2024-08-30 - tech

Embarrassing Kioxia Semiconductor

In February, a key figure in Japan's storage industry, Sakae Sakashita, known as the "Godfather of Japanese Semiconductors," passed away due to health issues. This veteran had served as the President, Representative Director, and CEO of the Japanese DRAM manufacturer Elpida. Elpida was once an important semiconductor company in Japan, originating from the merger of DRAM businesses of NEC and Hitachi in 1999, forming NEC Hitachi Memory. In 2000, NEC Hitachi Memory was renamed Elpida Memory, and at its peak, Elpida's global market share in memory storage was close to 20%. However, after experiencing severe market fluctuations, the Lehman crisis, and the impact of the yen's depreciation, Elpida faced financial difficulties, eventually declaring bankruptcy and being acquired by Micron.

Elpida's experience is a microcosm of the decline of Japan's semiconductor industry. Another Japanese storage semiconductor company has also fallen into a developmental dilemma, which is Kioxia Semiconductor.

01

The Era of Kioxia Semiconductor's Dominance

The predecessor of Kioxia Semiconductor was also a major Japanese semiconductor company—Toshiba (which has since announced its delisting).

In 1984, Toshiba developed a new type of semiconductor memory (NOR) and named it flash memory, thus ushering in a new era. As one of the five main companies selected by the Japanese government at the time, Toshiba established a W plan in 1982, investing 34 billion yen and 1,500 technicians into the research and development and production of DRAM. In 1985, Satoru Kito of the semiconductor team developed the largest capacity 1M DRAM in the world at that time, and Toshiba overtook the United States to become the world's leading semiconductor manufacturer. Seven years after the implementation of the W plan, Toshiba ranked second among the top ten semiconductor manufacturers globally, just behind NEC (Nippon Electric Corporation). At that time, Toshiba was at its peak, with its stock price reaching a historical high of around 1,500 yen.

Advertisement

In 1987, Toshiba proposed the concept of NAND flash memory, which is more suitable for data storage and has a lower cost. However, for various reasons, NAND flash memory was not given much attention until the personal computer began to popularize in the 1990s, and Toshiba put the research and mass production of NAND back on the agenda. In 1989, Toshiba's first NAND flash memory was launched and achieved success. But the good times didn't last long, as Samsung began to focus on semiconductors at that time, with greater investment and more decisive actions. In 1993, Samsung took the lead in global memory, relegating Toshiba to second place.

In 2015, Toshiba was embroiled in a financial fraud scandal. Toshiba, which had purchased Westinghouse Electric at all costs, faced a financial deficit of 8.8 billion US dollars. To repay its debts, Toshiba chose to sell its most profitable semiconductor business.

In 2017, Toshiba Memory was established, with the original Toshiba flash memory research and development and manufacturing departments becoming independent from Toshiba, bringing in 18 billion US dollars in revenue for Toshiba's semiconductor business. The consortium that took over, Bain Capital, included companies such as SK Hynix, Apple, Dell, Seagate, and Kingston, and at that time, Toshiba also invested in shares of Toshiba Memory.

In July 2019, Toshiba Memory announced that it would change its name and officially became "Kioxia" on October 1st, with the Chinese name "Kioxia Corporation," marking the birth of Kioxia.Why Did Kioxia Decline?

With its leading three-dimensional (3D) vertical flash memory cell structure BiCS FLASH, Kioxia's flash memory density ranks among the top in the market. Kioxia was the first industry participant to envision and prepare for the successful migration of SLC technology to MLC, then from MLC to TLC, and now from TLC to QLC. In the fiscal year of 2021, Kioxia achieved sales revenue of 1,526.5 billion yen (approximately $11.754 billion), a significant increase of 29.53% compared to the 1,178.5 billion yen in the fiscal year of 2020; operating profit was 216.2 billion yen (approximately $1.665 billion), soaring 3,176% compared to the 6.6 billion yen in the fiscal year of 2020; net profit was 106.9 billion yen (approximately $815 million), a substantial increase of 532% year-over-year.

The good times did not last, as Kioxia faced dual challenges from both internal and external factors.

The 2022 chip market experienced a severe winter.

The chip market in 2022 changed from the strong growth of 26.2% in 2021, with industry giants such as Intel, Samsung, Nvidia, and TSMC facing pressure from declining revenues and shrinking stock prices. The "supply shortage" of chips has reversed to "oversupply," and the global semiconductor industry has entered a downturn. According to Gartner data, the total global semiconductor revenue in 2022 was $601.7 billion, only a 1.1% increase from the $595 billion in 2021. The total revenue of the top 25 semiconductor manufacturers only grew by 2.8% in 2022.

In the second half of 2022, the global economy began to slow down under the multiple pressures of high inflation, interest rate hikes, rising energy costs, and China's pandemic lockdowns, affecting many global supply chains. Consumers also started to cut back on spending, leading to a decline in demand for personal computers (PCs) and smartphones. Subsequently, businesses also began to reduce spending due to expectations of a global economic recession. The sales volume of memory and NAND flash memory decreased by 10.4%.

Kioxia faced the most severe quarterly loss since its establishment during this winter. In the first three months of 2022, Kioxia's operating revenue was 245.2 billion yen, compared to 278.2 billion yen in the previous fiscal quarter, a year-on-year decrease of about 12%. In addition to the sluggish sales of storage chips for smartphones and computers, the inventory of memory for data center applications also began to be reduced.

The situation did not improve in the second half of 2022.

