Is the merger between NAND flash memory chipmakers Kioxia and Western Digital (WD) , which was expected to be finalized this month, facing a twist? According to media reports, South Korean memory giant SK Hynix is opposing this merger and is considering a collaboration with Japan’s SoftBank to invest in Kioxia.

As reported by Yomiuri Shimbun on the 18th, insiders reveal that the merger negotiations between Kioxia and WD might be at an impasse. Although both sides aimed to reach a merger agreement this month, SK Hynix, which plans to indirectly invest in Kioxia, doesn’t concur with the merger. In anticipation of negotiations collapsing, SK Hynix considers partnering with SoftBank to invest in Kioxia.

Kioxia, a spin-off from Toshiba’s memory business, was sold in 2018 to a Bain-Capital-lead consortium, including SK Hynix, and Hoya. At that time, regulations stipulated that the largest shareholder, Bain, must secure the consent of contributors like SK Hynix to promote this merger. It’s reported that SK Hynix is apprehensive that WD’s dominance will increase in this merger.

Reports suggest that SoftBank is currently bolstering its AI-related ventures. Therefore, by investing and enhancing relationships, SoftBank may secure a stable memory supply from Kioxia and SK Hynix.

Throne Shift for memory business? Mergers May Reshape Rankings

Nikkei reported on the 17th that SK Hynix does not approve of the Kioxia and WD merger. At this stage, SoftBank is not directly involved in the merger talks between Kioxia and WD. Kioxia and WD aim to secure a financing agreement with financial institutions this week to facilitate the merger. However, the lack of consent from SK Hynix may impact negotiations with financial institutions.

The report points out that in the NAND flash market, SK Hynix is the world’s second-largest manufacturer, trailing only Samsung. If Kioxia, the world’s third-largest manufacturer, and WD, the fourth-largest, were to merge, they would nearly match Samsung’s scale. This would create a significant gap between SK Hynix, which holds the third position, and raise concerns for SK Hynix.

Toshiba, currently holding approximately 40% of Kioxia, will also become a shareholder in the holding company, with Kioxia’s President, Nobuo Hayasaka, assuming the role of President for the holding company. Additionally, Kioxia will have the majority of seats on the board, granting substantial operational authority

According to data from TrendForce, In Q2 2023, Samsung leads the NAND memory market with 31.1% market share, while Kioxia holds 19.6%, and Western Digital has 14.7%. After the Kioxia-WD merger, their combined market share could exceed 34.3%, establishing them as the dominant force in the NAND memory market.

Explore more

• [News] Updates: Kioxia and Western Digital’s NAND Flash Progress towards Merger
• Q2 NAND Flash Revenue Up 7.4%, Anticipated to Exceed 3% Growth in Q3, Says TrendForce

(Image: SK Hynix )

As the global economy enters the post-pandemic era, technologies including 5G, WiFi6/6E, and HPC (high-performance computing) have been advancing rapidly, in turn bringing about a fundamental, structural change in the semiconductor industry as well, according to TrendForce’s latest investigations. While the demand for certain devices such as notebook computers and TVs underwent a sharp uptick due to the onset of the stay-at-home economy, this demand will return to pre-pandemic levels once the pandemic has been brought under control as a result of the global vaccination drive.

Nevertheless, the worldwide shift to next-gen telecommunication standards has brought about a replacement demand for telecom and networking devices, and this demand will continue to propel the semiconductor industry, resulting in high capacity utilization rates across the major foundries. As certain foundries continue to expand their production capacities this year, TrendForce expects total foundry revenue to reach a historical high of US$94.6 billion this year, an 11% growth YoY.

TrendForce’s latest analysis also finds that shipments and production volumes of end products will continue to grow in the post-pandemic period. Regarding host computers, the total (or global) shipments of servers and workstations are forecasted to undergo a yearly growth mainly driven by applications that are enabled by 5G and HPC. As for various types of client (or end-user) devices, the annual total production volume of 5G smartphones, in particular, is forecasted to increase by around 113% YoY. The penetration rate of 5G models in the smartphone market is also forecasted to rise to 37% in the same year. Turning to notebook (or laptop) computers, their total shipments in 2021 will register a YoY growth rate of about 15% thanks to the proliferation of the stay-at-home economy.

