TOKYO - SoftBank Group announced its consolidated earnings for the fiscal year ending March 2026 on May 13th, posting what was described as the highest quarterly profit ever recorded by a Japanese company, as attention increasingly focuses on Chairman and CEO Masayoshi Son’s aggressive strategy of betting heavily on artificial intelligence.
Speaking at last year’s shareholder meeting, Son outlined his long-term vision for the company, saying SoftBank aims to become “the world’s number one platformer in the field of ASI” within the next decade.
ASI, or Artificial Super Intelligence, refers to AI systems capable of surpassing human intelligence and potentially improving themselves autonomously. While experts remain divided on when such technology may become reality, Son has made clear that dominating this field has become SoftBank’s central ambition.
To support that goal, SoftBank has accelerated a series of massive investments in AI-related businesses. In 2025, the company announced plans to acquire U.S.-based semiconductor design firm Ampere Computing for 6.5 billion dollars, while also investing 20 billion dollars into OpenAI. This year, SoftBank unveiled an additional financing package for OpenAI alongside plans for a 500 billion dollar AI-focused data center project.
The group is also advancing “Stargate,” a large-scale infrastructure initiative aimed at building AI computing foundations across the United States to support OpenAI’s future operations.
Takashi Sugimoto, commentator for The Nikkei, said Son’s recent moves may appear sudden but are actually the result of preparations stretching back more than a decade.
According to Sugimoto, SoftBank’s strategy rests on four pillars: AI models, physical AI, AI chips, and infrastructure. AI models include systems such as ChatGPT, while “physical AI” refers to robotics and autonomous driving technologies. The infrastructure side primarily centers on data centers and semiconductor networks.
Sugimoto argued that Son’s approach has consistently focused on controlling both the “top” and “bottom” layers of technological ecosystems — meaning both the infrastructure and the services running on top of it.
He pointed to SoftBank’s early broadband business, Yahoo BB, in the early 2000s as an example. At the time, the company famously distributed free broadband modems outside train stations across Japan, rapidly expanding its internet infrastructure footprint. Simultaneously, SoftBank invested heavily in internet platforms such as Yahoo and Alibaba.
The same strategy later emerged in mobile communications. Before acquiring Vodafone Japan, Son reportedly approached Steve Jobs about developing what would later become the iPhone, showing that SoftBank’s ambitions extended far beyond simply owning telecom networks.
Sugimoto said Son was never interested in becoming “just a mobile carrier,” but instead sought to dominate the broader mobile internet platform ecosystem.
The company has also historically avoided large-scale manufacturing operations, preferring to outsource hardware production while focusing on the “brains” behind technology. Pepper, SoftBank’s humanoid robot introduced more than a decade ago, was manufactured by Taiwan’s Foxconn, while SoftBank concentrated on software and platform development.
Sugimoto noted that this philosophy traces back to SoftBank’s founding in 1981 as a software distribution company. Even then, Son sought exclusive agreements with both top software creators and distributors, reflecting the same “control both layers” strategy seen today in AI.
As for Son’s role model, Sugimoto said the SoftBank founder has long admired oil magnate John D. Rockefeller, founder of Standard Oil.
Rockefeller not only controlled oil production but also built the transportation infrastructure needed to support America’s rise as an automotive superpower. Sugimoto said Son has attempted to replicate this model in the information and AI industries by securing both the infrastructure and the applications before the market fully matures.
One key aspect of Son’s philosophy is making major investments before a technology reaches mainstream adoption. Rockefeller began building Standard Oil decades before automobiles became widespread, and Son appears determined to position SoftBank ahead of the anticipated AI boom in similar fashion.
Still, the strategy carries risks.
Sugimoto said Son himself once described the company’s greatest management risk as “if I suddenly die,” highlighting concerns over succession planning at SoftBank.
The company previously appeared to have a successor in Nikesh Arora, who joined SoftBank roughly a decade ago, but Arora eventually departed the group. At present, no clear successor has emerged.
Son had previously suggested he would retire in his 60s, but as he approaches 70, analysts increasingly believe he has abandoned any concrete retirement timeline.
Given his stated ambition of making SoftBank the world’s leading ASI platform within 10 years, Sugimoto suggested Son is likely to remain actively involved for at least another decade as he attempts to turn his long-envisioned AI empire into reality.
Source: テレ東BIZ