In the Q3 of 2022, Kioxia achieved revenue of 391.4 billion yen (approximately $2.835 billion), a 6.6% increase compared to the previous quarter and a 2.3% decrease year-on-year. The operating profit was 80.6 billion yen (approximately $584 million), a 5.3% decrease compared to the previous quarter and a 3.3% increase year-on-year; the net profit was 34.8 billion yen (approximately $252 million), an 18.3% decrease compared to the previous quarter and a 22.1% decrease year-on-year.In response to the sluggish market, Kioxia has taken measures such as reducing production, destocking, and lowering costs. Kioxia initiated a production cut plan in October 2022, reducing production by about 30% at its NAND Flash wafer factories in Yawata and Kitakami, Japan. Kioxia has also accelerated the research and development and production of the new generation BiCS FLASH 3D flash memory technology to enhance product performance and competitiveness. Kioxia has now launched BiCS FLASH 3D flash memory products with 112 layers, 162 layers, and 218 layers, and is preparing to produce higher-level products. Furthermore, Kioxia has introduced a plan to encourage early retirement for employees over 56 years old to optimize the human resource structure and reduce labor costs.

The prolonged downturn has led Kioxia to continuously reduce production, with executives stating, "There is hardly any time to make money now." Kioxia (formerly Toshiba Memory) became independent from Toshiba in 2018 with investment from U.S. investment fund Bain Capital and others. The power semiconductor and hard disk drive (HDD) businesses remained with Toshiba. At the time of its independence from Toshiba, Kioxia was burdened with about 600 billion yen in bank loans. Its poor financial situation made it difficult to make bold investments. This situation made it challenging for Kioxia to compete with "deep-pocketed" South Korean companies.

03

Kioxia Seeking a Way Out

In an effort to revitalize the market, Kioxia hoped to acquire Western Digital. As of the second quarter of 2023, Kioxia held a 20.0% share of the global NAND Flash market, while Western Digital held a 15.1% share. Together, they accounted for 35.1% of the market, higher than Samsung's leading 30.2% share. However, this merger was not a "strong alliance." The reason Kioxia and Western Digital accelerated merger negotiations in 2023 was due to poor earnings for both parties, with Kioxia reporting a loss of $1.3 billion in the first quarter of 2023, and Western Digital announcing a loss of $427 million.

News about the merger of the two companies has been ongoing, with information indicating that during negotiations, both companies preferred Kioxia's team to lead the management of the new company after the merger, while Western Digital's executives would also play a relatively important supportive role. It is expected that the new company after the merger will adopt a dual-board system, with senior executives from both flash memory chip manufacturers as members. The new company's headquarters will be located in Japan, and it is anticipated to be listed on the NASDAQ and Tokyo Stock Exchange in the future.

The idea of cooperation and mutual benefit is beautiful, but on October 26, 2023, SK Hynix expressed opposition to the merger between Kioxia and Western Digital. Subsequently, Western Digital announced that after the merger negotiations with Kioxia reached an impasse, the company would split into two independent publicly traded companies, focusing on the hard disk and flash memory markets. It is not difficult to understand SK Hynix's opposition, as a successful transaction between Western Digital and Kioxia would inevitably threaten the position of SK Hynix, the world's third-largest NAND flash memory manufacturer. SK Hynix's CFO stated, "This deal undervalues its stake in Kioxia. Given the impact on the overall value of the company's investment in Kioxia, we currently do not agree with this transaction."

As mentioned earlier, SK Hynix is a major shareholder of Kioxia. However, since 2021, Western Digital and Kioxia have been intermittently negotiating the merger of their NAND flash memory production businesses to create the world's largest NAND flash memory manufacturer. The negotiations in 2023 were more aggressive, seemingly determined to succeed. Kioxia's "restless" heart is already well-known, and now that the merger has been put on hold, its relationship with SK Hynix seems to have become awkward. Having offended its financial backer, Kioxia is in urgent need of demonstrating loyalty.

04

Where Will Kioxia Go From Here?Recently, reports have indicated that Kioxia Corporation has proposed to produce chips for SK Hynix. According to the reports, Kioxia has offered a deal to SK Hynix to use its wafer fabrication facility in Japan to manufacture NAND flash memory for the latter, hoping that SK Hynix might reconsider its stance on the merger plan between Western Digital and Kioxia's NAND flash memory production business. However, it remains to be seen whether SK Hynix is interested in this proposal.

Kioxia, which has been ceding territory and paying compensation, is merely a microcosm of the declining Japanese semiconductor industry. Since their defeat in the storage market battle in the 1990s, it seems that Japanese chip companies have lost their fighting spirit. As the importance of semiconductors grows, the Japanese government has bet $67 billion in an effort to become a global chip powerhouse once again. Japan has brought in TSMC, and on February 24, TSMC's Kumamoto factory is set to open. Reports suggest that the expected total capacity of the Kumamoto plant will reach a scale of 40-50Kwpm, with the main process being 22/28nm, and a small amount of 12/16nm. At the same time, Japan is also collaborating with global giants in the hope of producing 2nm chips in the country.

On February 23, Asahi Shimbun reported that merger negotiations between semiconductor manufacturer Western Digital and Kioxia are expected to resume by the end of April. This may be significantly related to the various efforts of the Japanese government. However, whether this merger will be successful and what changes it might bring after success are still uncertain. In this back-and-forth scenario, how much of Kioxia's competitiveness remains is a question that may not have an optimistic answer.