SoftBank Confirms Entry into AI Computing Power Leasing in the U.S.
SoftBank Group, led by Masayoshi Son, has confirmed that it will start leasing AI computing power to U.S. companies in the next fiscal year.
The group has decided to establish a new company, SB Neo, focused on providing GPU computing power for AI training and inference, aiming to achieve stable cash flow.
This move aims to seize the growing demand for AI infrastructure. SoftBank has previously made deep investments in the semiconductor and computing ecosystem through Arm and other ventures.
Source: Public Information
ABAB AI Insight
Masayoshi Son has made large-scale investments in technology and AI-related companies since the 2010s through the Vision Fund. SoftBank's acquisition of Arm has strengthened its chip design capabilities, laying the technical foundation for computing power leasing.
In terms of capital strategy, SoftBank is leveraging existing GPU resources and infrastructure to establish SB Neo, converting idle or new computing power into leasing revenue, reducing single investment risks, and capturing recurring revenue amid the AI boom.
Similar to Amazon AWS transitioning from internal infrastructure to cloud services, SoftBank is in the process of transforming from an investment holding company to an infrastructure-as-a-service model, expanding its global computing power footprint by tapping into U.S. market demand.
Structural assessment: This essentially belongs to the restructuring of the industry chain. Traditional investment groups are reshaping the AI computing power supply chain through vertical integration of GPU resources and leasing models, shifting from capital allocation to direct infrastructure operation, achieving stable cash flow and ecological control.
ABAB News · Law of Cognition
Excess computing power is a cash machine, leasing converts investment returns.
Infrastructure pioneers lead the wave of downstream AI applications.
In capital-intensive industries, stable cash flow outweighs single bets.