oftBank’s strategic direction has undergone a fundamental transformation, moving away from a traditional venture portfolio model toward a long-term provider of Artificial Super Intelligence (ASI) infrastructure. The core of this strategy is the metaphor of the 'Golden Goose,' which represents the internal organizational factory capable of continuously generating value (the 'eggs'). This approach addresses long-standing investor frustration regarding how to value a conglomerate that holds massive, volatile tech positions. By focusing on the 'goose,' Son argues that shareholders should value the company’s recurring ability to innovate rather than the snapshot valuation of its current holdings.
Masa Son’s 50-year vision is underpinned by the conviction that ASI will become the defining commodity of the 21st century. To capture this, SoftBank is directing unprecedented levels of capital into global infrastructure, specifically 5-10 gigawatt data center clusters and robotics manufacturing. The scale of these investments suggests that SoftBank is betting on power availability and compute density becoming the primary bottlenecks for ASI development globally. This is a departure from pure-play venture capital into a more capital-intensive, industrial model.
A key component of this vision is the idea that humanity will evolve alongside ASI rather than be replaced by it, with robotics playing a major role in solving global labor shortages. This belief justifies the aggressive focus on the 'Physical AI' layer—integrating software intelligence with robotic hardware to act upon the physical world. The discussion suggests that while the current market remains focused on digital assistants, the next phase of the investment cycle will shift toward autonomous labor in healthcare, research, and physical caregiving.
Despite the clear strategy, uncertainty remains regarding execution. The reliance on massive, unprecedented power projects poses significant political and regulatory risks that SoftBank must navigate across multiple jurisdictions. Furthermore, the market's historical difficulty in pricing intangible 'innovation factories' may continue to create valuation gaps. Nevertheless, Masa Son remains committed to the belief that the current AI build-out is merely the early stage of a much larger, multi-decadal shift that will eventually manifest in a quadrillion-yen economy.