Although the demand for end products related to the stay-at-home economy slowed down as many countries saw rising vaccination rates and were partially lifting social distancing restrictions, the decline in foundry orders from this source was more than offset by the traditional peak season for smartphones, according to TrendForce’s latest investigations. At the same time, OEMs for notebook (laptop) computers, networking devices, automotive electronics, and IoT devices kept vigorously building up their inventories because the earlier capacity crunch in the foundry market was constraining them from reaching their shipment targets. Because of these developments, demand continued to outstrip supply in the foundry market during 3Q21. As for foundries, they have been gradually taking on new production capacity in the recent period and gaining from the ongoing rise in the ASP. Thanks to robust demand, new production capacity, and rising wafer prices, the quarterly total foundry revenue rose by 11.8% QoQ to reach a new record high of US$27.28 billion for 3Q21. This result indicated nine consecutive quarters of revenue growth.

Top four foundries posted double-digit revenue growth for 3Q21 due to peak season for smartphones; SMIC’s revenue growth was slightly limited by restrictions imposed on its capacity expansions

TSMC raised its quarterly revenue by 11.9% QoQ to US$14.88 billion as it benefited from the release of new iPhone models. The foundry remained firmly at the top of the ranking in 3Q21. Regarding TSMC’s revenue generation by node, the combined revenue share of the 7nm and 5nm nodes has already surpassed 50% and is still expanding thanks to continued demand for smartphone chips and HPC chips. Samsung raised its revenue by 11% QoQ to US$4.81 billion for 3Q21 and sat firmly in second place. The revenue growth was attributed to several factors. First, the releases of new smartphone models during the second half of the year has spurred the demand for SoCs and DDIs. Second, fab Line S2 in Austin has returned to its normal level of revenue contribution following the recovery from the winter storm that struck Texas in the earlier part of this year. Third, fab Line S5 in Pyeongtaek has activated its newly added production capacity. And finally, the revenue result for 2Q21 was a low base for comparison and thus led to a rather impressive performance for 3Q21.

UMC made significant gains in 3Q21 because the activation of new production capacity for its 28/22nm nodes led to an increase in wafer input for OLED driver ICs and other components. This also caused a rise in its blended ASP. UMC’s revenue went up by 12.2% QoQ to US$2.04 billion for 3Q21. With a growth rate that surpassed the top two ranking leaders, UMC retained third place by overtaking GlobalFoundries in the ranking for the first time in 1Q20, and its lead has been gradually widening since then. GlobalFoundries posted a QoQ increase of 12% in revenue to US$1.71 billion for 3Q21 and kept fourth place in the ranking. To address the worldwide chip shortage, GlobalFoundries has announced a series of capacity expansions and greenfield projects this year. Existing plants including Fab1 in Dresden and Fab8 in Malta (which is a town in the state of New York) will take on new production capacity. New plants will also be built in Singapore and Malta. It is worth noting that the capacity expansions and greenfield projects that GlobalFoundries has revealed so far for this year will be financed via a public-private partnership model. GlobalFoundries will be leveraging funding from governments and advance payments from its clients to reduce the pressure of rising capital expenditure and ensure that the new production capacity will operate at a high utilization rate in the future.

SMIC increased its revenue by 5.3% QoQ to US$1.42 billion for 3Q21 and was ranked fifth. Two reasons were behind the revenue growth. First, there is a stable level of demand for its PMICs, Wi-Fi chips, MCUs, and RFICs. Second, SMIC has been steadily raising wafer prices. It is also worth pointing out that SMIC has been adjusting its product mix and client base due to geopolitical factors. Growing consistently over the quarters, the share of Chinese clients in SMIC’s client base came to almost 70% in 3Q21. Under the impetus of the semiconductor policies of the Chinese government, SMIC will continue to give priority to the demand from domestic clients. Hence, the portion of foreign clients in its incoming orders will gradually shrink relative to that of domestic clients.

Second- and third-tier foundries posted higher revenue growth rates compared with first-tier counterparts because of strong demand for mature nodes

HuaHong Group posted a QoQ increase of 21.4% in revenue to US$799 million for 3Q21, thereby taking sixth place in the ranking. HuaHong continues to raise its ASP as it production capacity is expected to be fully loaded through the whole 2021. This development, together with the successful capacity expansion undertaken at its Fab7 in Wuxi, contributed to the above-expected revenue result for the foundry. PSMC’s revenue growth continued to pick up pace in 3Q21 thanks to the general rise in wafer prices and the robust demand for the main categories of chip products (e.g., DDIs, PMICs, CIS, and power discretes such as MOSFETs and IGBTs). PSMC raised its quarterly revenue by 14.4% QoQ to US$525 million and was ranked seventh.

