2 hours ago
1
SoftBank Group and its telecom subsidiary are gearing up to lease AI computing resources to American companies beginning next fiscal year.
The infrastructure play behind the rental model
SoftBank has been writing some very large checks to make this happen. The company is pursuing a $100B AI infrastructure investment plan in the US, which involves massive GPU acquisitions and the leasing of large language models to enterprise customers.
SoftBank has also committed to investing around 45 billion euros over five years in AI data centers in France, with the potential to scale that figure to 75 billion euros for 3.1 gigawatts of capacity by 2031.
To bolster its hardware capabilities, SoftBank agreed to acquire Ampere Computing for $6.5B. Ampere designs energy-efficient server processors, and the deal is expected to close in the second half of 2025.
Son’s AI agent army and the compute it needs
Masayoshi Son’s stated ambition is to deploy up to one billion AI agents by 2025, with future goals stretching into the trillions.
SoftBank reported a net profit of roughly 1.153 trillion yen (approximately $7.77B) for the fiscal year ending March. That marked its first annual profit in four years, driven largely by growing AI-related optimism across its portfolio.
What the AI compute rental market looks like
SoftBank is entering a US compute rental market that already includes Amazon Web Services, Microsoft Azure, Google Cloud, Nvidia’s DGX Cloud service, and specialized GPU cloud providers such as CoreWeave and Lambda.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
English (US) · 