Per-MW pricing, regional variance, and cost drivers for owners scoping hyperscale & AI builds.
Salary benchmarks across the 14 mission-critical disciplines.
SoftBank Group said it spent approximately ¥968.9 billion during the first quarter of fiscal 2026 on U.S. assets tied to power generation and data centers, underscoring the scale of the company’s artificial intelligence infrastructure buildout.
The spending came during the three months ended June 30 and included approximately ¥637.8 billion in advance payments. The outlay pushed SoftBank’s purchases of property, plant and equipment and intangible assets to approximately ¥1.21 trillion for the quarter.
The investments were tied to Energy Global, a SoftBank subsidiary that develops, constructs and operates solar power plants and also develops and constructs data centers in the United States.
Hiring for a mission-critical build? Get a pre-qualified shortlist.
iRecruit.co specializes in construction recruiting for data center, energy, and advanced-manufacturing projects — project managers, MEP coordinators, commissioning leads, and more. We pre-screen every candidate so only qualified professionals reach your hiring team.
Get Started
Success-based pricing · 90-day replacement credit · No upfront fee on single roles
SoftBank said the spending was already visible on its balance sheet. Property, plant and equipment rose by approximately ¥725.9 billion from the end of March, mainly because of the acquisition of assets related to U.S. power generation and data centers.
The buildout also had a significant effect on reported liabilities and expenses. SoftBank recorded a ¥453.4 billion derivative loss tied to an increase in the fair value of warrants granted by Energy Global as incentives under lessor lease arrangements involving data centers. Those warrants can be converted into ordinary equity interests in Energy Global.
At June 30, derivative financial liabilities linked to the Energy Global warrants stood at approximately ¥1.05 trillion. That total included ¥177.3 billion in current derivative liabilities and ¥873.5 billion in non-current liabilities.
Expenses at the business increased by ¥183.6 billion, primarily due to higher cash-settled share-based compensation expenses after an increase in Energy Global’s valuation.
The U.S. infrastructure spending is taking place alongside SoftBank’s wider effort to build what it described as an AI ecosystem spanning capital, computing and physical infrastructure. Its AI Computing segment combines Arm, Ampere and Graphcore. Arm has also moved beyond its traditional semiconductor intellectual property model with the introduction of its first proprietary silicon chip, the Arm AGI CPU, designed for cloud and agentic AI workloads.
For the quarter, SoftBank reported net sales of ¥2.02 trillion, up 10.9% year-over-year. The group posted ¥1.86 trillion of investment gains, while net income attributable to owners of the parent fell 17.7% to ¥347.3 billion as higher expenses, financing costs, foreign exchange losses and derivative losses offset part of those gains.
Read the source