After surpassing Tower Semiconductor in the ranking for the first time in 2Q21, VIS maintained its strong growth momentum by posting a QoQ increase of 17.5% in revenue to US$426 million in 3Q21 on account of several factors. First, VIS increased its products shipments through capacity expansion. Furthermore, VIS was able to optimize its product mix and raise its ASP. It secured eighth place in the ranking. Occupying ninth place in the ranking, Tower Semiconductor’s performance exceeded expectations for 3Q21 with its revenue climbing 6.9% QoQ to US$387 million. Tower’s revenue generation mainly benefited from the stable demand related to RF-SOI chips, industrial sensor chips, and PMICs.

Taking the tenth place in the ranking, DB HiTek registered a 15.6% QoQ increase in revenue to a record high of US$283 million for 3Q21 because of the rising ASP. In the past year, DB HiTek kept its capacity utilization rate at almost 100%. To raise its overall output, the foundry has decided to focus its expansion efforts on its existing wafer production lines. As a result, its production capacity has been increasing slightly since 2Q21. The additional production capacity will effectively contribute to its revenue generation in 4Q21.

Moving into 4Q21, although foundries have undertaken various capacity expansions and greenfield projects, their new production capacity that has been activated this year is already completely booked. The new fabs that foundries have announced will need some time to get built and fully set up, so the chip shortage on the whole will unlikely ease off anytime soon. On the demand side, sales have weakened a bit for TVs and other end products associated with the stay-at-home economy. However, the hardware and infrastructure demand related to 5G, Wi-Fi 6, and IoT continues to gain momentum. Moreover, OEMs for consumer electronics are still stocking up on components in preparation for the year-end holiday sales. Based on the latest examination of incoming foundry orders, TrendForce finds that foundries will continue to operate at fully-loaded capacity. Due to the undersupply situation, the overall ASP of the foundry market has also been climbing. Meanwhile, foundries have been optimizing their product mixes to boost their financial performances. Taking account of this and other aforementioned developments, TrendForce believes that revenue growth will continue for the top 10 foundries in 4Q21. However, 4Q21 will also see more moderate growth compared with the previous quarter because there is a shortage of peripheral ICs made using mature process nodes. Additionally, demand has slacked a bit for some SoC products.

For more information on reports and market data from TrendForce’s Department of Semiconductor Research, please click here, or email Ms. Latte Chung from the Sales Department at lattechung@trendforce.com

Owing to the EV market’s substantial demand for longer driving ranges and shorter charging times, automakers’ race towards high-voltage EV platforms has noticeably intensified, with various major automakers gradually releasing models featuring 800V charging architectures, such as the Porsche Taycan, Audi Q6 e-tron, and Hyundai Ioniq 5. According to TrendForce’s latest investigations, demand from the global automotive market for 6-inch SiC wafers is expected to reach 1.69 million units in 2025 thanks to the rising penetration rate of EVs and the trend towards high-voltage 800V EV architecture.

The revolutionary arrival of the 800V EV charging architecture will bring about a total replacement of Si IGBT modules with SiC power devices, which will become a standard component in mainstream EV VFDs (variable frequency drives). As such, major automotive component suppliers generally favor SiC components. In particular, Tier 1 supplier Delphi has already begun mass producing 800V SiC inverters, while others such as BorgWarner, ZF, and Vitesco are also making rapid progress with their respective solutions.

At the moment, EVs have become a core application of SiC power devices. For instance, SiC usage in OBC (on board chargers) and DC-to-DC converters has been relatively mature, whereas the mass production of SiC-based VFDs has yet to reach a large scale. Power semiconductor suppliers including STM, Infineon, Wolfspeed, and Rohm have started collaborating with Tier 1 suppliers and automakers in order to accelerate SiC deployment in automotive applications.

It should be pointed out that the upstream supply of SiC substrate materials will become the primary bottleneck of SiC power device production, since SiC substrates involve complex manufacturing processes, high technical barriers to entry, and slow epitaxial growth. The vast majority of n-Type SiC substrates used for power semiconductor devices are 6 inches in diameter. Although major IDMs such as Wolfspeed have been making good progress in 8-inch SiC wafer development, more time is required for not only raising yield rate, but also transitioning power semiconductor fabs from 6-inch production lines to 8-inch production lines. Hence, 6-inch SiC substrates will likely remain the mainstream for at least five more years. On the other hand, with the EV market undergoing an explosive growth and SiC power devices seeing increased adoption in automotive applications, SiC costs will in turn directly determine the pace of 800V charging architecture deployment in EVs.