Finally, the governments of many countries introduced consumption subsidies during the pandemic so as to stimulate the domestic economy. Video streaming services have also grown dramatically with respect to content and demand because of the pandemic. As a result, the TV market is seeing a wave of replacement demand as consumers want to purchase the latest models that offer higher resolutions (e.g., 4K and 8K) and network connectivity (i.e., smart TVs). The total shipments of digital TVs in 2021 are forecasted to undergo a YoY growth rate of around 3%.

The high demand for the aforementioned end devices has therefore resulted in a corresponding surging demand for various ICs used in these devices, including CIS, DDI, and PMICs. In addition, the increasing adoption of cloud services, including IaaS, PaaS, and SaaS, has also generated a massive demand for various high-end CPUs and memory products used in the HPC platforms that power said cloud services.

On the whole, TrendForce believes that, with demand maintaining a healthy growth momentum for many kinds of end products, semiconductor components that are manufactured with the same foundry nodes will be competing for production capacity. Some categories of ICs will therefore experience a more severe capacity crunch due to the product mix strategies of respective foundries. In the short term, no effective resolution is expected for the undersupply situation in the foundry market.

Certain foundries will continue to expand their production capacities in 2021 as the semiconductor industry undergoes a structural change

With regards to the expansion plans of various foundries this year, tier-one and tier-two foundries will prioritize the development of different process nodes. More specifically, tier-one foundries, including TSMC and Samsung, will focus on the R&D, fab build-out, and capacity expansion for the 5nm and below nodes in response to the growing chip demand for HPC-related applications. On the other hand, tier-two foundries, including SMIC, UMC, and GlobalFoundries will primarily focus on expanding their production capacities of the 14nm to 40nm mature process nodes in order to meet the massive demand for next-gen telecom technologies (such as 5G and WiFi6/6E) and other diverse applications (such as OLED DDI and CIS/ISP).

Incidentally, it should be pointed out that SMIC’s capacity expansion plans have been constrained after the US Department of Commerce added SMIC to the Entity List, which prohibited the company from procuring US semiconductor equipment. However, SMIC still possesses enough funds for procuring non-US equipment and building new fabs, as the company is not only actively expanding its existing 8-inch and 12-inch wafer capacities, but also proceeding with the construction of its new fab in Beijing.

Apart from the aforementioned companies, other foundries, including PSMC, Tower Semiconductor, Vanguard, and HHGrace, will prioritize the capacity expansion of their 8-inch wafers (which are used for the 55nm and above nodes) to meet the demand for large-sized DDI, TDDI, and PMICs. These foundries, in contrast with their larger competitors, are primarily focusing on 8-inch capacity expansion due to the relatively high cost of DUV immersion systems used for the 40/45nm and below processes. For these companies, it is much more economically feasible to instead undertake capacity expansions for the 55/65nm and above nodes.

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 Samsung Display (SDC) decided to extend the manufacturing operations of its Korea-based Gen 8.5 LCD fab, and tier-two panel suppliers are still slow to reassign their production capacities from TV panels to IT panels, TrendForce expects total TV panel shipment for 2021 to reach 269 million units, which is relatively unchanged compared to 2020 levels. Panel suppliers will continue to focus on large-sized TV panels this year in response to several industry-wide developments, including M&A, reduced production capacities, improved manufacturing technologies, and increased panel demand. Furthermore, as the persistent price hike of TV panels continues to reduce the profit margins of TV sets, TV brands have started to gravitate towards larger, more profitable TV sizes. TrendForce therefore expects the average TV panel size this year to increase by 1.6 inches and move towards 50 inches.

TrendForce analyst Jeanette Chan indicates that the shift towards large-sized panels is an effective means of expending the production capacity of panel suppliers. Case in point, due to the limited production capacity for TV panels in 1H21, not only are TV panels currently in short supply, but TV panel prices are also on the rise. On the other hand, the demand for TV panels in 2H21 will depend on several key factors: first, whether the increased retail price of TV sets will hamper consumer demand; second, whether the pandemic will be effectively brought under control as more countries begin vaccinations; third, whether the impending global economic recovery will be a significant one. And finally, whether a market bubble will appear as a result of TV manufacturers’ overbooking panel orders in anticipation of potential hindrances including the price hike of materials in the upstream supply chain, the shortage of glass substrates due to such accidents as facility fires, the shortage of IC supply, and the extended shipping times.

Thanks to their persistently rising production capacity and successful acquisitions, China-based BOE and CSOT, the two largest panel suppliers in the world, are expected to collectively account for about 40% of total TV panel shipment this year. At the same time, BOE and CSOT are actively improving their technologies and making a push for high-end products, such as 8K, ZBD, and AM Mini LED. By leveraging their improved technologies and available funds, the two companies are likely to extend their operations upstream by systematically undertaking vertical integrations.

