A Burgeoning Secondary SME Market Is Inevitable
By Steven Zhou, CEO of Moov Technologies
Many variables impact supply and demand in the semiconductor market, and as a consequence, the semiconductor equipment manufacturing (SME) market. Let's look at an anecdote about lithography equipment, the most expensive category of equipment in a modern fabrication center (fab). In 2012, Nvidia and Qualcomm, clients of Taiwan Semiconductor Manufacturing Company (TSMC), the largest dedicated independent semiconductor foundry in the world , complained about capacity shortages for 28 nanometer (nm) nodes. The 28nm process technology supports a wide range of applications, notably smartphones. TSMC ramped up 28nm production, but within the span of five years reported overcapacity in the industry for the 28nm node.
A number of factors are attributed to the quick cycle that led to overcapacity of the 28nm node: smartphone growth peaked faster than expected, better power-to-performance ratio and enhanced RF capabilities made the 22nm node preferable, and quicker than expected migration from 28nm to 16nm and below rendered 28nm over-supplied.
In reality, nothing has to “go wrong” for market dynamics to substantially alter the semiconductor supply chain. In 2019, the market downturn resulted in the industry's overall equipment utilization rate dropping to 86% from 94% in 2018. (IC Insights).
That said, while TSMC discussed 28 nanometer overcapacity on their Q4 2019 earnings call, another large Taiwan foundry, UMC, stated on their Q4 2019 earnings call that they would be spending ⅓ of their 2020 $1B capex budget to ramp up 28 nanometer to full capacity in their Xiamen factory. Why is UMC investing in additional capital equipment for 28nm process technology if the world's largest chip producer is indicating industry-wide overcapacity?
One reason is that an idle or underutilized fab is an expensive proposition. A single fab operating at full capacity is still better for profit margins (even if the industry as a whole has an oversupply in capacity for that technology) than a single fab operating at 70% capacity. As Jason Wang, co-President of UMC said in the company's Q4 2019 earnings call, “[UMC] will see pretty significant profitability improvements because of the economy of scale benefit for that one particular fab.”
Another reason is that international trade dynamics created an opportunity for UMC to sell older technology to Chinese clients. In 2019, UMC showed a 19 percent sales increase into China (as reported by EE news).
Demand for older technologies can be both a market specific and global phenomenon. Demand for legacy fabs (200mm), which increased globally from 2016 to 2018, but slowed down in 2019 due to the overall industry slowdown and trade disputes, is expected to heat back up fueled by applications related to mobility, sensing, and the Internet of Things (IoT). Notably, demand for 200mm process technology is challenging not only in the ability to acquire legacy equipment but also in that chips made in these fabs have a lower average selling price than cutting edge chips, rendering alternative solutions to full-priced tooling enticing.
While Moore's Law and the rapid technology cycles that render billion-dollar fabs obsolete in five years have led to consolidation at the cutting edge, differentiated end applications (it's not all about smartphones anymore) are creating opportunities for profit that smaller foundries and manufacturers could take advantage of — that is, if they can find a way to more nimbly and flexibly respond to the longtail demand created by changing global trade and technology trends.
Opportunities and Challenges in the Secondary SME Market
Nimbleness is not a word often associated with the semiconductor supply chain. Relatedly, IC manufacturing is not a market with a robust secondary market to help add liquidity to the primary market. Secondary markets run countercyclical to primary markets and exist because the value of an asset changes in the primary market . To illustrate this dynamic, at Moov (a global online marketplace for pre-owned semiconductor manufacturing equipment) we think about the total secondary market value for SME as:
Retail Spend on New Equipment x Fair Market Valueused x Unused Capacity
for some subset of all capital assets where the delta between equipment usable life in a
particular fab and full depreciation is significant enough to merit resale (i.e. a piece of equipment
has a substantial usable life left, just not at its first home fab due to changes in production lines,
utilization of a particular fab, etc.).
In this way, as unused capacity rises so too does the opportunity for manufacturers and foundries to access equipment without investing millions of dollars, and for fabs to recoup at least a portion of their initial capital investments.
By Moov's calculations, the total global value of the secondary semiconductor equipment market is around $50B. However, by our estimations only about $5-$8B per year of that potential market is being moved so far. Since lead times for new equipment can run anywhere from 12-24 months on average, the secondary market would seem to be an appealing alternative for manufacturers and foundries looking to replace or supplement tooling fast. So why is only 10-16% of the secondary market being activated?
Buying equipment on the open market or from brokers poses challenges that have historically limited the growth of a truly robust secondary market:
● The secondary market remains fragmented. Reliability of brokers and third party marketplaces can vary greatly leaving buyers to face a number of marketplace problems such as unreliable or outdated listings, and listings posted by a series of intermediaries removed from the asset's owner.
● There are no guarantees on the condition of used equipment purchased through a broker or the open market, leaving the buyer to take on 100% of the risk.
● Equipment is hard to find, not only because of decentralization, but also because general purpose marketplaces and even standard search engines have trouble reconciling different nomenclature used to identify a machine.
● Assets are typically available “as-is” requiring the buyer to invest significantly in refurbishment, and assume additional risk as quality of third party refurbishers can vary.
● Logistics - packaging, crating, shipping, installing tooling - is typically not accounted for by the seller and/or is a manual process of finding different third-party vendors.
Looking Toward the Future
Moov envisions a future global secondary SME market in which manufacturers and foundries can directly buy and sell tooling from each other, facilitated by robust digital marketplaces.
That said, none of these are insurmountable challenges. For example, OEMs are slowly improving support for legacy systems (hardware and software) and introducing buy-back programs to play a more direct role in the evolving equipment aftermarket. At Moov, we partner with top OEMs to help facilitate buy-back initiatives at a greater scale, while also facilitating transactions between end-users directly. We have introduced the industry's first buyer satisfaction guarantee program for used equipment to help alleviate the risk of buying used equipment from fabs across the globe. Digital-first marketplaces, like Moov, are also leveraging data and automation to streamline the procurement process. For example, at Moov, we're leveraging machine learning on our massive dataset of make/model searches to create a better search algorithm for reconciling equipment nomenclature. Put simply, we're leveraging data and technology to help users of our platform locate the equipment they are looking for, fast, by creating a product search graph unique to our industry. A number of companies have emerged, some partners with Moov, others not, to build value in the aftermarket services sector offering managed refurbishment, logistics, and other aftermarket services. Meanwhile, the Surplus Equipment Consortium Network (SEC/N) established industry guidelines for buying used equipment before their acquisition/merger with SEMI back in 2008. We applaud the early efforts here, and hope to see these standards continue to actively evolve to keep up with the changing landscape of our industry.
The changing landscape of the semiconductor industry - namely the proliferation of end applications for ICs, consolidation of competition at the cutting edge leading smaller foundries and manufacturers to seek opportunities in the longtail, and industry-wide need for economic efficiencies will drive more and more buyers and sellers (and thus innovation) in the secondary SME market. Whether progress in IC production is achieved through “more Moore” or, as many believe, “more than Moore” advancements (or most likely a combination of both in the near-term), one thing remains certain: both innovation in SME technology and end-market applications for ICs are more diversified than ever. This trend of diversification in the supply chain will only continue, with a burgeoning secondary SME market its inevitable outcome.