For more information on reports and market data from TrendForce’s Department of Semiconductor Research, please click here, or email Ms. Latte Chung from the Sales Department at lattechung@trendforce.com

As the global vaccination rate rose, and border restrictions in Europe and North America eased, social activities also began to enter a period of recovery, with the consumer electronics market seemingly ready for the arrival of the traditional peak season in 2H21, according to TrendForce’s latest investigations. At the same time, however, the global supply chain was affected by delays in maritime transport, skyrocketing shipping costs, and component shortages, in addition to already-prohibitive price hikes for certain components in 1H21. Given the parallel rise in both material and manufacturing costs, the market for end products has not undergone the expected cyclical upturn in 2H21. Even so, the overall demand for and shipment of smartphones, notebook computers, and monitors experienced QoQ increases in 3Q21, thereby driving up businesses for major OSAT (outsourced semiconductor assembly and test) companies. For 3Q21, the revenues of the top 10 OSAT companies reached US$8.89 billion, a 31.6% YoY increase.

Certain major OSAT companies experienced a slight dip in their capacity utilization rates due to the ongoing shortage of key components including chips and substrates, as well as the electricity rationing that limited both energy intensity and energy consumption in Jiangsu, Zhejiang, and Guangdong at the end of September. Nevertheless, this dip has had virtually no impact on the OSAT industry because certain OSAT companies shifted their operations to substrate-less packaging technologies and reallocated the previously affected capacities. Hence, TrendForce is bullish on the performance of the OSAT industry in 4Q21.

Market leaders ASE and Amkor registered revenues of US$2.15 billion and US$1.68 billion, which represent YoY increases of 41.3% and 24.2%, respectively, for 3Q21. While both companies had some of their capacities hindered due to the shortage of chips, lead frames, and substrates, ASE had its lead times further extended given that its Suzhou-based fab was affected by China’s power rationing. Notably, as the demand for packaging and testing smartphone APs, network chips, and automotive chips remains strong in 4Q21, ASE and Amkor will continue to expand in the 5G, IoT, and AI end-product markets in 2022.

SPIL is currently aiming to strengthen the R&D operations for advanced packaging technologies at its new fab in Erlin, Changhua, since it will be unlikely to compensate for the loss of smartphone AP packaging business from Huawei in the short run. SPIL’s revenue for 3Q21 reached US$1.04 billion, a 15.6% YoY increase. While KYEC previously suffered lowered capacity utilization due to the COVID-19 pandemic, it has since made a gradual recovery and registered a revenue of US$323 million, a 28.5% YoY increase, for 3Q21, thanks to testing orders for 5G chips from Qualcomm and MediaTek. PTI, on the other hand, primarily benefitted from its DRAM packaging and testing business. The company posted a revenue of US$802 million, a 24.0% YoY increase, for 3Q21. Nonetheless, PTI’s memory packaging capacity will likely undergo a sharp decline going forward, as Intel gradually finalizes the sale of its Dalian fab to SK Hynix by 2025, and the agreement between Micron and PTI regarding the assembly and testing services PTI provides at Micron’s Xi’an fab expires in 2Q22. In response, PTI reallocated some capacities at its new fab in Hsinchu to other strategic priorities such as CIS packaging and FOPLP technologies in 3Q21.

Major Chinese OSAT companies JCET and Hua Tian continued to benefit from China’s pursuit of domestic semiconductor substitutes. The two companies expanded their supply of OSAT services for 5G smartphones, base stations, automotive chips, and consumer electronics. As a result, JCET and Hua Tian registered revenues of US$1.25 billion and US$502 million, representing YoY growths of 27.5% and 57.6%, respectively, for 3Q21. Owing to strong sales by its client AMD this year, TFME recorded a revenue of US$636 million, an impressive 59.8% YoY increase, which represents the highest revenue growth among the top 10 OSAT companies in 3Q21.

Although ChipMOS and Chipbond, which specialize in packaging and testing display panel driver ICs, were affected by the slight drop in small-sized TV panel shipment in 3Q21, they were able to compensate for this loss owing to the gradual increase in packaging and testing demand for such driver ICs as TDDI and DDI. This increase can primarily be attributed to the growing demand for mid- and large-sized TV panels, as well as the ramp-up of OLED smartphone panels, which certain smartphone models began to adopt. For 3Q21, ChipMOS and Chipbond grew their respectively revenue by 32.5% YoY and 29.5% YoY to about US$257 million and US$255 million. At the same time, as IC design companies from the upstream supply chain redirected certain orders to ChipMOS and Chipbond in response to China’s power rationing at the end of September, these two companies will likely reach new revenue records in 4Q21.

