging, and backside power delivery) and the rise of AI drives ever-increasing need for compute. And it is becoming abundantly clear how important this industry is from a geopolitical standpoint as the U.S. and China square off.
In your opinion, what key factors are driving the global semiconductor market today?
In the near term, the industry is working through an asynchronous demand cycle as COVID-era euphorias wear off. Some end-markets (PCs, smartphones, etc.) have bottomed and are showing some signs of life. And AI-related spending is of course off the charts. At the same time though, traditional data center spending (servers, networking, etc.) remains weak and industrial and automotive markets are rolling over. And some industries (advanced AI, semicap equipment, and Huawei suppliers) are suffering through the impact of some rather extreme export controls as the U.S. attempts to stymie China’s ability to build out a local leading-edge ecosystem. Nevertheless, the industry appears poised for growth this year and over the long term almost certainly (many market forecasters are calling for a trillion dollar market by 2030, a level that we believe is absolutely plausible at some point, if not 2030 than at least not too far after).
We can’t talk about the semiconductor industry without mentioning artificial intelligence (AI). What do you think is going to be the biggest impact of AI on the industry?
The numbers at this point are absolutely staggering, at least for the (relatively small) subset of companies that are supplying in size and the prospect for AI and accelerated computing has the potential to form the basis for the next leg of growth in the industry. I see significant long-term potential for the space as AI deployment (both in the cloud and the edge) becomes more ubiquitous and the rise of accelerated computing in the data center drives much needed computing efficiencies as installations grow. That being said, the potential return required to justify the current magnitude of investment is starting to draw more queries, but we remain early on the journey and I remain positive on the opportunity.
What is your view on the issue of excess inventory, how companies are dealing with it, and what effects it will have on the overall supply of wafers moving forward in our current market?
Overall inventory levels, both on company balance sheet and (importantly) in the channel remain extremely high, much higher than they were pre-COVID. While some of this might be structural given the shortages the supply chain experienced during the pandemic, I do worry somewhat (particularly in broad industrial markets, automotive, and PC CPUs). Currently many analog (diversified) companies are in the process of trying to drain their channels somewhat and are seeing headwinds as a result though there is reasonable hope that Q2 might represent the bottom.
If you had a few pieces of advice for semiconductor investors today, what would they be?
- Watch end-markets, everything ultimately stems from them.
- Watch the channel, semis are at the back of the supply chain and inventory can build or bleed in many places.
- Semis are the fundament of all technology – there is a read-across from everything to everything and back again.
- Semiconductor management teams’ visibility into end-demand in the short term is precisely zero (all they see are the orders in front of their face). Note that this is not “bad,” it just is. Don’t forget it.
- Watch out when stocks stop going up on good news or stop going down on bad news.
- But over the long term, enjoy the ride. This is an absolutely remarkable space with companies doing amazing things (I’m frankly amazed that any of it works at all!)
Some analysts predict that the tech sector could return to modest growth in 2024. Do you agree?
Semiconductors themselves should show some growth this year, though it’s still a very mixed environment. But PCs/smartphones (historically about half of semiconductor demand) have bottomed, AI is growing like a weed, and memory should have a strong year (following a horrendous 2023), all of which should likely deliver revenue growth for the industry overall. Additionally, units themselves appear to have been shipping reasonably well below trend (while non-memory revenues declined only ~2% in 2023 units fell almost 20%, the worst year for unit growth since 2001, and suggesting potential for catch-up). Pricing in fact saved the day as companies passed along inflationary input costs (and at their margin structure), we will be watching the trajectory of pricing going forward, but as of right now, things do not appear hugely out of the ordinary on that front.
In your opinion, what is the inflection point in the semiconductor cycle?
From a revenue standpoint, we’re likely past the inflections. Stock behavior tends to follow a different trend. It turns out that normally the best time to buy stocks is after numbers come down, but before they trough (just before the last estimate cut if you could time it perfectly). Hence stocks tend to be anticipatory and have in many cases already started to move on anticipation of an approaching bottom (we shall see if that is justified or not).