On the other hand, HKC, which is currently raising its production capacity, has garnered much attention in the market amidst the current shortage situation of TV panels. Along with its Changsha-based H5 fab, which is set to kick off mass production shortly, HKC possesses four Gen 8.6 fabs in total. By raising its production capacity and engaging in additional strategic partnerships with tier-one TV brands, HKC is expected to enter the top three ranking of panel suppliers by TV panel shipment for the first time ever, with a shipment of about 41.91 million units this year, a 33.7% increase YoY.

Taiwan-based AUO and Innolux are expected to experience YoY decreases in their shipments this year as their production capacities are relatively limited, although both companies’ efforts to optimize their products and engage in cross-industry partnerships have brought them certain competitive advantages. In particular, AUO is leading the panel industry in developing not only ultra-high-end products, such as 8K+ZBD, but also Micro LED displays, whereas Innolux holds competitive advantages in product diversity and in-house ODM services. It should be pointed out that these two Taiwanese companies are able to deal with the current IC shortage situation better than their competitors because their parent companies have longstanding business relationships with IC design companies.

With regards to Korean suppliers, although LGD and SDC have both prolonged their LCD manufacturing operations in Korea in order to satisfy the current bullish market demand, the two companies are primarily focusing on transitioning their offerings to new products. LGD will expand the OLED production capacity of its Guangzhou fab in 2Q21 as part of its effort to dominate the OLED market. As for SDC, the company has dropped out of the top six ranking this year as a result of its lowered production capacity. However, new TV sets featuring SDC’s QD-OLED panels are expected to officially hit the market in 4Q21, in turn driving SDC’s yearly TV panel shipment to 2 million units in 2022.

For more information on reports and market data from TrendForce’s Department of Display Research, please click here, or email Ms. Vivie Liu from the Sales Department at vivieliu@trendforce.com

TrendForce’s investigations finds that LG manufactured merely 30.6 million smartphones last year, which represented a 2.4% market share. The Korean company took ninth place in the global ranking of smartphone brands by production volume in 2020. At the start of this year, LG began to consider either selling or shuttering its mobile phone unit.

Around that same time, it also suspended the R&D of new smartphone models. On April 5, LG announced the decision to fold up the mobile phone unit as it was ultimately unable to offset consecutive years of financial losses it suffered in the smartphone market. Based on the company’s current plan, the mobile phone unit will wind down its operation by the end of July, while its smartphone manufacturing operations will cease by the end of 2Q21.

TrendForce therefore expects LG to occupy a lower than 1% share in the smartphone market this year. Incidentally, the conditions for survival for smartphone brands have further deteriorated on account of the increasingly fierce market competition as well as the recent and continuing hike in component prices. Taken altogether, these developments will reinforce the trend of the dominant brands having more and more market share in the future at the expense of the smaller brands.

Regarding LG’s performance in the smartphone market during the recent years, the company spared no effort in high-end R&D, with such results as the LG Wing with a rotating screen and the LG Rollable, which, as the name suggests, features a side-rolling display. The latter model remained a concept and did not enter mass production.

Despite its efforts, LG however continued to lag behind in sales when compared with the other major brands, such as Samsung and Apple. LG had a relatively weak position in the high-end segment of the smartphone market. As for the mid-range and entry-level segments, LG could not match Chinese brands in terms of pricing. To optimize its cost structure, LG expanded the share of device production going to ODMs. Nevertheless, this action was too late to turn things around.

Samsung, Lenovo, and Xiaomi are likely to benefit from LG’s exit from the North and Latin American markets

LG’s smartphone business has become unprofitable since 2Q15; and its financial losses were further exacerbated after it made a gradual exit from the Chinese market in 2016. As of 4Q20, LG’s smartphone business suffered 23 consecutive quarters of financial losses, which totaled about 5 trillion KRW. Despite LG’s limited market share, however, its exit from the various regional smartphone markets will still benefit its competitors in those markets, in particular, the mid-range segment in North America and Latin America.

With regards to North America, LG’s market share there will be split among its Android-based competitors, including Samsung, Lenovo (Motorola), and other in-house, private brands owned by domestic telecom operators. With regards to Latin America, on the other hand, LG’s exit will more noticeably benefit Lenovo (Motorola) and Xiaomi.