For more information on reports and market data from TrendForce’s Department of Semiconductor Research, please click here, or email Ms. Latte Chung from the Sales Department at lattechung@trendforce.com

As the COVID-19 pandemic wreaked havoc on the global electronics supply chain, the packaging and testing operations of mid-range and high-end chips were subsequently confronted with prolonged lead times. This can primarily be attributed to the fact that IC substrate suppliers were unable to raise output or expand their production capacities in the short run in order to meet the skyrocketing volume of client orders. Hence, products that are packaged using BGA (Ball Grid Array), Flip Chip, or SiP technologies, all of which require the use of IC substrates, had their lead times lengthened. Certain IC design companies are therefore considering the feasibility of packaging technologies that do not require substrates.

Regarding the trend of advanced packaging development, technologies such as 2.5D/3D IC, SiP, and FOPLP (Fan-out Panel Level Packaging) remain the current mainstream R&D targets. Given the ongoing shortage of semiconductor components, including IC substrates, FOPLP, in particular, has garnered the most attention among the aforementioned three packaging technologies as it can be operated without substrates. At present, most OSAT companies and other chipmakers have successively invested in FOPLP-related technological and manufacturing development in order to capitalize on potential new commercial opportunities.

Despite FOPLP’s advantage of packaging chips across large areas, technological development remains problematic

Regarding the history of FOPLP development as well as the technology’s evolution going forward, its earliest roots can be traced to existing packaging applications including Flip Chip and BGA. As end-products continued to experience performance upgrades, leading to the number of I/O pins being insufficient for meeting the increase in performance demand, new types of wafer-level packaging technologies such as Fan-in and Fan-out subsequently emerged to fulfill the packaging demands of mid-range chips, high-end chips, and other emergent applications.

Although Fan-in and Fan-out packaging technologies are able to effectively raise the number of I/O pins, they also result in a substantial increase in manufacturing costs compared to previous-generation technologies such as Flip Chip and BGA. For both 8-inch wafers and 12-inch wafers, packaging costs have only been on a very slight downtrend. That is why the packaging industry has placed a top priority on simultaneously lowering production costs while raising the number of chips packaged at once. Hence, FOPLP technology has emerged in response to this demand for large-area packaging technology.

Regarding the actual implementation of FOPLP, a potential solution may be found in wafer-level packaging RDL (Redistribution Layer) designs, such as chip first or chip last. It should be noted that chip first FOWLP or chip last FOWLP processes do in fact serve as feasible concepts for FOPLP development. However, the FOPLP process involves stacking massive amounts of packaging materials and chips together, and their combined weight may lead to such issues as panel warpage. In addition, it remains difficult to maintain a consistent uniformity and yield rate during the FOPLP process, meaning further collaborations and optimizations on the parts of OSAT companies and semiconductor equipment suppliers are necessary for FOPLP to succeed going forward.

（Image credit: Unsplash）

While the global electronics supply chain experienced a chip shortage, the corresponding shortage of foundry capacities also led various foundries to raise their quotes, resulting in an over 20% YoY increase in the total annual revenues of the top 10 foundries for both 2020 and 2021, according to TrendForce’s latest investigations. The top 10 foundries’ annual revenue for 2021 is now expected to surpass US$100 billion. As TSMC leads yet another round of price hikes across the industry, annual foundry revenue for 2022 will likely reach US$117.69 billion, a 13.3% YoY increase.

Foundries will gradually kick off production with newly added capacities in 2H22 in response to the ongoing chip shortage

TrendForce indicates that the combined CAPEX of the top 10 foundries surpassed US$50 billion in 2021, a 43% YoY increase. As new fab constructions and equipment move-ins gradually conclude next year, their combined CAPEX for 2022 is expected to undergo a 15% YoY increase and fall within the US$50-60 billion range. In addition, now that TSMC has officially announced the establishment of a new fab in Japan, total foundry CAPEX will likely increase further next year. TrendForce expects the foundry industry’s total 8-inch and 12-inch wafer capacities to increase by 6% YoY and 14% YoY next year, respectively.