What is your take on the evolution of the memory market over the coming months?
Memory had a horrible, horrible year in 2023 (the worst memory down cycle since the financial crisis at least, and maybe even since the tech bubble) though the companies are in much better shape this time (we had bankruptcies before!). However, pricing has started to rip higher, partially on demand (especially in HBM DRAM for AI applications) and partially on supply as the companies took capacity offline in the wake of the downturn. This year should likely be much better as pricing has already started to move rapidly higher and HBM demand is off the charts.
Geopolitical tensions between the U.S. and China do not show any signs of abatement. Do you anticipate a further escalation of this trade war in the near future, and if so, what is the impact going to be on semiconductor companies?
We have seen the implementation of tariffs, export controls, and specific sanctions on Huawei entities and I doubt tensions are going to get better anytime soon. Export controls have primarily hit AI suppliers and semicap companies (the former have been almost completely cut off from selling to China and the latter have been relegated to selling only trailing-node tools and servicing into the country). I am not currently in the camp that we will see, for instance, extension of controls to impact lagging-edge manufacturing, but I do think we may continue to see tightening of the controls at the boundary and would not be surprised to see other individual companies targeted in a manner similar to Huawei (who was placed on the entity list and subjected to the Foreign Direct Product Rule which effectively eliminated their ability to buy or use anything that leverages any U.S.-based technologies). There has also been recent news flow about potentially restricting access to gate-all-around technology, which I would interpret as more likely preventing local Chinese companies from sourcing leading-edge manufacturing capability from foundries outside China (rather than, say, banning sales of GAA chips for acceptable applications into China).
There is not much China can do in response directly in semiconductors (they mostly need what they can buy and don’t have enough leadership to significantly impact things on their own) though retaliation through other means where their positioning is stronger (batteries? Rare earths? Slow-walking M&A? etc.) I prefer not to think about the broader Taiwan risk though (which has potential to be a black swan not just for semis, but frankly for everything). I also do not know what the upcoming U.S. election results might signify regarding possible escalation risk.
How do you anticipate U.S. companies should respond to increasing technological competition from their Asian competitors?
Chinese companies (semicap, analog, etc.) will likely gain more share than they “deserve” to gain inside China (for locally consumed applications at least) as there is clearly a mandate to go local wherever possible. Local Chinese AI chips are already available (not as good as their U.S. counterparts, but good enough if the alternative is nothing). China is doubling down on trailing node capacity which could continue for quite some time, assuming self-sufficiency is the goal. And we are seeing efforts to work around leading-edge restrictions by leveraging local capabilities (multiple patterning vs. EUV, remaining on FinFET vs. GAA, etc.), something that will undoubtedly be deficient vs. what is available outside the country, but presumably suitable if one has no choice (which they don’t). Hence while I understand the national security rationale for the U.S. efforts to hamstring China, I remain less than convinced that it is the right move (though frankly I don’t know what the right move is!) as we are forcing China to become creative (in all senses of that word).
CHIPS Act public funding allocations have started. Do you believe these investments will be enough to bring semiconductor manufacturing back to U.S. shores?
Yes and no. If the goal is really to significantly increase the fraction of capacity installed in the U.S. vs. elsewhere, I don’t think the numbers are nearly big enough, though they are absolutely a good start. They have brought some projects that might not have gotten built otherwise and have jump-started more investment from the companies themselves (the CHIPS Act funds only a portion of the project base). In practice, I believe the investment tax credit was better thought out (little or no red tape, a much easier process to receive the cash and actually the bigger portion at ~25% of qualified costs). The grants process however appears far slower than it could have been (and at 5-15% of project costs is actually not that big). But I am happy to see the investments get started and would love to see a second (possibly streamlined?) round, assuming we have the political will remaining to do it.
