Masayoshi Son, the chairman of SoftBank, explains the performance of corporations in Tokyo in Jul. 2016. /Courtesy of Reuters.

ARM, a subsidiary of SoftBank led by Masayoshi Son, is known to shift its corporate identity from a business model that sells semiconductor design only to launching its own chips, drawing attention to the impact it will have on the global semiconductor industry.

ARM, dubbed the hidden ruler of the mobile semiconductor market, boasts an overwhelming market share, with over 90% of the chips produced worldwide based on ARM architecture, including the application processors (AP) used in smartphones.

◇ ARM has radically changed under Masayoshi Son

Historically, ARM has pursued a low-cost high-volume business model that sells semiconductor design assets (IP) for various fields such as mobile, automotive, and the Internet of Things (IoT). Concerns have often been raised over the past decade that ARM might begin designing chips directly or launch its own manufactured chips, but ARM has dismissed these concerns by adhering to the principle of "not competing with clients."

However, since SoftBank Group, led by Son, acquired ARM in 2016, the nature of the company's business has gradually begun to change. As the artificial intelligence (AI) semiconductor boom started, ARM, which specializes in low-power design, has also been developing licenses necessary for high-performance designs for servers and data centers to secure future competitiveness.

Recently, as the AI market began to stir with the ChatGPT and DeepSeek phenomenon, ARM started seeking more dramatic changes. The Financial Times reported on the 14th, citing sources, that ARM Chief Executive Officer (CEO) Rene Haas is expected to unveil its first self-manufactured chip as early as this summer. ARM is said to have secured its first customer for the new chip, Meta Platforms, the parent company of Facebook.

◇ Diversifying from mobile-centric business to participate in the AI market as a competitor

The sudden shift in ARM's direction can be interpreted in two major ways. First, it is seen as a move to officially enter the rapidly expanding AI chip market propelled by the AI semiconductor boom. The industry analyzes that ARM's launch of its own chips is part of a larger plan to transition to AI chip production. In fact, ARM's first product is expected to be a central processing unit (CPU) for servers in large data centers.

In this case, the likelihood of directly competing with key clients in the mobile market, such as Samsung Electronics, Apple, MediaTek, and Qualcomm, seems minimal. Since ARM will be producing chips for servers, it is expected to encroach on the market share of companies like Nvidia, Intel, and AMD.

Another analysis posits that this is a long-term plan to change ARM's structural issue of a low-cost high-volume business to elevate revenue and profit scales. ARM's revenue and profit levels are relatively low compared to its market share. This is because it has only provided design licenses and has not been involved in the direct design, optimization, or manufacturing of chips. Recently, there have been consistent reports suggesting that ARM is significantly increasing its licensing fees.

Meanwhile, the industry is paying attention to the cooperative relationship among the three companies following the recent meeting between Lee Jae-Yong, Chairman of Samsung Electronics, Sam Altman, CEO of OpenAI, and Son. An industry source noted, "It has been suggested that Samsung Electronics might invest in the Stargate project, integrating Samsung's technological capabilities into OpenAI and ARM," explaining that "OpenAI could develop AI accelerators based on ARM's design assets, with Samsung Electronics handling production."