Post

Conversation

Japan's less-recognized dominance in legacy power semiconductors is being challenged by emerging 🇨🇳 companies but local producers have been slow to form a united front. A major power chip alliance between Toshiba and Rohm has struggled to produce tangible results beyond a co-manufacturing project. Discussions on deeper collaborations, initially announced in early 2024, have stalled. Japan's power chip industry boasts 5 major players: Mitsubishi Electric, Fuji Electric, Toshiba, Rohm, and Denso, each of which has < 5% of the global market. Power chips are vital components in everything from power grids to EVs. Advanced power chips can significantly boost energy efficiency. Rohm invested ~$2B in Toshiba in 2023, as part of a 2 trillion yen acquisition led by Japan Industrial Partners and other, mainly domestic, companies and banks that took Toshiba private. Rohm and Toshiba have complementary strengths: EV chips and industrial products respectively. Rohm has struggled to make its next-generation silicon carbide (SiC) power chip profitable, as the EV market, the biggest user of the devices, has slowed. Fierce competition from emerging 🇨🇳 companies has also eaten into earnings. 🇯🇵 Renesas has abandoned plans to enter the SiC market due to sluggish EV market growth and competition from China. It was also hit by the bankruptcy of 🇺🇸 Wolfspeed, a maker of substrates for SiC chips that the Japanese company had a supply agreement with. The biggest reason power chip makers are struggling is the price race with 🇨🇳 companies. Rohm and Renesas both tried to expand their power chip manufacturing capacity in recent years, in hopes that rising EV demand would boost the market for chips built on SiC. But Japan has been much slower than China or Europe to embrace fully electric cars, a blow for power chip companies closely connected with 🇯🇵 automakers. 🇨🇳 companies have gradually built up the capability to manufacture more complex products and are now able to make end-to-end power chip devices on conventional single-crystal silicon. Some, such as TankeBlue and SICC, have even entered the market for making SiC substrates, which are higher grade than conventional silicon and primarily used in EVs. China's cheaper energy is helping these companies produce substrates at competitive prices. Energy accounts for 30% to 40% of the overall cost of making a substrate. Producing substrates is just half the battle. Printing circuits on them is much more technically challenging, especially for SiC, and China has yet to achieve the capabilities to make automotive-grade power chips. Still, given that China is the largest EV market, the country's up-and-coming chip makers have ample demand to feed, allowing them not only to mass-produce and reduce costs but also leverage client data to improve product quality. 🇨🇳 companies are also challenging the vertical integration model of existing companies. 🇨🇳 manufacturers are organized by process, while 🇯🇵 companies typically carry out end-to-end processes in-house. This means 🇨🇳 companies can focus their resources on specific production processes and improve efficiency. “The Chinese basically already dominate the silicon carbide substrate market, which means that nobody makes any profits in substrate ... The vertically integrated model is having a lot of issues.” “The entire substrate market is already disrupted by the business model of the Chinese.” While today's Japanese power chip companies are profitable, typically with operating profit margins of ~10%, they each held < 5% of the global market in 2024. “For Japan, it will be difficult unless companies join forces rather than going it alone, to gain the cost competitiveness needed to counter Chinese players.” The technological gap between Japanese and Chinese companies in silicon chips is likely 1 or 2 years and 3 in SiC chips, at the longest. That would mean there is little time for Japan to come up with a united front. asia.nikkei.com/business/techn
Image
Image