About Stacy Rasgon
Stacy serves as Senior Analyst at Bernstein covering U.S. semiconductors. Prior to joining Bernstein in April 2008, he worked as a management consultant at McKinsey & Company where he advised clients across the semiconductor value chain on three continents in matters of strategy, operations, and M&A. Stacy also spent time at IBM’s TJ Watson Research Center where he examined line edge roughness formation mechanisms during plasma etching of semiconductor devices. He holds a PhD in Chemical Engineering from the Massachusetts Institute of Technology (MIT), as well as a Certificate in Financial Technology from the MIT Sloan School of Management. Stacy earned a BS summa cum laude in Chemical Engineering from the University of California, Los Angeles. Institutional Investor‘s annual survey has recognized Stacy’s research since 2010. In 2015, 2016, 2017, and 2018, II ranked Stacy as the No. 1 analyst in the semiconductor space.
THE INCREASING ROLE OF GOVERNANCE IN SEMICONDUCTOR FOREIGN DIRECT INVESTMENT
By Annie Rothrock, Senior Vice President, ATREG, Inc.
Escalating geopolitical tensions are increasingly manifesting in enormous government subsidy packages to semiconductor firms, most recently seen in China’s announcement of a $47.5 billion dollar fund to boost its advanced chip capabilities. Such subsidies, undertaken by the likes of Malaysia, South Korea, Japan, and of course the United States, reflect the renewed importance governments are placing on national chip manufacturing and semiconductor value chains. Amidst this climate, companies must navigate new political realities as they look to buy or sell fab manufacturing assets.
China and the U.S. have approached subsidies differently. While China’s subsidies are almost entirely allocated and directed inwardly to local firms trying to catch up to foreign competitors, the U.S. has taken a broader approach, rewarding foreign semiconductor companies such as SK Hynix, Samsung, and TSMC who have shown a commitment to integrate into the U.S. chip supply chain. For these companies, increasing foreign direct investment (FDI) into the U.S. is not chiefly a matter of political expediency, but of strategic necessity – both SK Hynix and Samsung, for example, worry that Micron has a strategic edge in the race for High Bandwidth Memory (HBM) due to its CHIPS Act funding, leading each of them to increase their FDI into the U.S. in a race for market share.
The numbers speak for themselves. According to a report published in May by research and analysis company GlobalData, the U.S. semiconductor industry has garnered more foreign investment – over $100 billion – than any other country, with top investors TSMC, Infineon, Bosch, and ASML putting the U.S. far ahead of the next closest recipients, Germany and Japan, as shown in the diagram below. Even so, China is not afraid to flex its muscles where it can. In August 2023, Intel aborted its proposed $5.4 billion acquisition of Tower Semiconductor, citing more than 18 months of delay by Chinese regulatory authorities, a situation many observers attributed to intentional disruption rather than simple bureaucratic sluggishness.
In a twist that embodies the delicate nuance of today’s FDI landscape, however, Intel’s commitment to China has remained steadfast. Despite U.S. President Biden’s executive order last year to create a mechanism for restricting outbound investment in the semiconductor, quantum information, and AI sectors in foreign “countries of concern,” a designation that includes China, Intel is still focusing on China as a huge market for its chips. Intel CEO Pat Gelsinger has recently pushed back against U.S. export controls, arguing that overly strict controls simply incentivize the Chinese to create their own national chip champions, hurting Intel’s market share in the process.
Though Intel is foremost among a cadre of Western companies that still want access to the Chinese market, there is no doubt that stringent export controls and FDI review mechanisms are having a real impact on where companies are able to locate their actual manufacturing. In the aftermath of the failed Tower takeover, Intel forged ahead with a new foundry partnership with UMC who is using some of their fab capacity in Arizona. UMC is focusing on relatively more mature 12nm production which frees Intel up to continue focusing on its leading-edge 3nm and 2nm and beyond production – an interesting pivot by Intel after the Tower deal fell apart which offers some lessons in how companies can pivot in the face of FDI challenges.