Looking ahead to the rest of 2021, as vaccinations begin to take place around the world, TrendForce expects the smartphone industry, which fulfills a basic living necessity, to make a gradual recovery as well. Thanks to the general public’s cyclical replacement demand, as well as growing demand from emerging markets, total smartphone production remains unaffected by LG’s exit. As a result, TrendForce maintains its smartphone production forecast of 1.36 billion units for 2021, a 9% increase YoY.

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 three dominant DRAM suppliers (Samsung, SK Hynix, and Micron) are currently experiencing a shortage in their production capacities, the corresponding shortage situation in the DRAM market has yet to be resolved, according to TrendForce’s latest investigations. Taking advantage of the fact that the whole DRAM market has entered a period of cyclical upturn in 1Q21, DRAM suppliers have significantly raised quotes for specialty DRAM products. This extraordinary development has led to price hikes that are almost double digits for some specialty DRAM chips. Furthermore, the magnitude of the price hike especially widens for products belonging to the lower part of the density range and the more niche applications. Looking at MoM changes in contract prices of specialty DRAM products for February, DDR2 and DDR3 chips saw the largest price hikes. Prices of DDR4 chips also went up due to the influence of the rising quotes for DDR3 chips. The average contract price of DDR3 4Gb chips, which are still mainstream for specialty applications, jumped 6.8% MoM.

As for DDR3 2Gb chips that are primarily promoted by Taiwan-based suppliers, TrendForce indicates that there is not enough supply even as quotes are being offered. With this situation becoming the norm, the price range (i.e., the difference between the high and low prices) has also expanded dramatically. The average contract price of DDR3 2Gb chips rose by nearly 9% MoM in February. Samsung significantly raised quotes for DDR4 4Gb chips in response to the sharp upswing in prices of DDR3 chips. The low and average prices of DDR4 4Gb chips for specialty applications both climbed around 6% MoM. The average price of DDR4 8Gb chips went up by around 4% MoM as the price trend of this product aligns with the general price trends of mainstream PC DRAM and server DRAM products. However, it should be pointed out that the hikes in contract prices of specialty DRAM chips were mainly the result of the adjustments made to monthly contract deals and thus reflected the market situation of February. By contrast, prices held steady for quarterly lock-in deals with tier-1 clients.

While the three dominant DRAM suppliers have been slowing down their DDR3 manufacturing, Taiwanese suppliers are constantly adjusting their capacity allocation to maximize profits

With demand getting hotter in the specialty DRAM market, DDR3 products are starting to surpass DDR4 products and logic ICs in profit margin. Consequently, DRAM suppliers are also changing their strategies. Looking at South Korean suppliers, Samsung will continue shifting the wafer production capacity of Line 13 to CMOS image sensors over the long term. However, this reassignment has now been scaled down for this year due to the recent surge in specialty DRAM prices. Likewise, SK Hynix will keep the DRAM production capacity of its older fab M10 relatively constant through 2021 after reducing it in 2020. As for Micron, it has raised the yield rates of the 1Z-nm and 1-alpha processes, so the output shares of products based on these more advanced technologies will gradually expand. Since the available fab space in Taiwan is limited, Micron will relocate the 20nm and more mature processes to Fab 6 in the US. In sum, the output of DDR3 products from the three dominant suppliers will continue to shrink, but the pace of the reduction is now slower than originally expected.

Regarding Taiwan-based suppliers, Nanya has shifted some 20nm and 30nm production capacity from DDR4 products back to DDR3 products. Winbond has been focusing on Flash products in the recent years, and its DRAM production capacity will remain fairly limited until the completion of its new fab in Kaohsiung. Nevertheless, Winbond is concentrating its DRAM production efforts on low-density DDR2 and DDR3 products (i.e., 1Gb and 2Gb chips). It actually has the advantage of being able to raise prices as its market share for low-density products is fairly large. As for PSMC, it has been focusing on foundry manufacturing of logic ICs for a while. However, with prices now rising for foundry manufacturing of DDR3 products, PSMC now wants to shift some wafer production capacity back to DRAM. Going forward, the three Taiwan-based suppliers will keep adjusting their capacity allocation strategies in accordance with changes in the profit margins of different products. Nevertheless, even as suppliers are now changing their product mixes to take advantage of the latest market situation, TrendForce projects that specialty DRAM products will be in undersupply at least through 1H21. The magnitudes of price hikes for various types of specialty DRAM products will depend on suppliers’ capacity allocation strategies.

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