Although the manufacturing costs of 8-inch and 12-inch wafer fabrication equipment are roughly equal, the ASP of 8-inch wafers falls short compared with 12-inch wafers, meaning it is generally less cost-effective for foundries to expand their 8-inch wafer capacities. That is why the increase in 8-inch capacity is also expected to fall short of the increase in 12-inch capacity next year. Regarding 12-inch wafer foundry services, the 1Xnm and more mature nodes, which currently represent the most severe shortage among all manufacturing process technologies, will account for more than 50% of the newly added wafer capacities next year. On the other hand, while Chinese foundries, such as Hua Hong Wuxi and Nexchip, account for most of the newly added 12-inch wafer capacities this year, TSMC and UMC will comprise the majority of 12-inch wafer capacity expansions in 2022. These two foundries will primarily focus on expanding the production capacities allocated to the 40nm and 28nm nodes, both of which are currently in extreme shortage. As a result, the ongoing chip shortage will likely be alleviated somewhat in 2022.

Chip shortages will show signs of easing, but component gaps will continue to impact the production of some end products

Application segments such as consumer electronics (such as notebook computers), automotive electronics, and most connected digital appliances are now being impacted by the shortages of peripheral components made with the 28nm and more mature nodes. The undersupply of the said components will probably begin to moderate somewhat in 2H22 if foundries proceed to activate their newly added production capacity. However, just as there will be signs indicating an easing of capacity crunch for the 40nm and 28nm nodes, the tightening of production capacity for 8-inch wafers and 1Xnm nodes is going to be an important development that warrants close attention in 2022.

Regarding 8-inch wafer foundry services, the overall production capacity growth has been limited while the demand related to PMICs has increased multiple folds. The growth of this particular application has to do with the increasing market penetration of 5G smartphones and electric vehicles. Under this circumstance, PMICs continue to take up the available production capacity of 8-inch wafers, and wafer production lines that deploy ≦0.18µm nodes are now expected to operate at fully-loaded capacity to the end of 2022. Hence, the capacity crunch for 8-inch wafers will not ease in the short term.

As for 1Xnm nodes, the number of foundries that are offering these more advanced process technologies is gradually shrinking. The reason is that following the migration to FinFET in the general development of semiconductor manufacturing, the costs associated with R&D and capacity expansions have risen higher and higher. TSMC, Samsung, and GlobalFoundries are now the only three foundries in the world that possess 1Xnm technologies. Also, GlobalFoundries is the only one among these three to undertake a marginal capacity expansion for its 1Xnm node next year. The other two currently have no plan to raise 1Xnm production capacity in 2022.

In the aspect of demand, the kinds of chips that are made with 1Xnm nodes include the following: 4G SoCs, 5G RF transceivers, and Wi-Fi SoCs equipped in smartphones, as well as TV SoCs, chips for Wi-Fi routers, and FPGAs/ASICs. Due to the increasing market penetration of 5G smartphones, 5G RF transceivers will take up a massive portion of the overall 1Xnm production capacity. This will, in turn, significantly limit the available wafer capacity allocated to other products. Furthermore, demand has been rising over the years for smartphones that are equipped with 1Xnm Wi-Fi SoCs and Wi-Fi routers that contain 1Xnm chips. The supply of these components is already very limited at this moment and will get tighter in 2022 because the overall 1Xnm production capacity will not be raised by a significant amount.

In sum, there are several takeaways from this focus on the potential developments in the foundry market next year. First, the major foundries have now announced capacity expansions with the emphasis on addressing the capacity crunch for the 40nm and 28nm nodes. Their newly added production capacity is expected to enter operation next year, following two consecutive years of chip shortages. This will bring some relief to the undersupply situation, which is already very severe at this moment. However, the actual chip output contribution from the newly added production capacity will mainly take place no earlier than 2H22, or during the middle of the traditional peak season. With stock-up activities across the supply chain expected to reach a higher level of intensity at that time because of preparations for holiday sales, the easing of the capacity crunch in the foundry market will not be especially noticeable.

Second, it is worth pointing out that even though supply will loosen slightly for some 40/28nm chips, the lack of production capacity for 0.1Xµm chips on 8-inch wafers and 1Xnm chips on 12-inch wafers will likely remain a serious bottleneck in the supply chain. Currently, production capacity is already quite insufficient for 0.1Xµm 8-inch wafers and 1Xnm 12-inch wafers. Next year, the related capacity growth is also expected to be fairly limited. In sum, TrendForce believes that the foundry market will continue to experience some tightness in production capacity during 2022. Although the undersupply situation will moderate for some components, the persistent issue of component gaps will also continue to adversely affect the production of certain end products.

For more information on reports and market data from TrendForce’s Department of Semiconductor Research, please click here, or email Ms. Latte Chung from the Sales Department at lattechung@trendforce.com