No doubt the lines are drawn more clearly today, in part due to lessons learned via chaotic FDI reviews of the past few years. In early 2022, the German government failed to approve Taiwanese firm GlobalWafers takeover of Siltronic before a regulatory deadline, a result that appears to have been driven more by an overabundance of caution than an intensive concern over Taiwanese involvement in the German semiconductor industry. In April 2021, the Italian Government vetoed the acquisition of a controlling stake in Italian semiconductor equipment firm LPE by Chinese company Shenzhen Investment Holdings Co., an outcome that would come to mark the beginning of more intense European scrutiny on Chinese chip deals in the region that continues today.
With China’s aforementioned funding push for semiconductor self-sufficiency and U.S. President Biden weighing additional limits to Chinese access to AI chips, companies now have a clearer picture of the FDI landscape. This does not mean they will have any easier a time deciding on where to locate their fab assets. Grey areas remain and it is likely that seemingly uncontroversial deals will run into unexpected regulatory review and obstruction. Perhaps in this environment companies will look to countries such as Malaysia who have touted their neutrality in a bid to attract both East and West. In any case, semiconductor manufacturers now have an added challenge in how they think about where to locate these increasingly expensive and vital assets.
About Annie Rothrock
Annie joined the ATREG team in January 2016. As Senior Vice President, she is responsible for advising the world’s most reputable chip makers on their global strategic manufacturing asset sales and acquisitions. Over the past seven years, Annie has successfully led both sell- and buy-side semiconductor fab and cleanroom transactions for such clients as Elmos, GlobalFoundries, Nexperia, Micron, onsemi, Renesas, TDK, Texas Instruments, and VIS amongst many others. In 2023, she assisted Nexperia with the sale of its operational 200mm Newport UK fab to Vishay Intertechnology.
In 2022, she facilitated the disposition of onsemi’s operational GaN-capable 150mm fab in Oudenaarde, Belgium to Belgian photonics company BelGaN Group B.V. In 2019, Annie was instrumental in Wolfspeed’s decision to select the Marcy Nanocenter in NY’s Mohawk Valley to build the world’s first, largest, and only 200mm greenfield silicon carbide (SiC) manufacturing facility. In 2017, she led the sale of Texas Instruments’ operational 200mm / 150mm fab in Greenock, Scotland, UK to Diodes.
Poised for recovery in 2024 and driving toward a historic $1 trillion in revenue, the global semiconductor industry has an incredibly promising future, backed by an unprecedented number of growth drivers, market opportunities, and technology advancements. Nevertheless, amid record greenfield capital investments and government-backed regional capacity expansion, global semiconductor manufacturing still needs to overcome perennial headwinds over the coming years.
Tracking the exchange of wafer fabs worldwide is an effective way to forecast where the global semiconductor industry is heading. In this article published in SemiWiki on May 26th, ATREG President & CEO Stephen Rothrock discusses the top three challenges semiconductor manufacturing will face in 2024, including geopolitical uncertainty, technological shifts, and capacity sourcing.
Click here to read the full article.
ATREG looks forward to seeing you at this year’s SEMICON West from July 9 to 11 in San Francisco, CA. Please email us to set up an appointment with one of our fab transaction advisors to discuss your global infrastructure-rich semiconductor manufacturing asset needs, whether brownfield, greenfield, or foundry. You can also connect with our team at the Fab Owners’ Alliance Speed Networking Event in Room 151 of the Moscone Center South the afternoon of Thursday, July 11th.
In addition, ATREG will attend the following global industry conferences until the end of the summer.
- FOA Q3 Meeting, July 9-11, San Francisco, USA
- Goldman Sachs Communacopia & Technology Conference, September 9-12, San Francisco, USA
- SEMICON India, September 11-13, Delhi, India
- GSA U.S. Executive Forum, September 26, San Francisco, USA
We hope to see you there!