Yoshimitsu Goto, CFO of the Japanese multinational conglomerate SoftBank Group, has highlighted an overwhelming supply shortage of AI data centers, stressing the urgent need for large-scale infrastructure investment. This move is driven by the explosive demand for computation required to realize Artificial Super Intelligence (ASI) and the structural bottlenecks in power supply and equipment procurement.

The Reality of the “Overwhelming Shortage” as Defined by CFO Goto

Yoshimitsu Goto, Director and Corporate Executive Vice President and CFO of SoftBank Group (SBG), has articulated a clear vision regarding the current state of data centers in the AI era: they are facing an overwhelming supply shortage. In the “SoftBank Group Report 2025” published in July 2025, Goto explained that for an information revolution to permeate society, three elements are indispensable: technical innovation, service deployment, and infrastructure development.

Just as the development of telecommunications infrastructure like broadband and mobile networks was the key to growth during the previous internet revolution, Goto argues that the modern AI revolution urgently requires an infrastructure environment capable of handling advanced generative AI technologies. Specifically, to support the evolution of large-scale language models developed by entities like OpenAI, an enormous amount of computing power is required—on a scale incomparable to previous cloud services. CFO Goto positions the resolution of this severe infrastructure deficit and the securing of massive computing resources as the most critical tasks in building the social foundation of the AI era. Please refer to the diagram below.

Realizing ASI Through the Four Core Themes of Infrastructure

SoftBank Group is aiming for the realization of Artificial Super Intelligence (ASI)—intelligence that far surpasses human capabilities. CFO Goto has organized the elements necessary for this goal into four strategic themes that must work in tandem to usher in the ASI era.

• First: Development of semiconductor chips, the heart of AI (expanding AI chips with a focus on Arm)

• Second: Development of AI robots

• Third: AI data centers as the physical foundation to support these technologies

• Fourth: Power, which is essential for operating the data centers

This vision has evolved from merely building computing hubs into a national-scale infrastructure project that vertically integrates power, chips, and AI models. Goto has stated that without AI data centers, the evolution of generative AI like OpenAI’s models cannot be achieved. Consequently, he has indicated a policy of providing full financial support for the development of computing infrastructure. SoftBank Group is accelerating its transformation from its previous framework as an investment holding company into a platformer that controls the very foundations of the AI era.

The Global Reality of AI Infrastructure Investment

The 80 Trillion Yen US Plan and Strategic Expansion into France

In the United States, the primary battlefield for generative AI, SoftBank Group announced a data center investment plan in March 2026 totaling approximately \$500 billion (about 80 trillion yen), described as the largest in human history. This plan was unveiled by Chairman Masayoshi Son in Ohio and will be propelled by the Portsmouth Consortium, a coalition of 21 Japanese and American companies including Panasonic HD and Mitsubishi UFJ. Furthermore, in May 2026, the group revealed plans to invest up to 75 billion euros (about 14 trillion yen) in France to develop 5GW (gigawatt) of AI infrastructure, making it one of the largest in Europe.

The primary reason for choosing France for this massive investment is its unique energy profile. France leads the world in nuclear power, which accounts for approximately 67% of its energy mix, providing a stable and low-carbon baseload power source. Unlike the software-led IT industry of the past, the AI era is transforming into a heavy-industry-style infrastructure business where those who secure physical resources like power and land emerge as the dominant suppliers. SoftBank Group is aggressively deploying capital worldwide in search of these optimal power environments.

The Vertical Integration Strategy Controlling Power, Chips, and Models

The essence of SoftBank Group’s strategy lies in building a vertical integration model that controls the entire AI value chain. CFO Goto explained that a massive economic zone is created by combining Arm’s semiconductor technology, OpenAI’s advanced AI models, and the gargantuan data centers and power infrastructure required to run them.

The specific components of this vertical integration include:

• Energy: Securing self-owned power sources, such as the construction of 9.2GW power plants in Ohio

• Chips: Supplying proprietary chips such as the Arm AGI CPU

• Compute: Implementing the latest GPU systems like NVIDIA GB200

• Models/Services: Maintaining a strong partnership with OpenAI through investments totaling over \$60 billion

By mastering everything from power procurement to semiconductor design and the utilization of AI models within the group, SoftBank aims to secure computing resources and cost competitiveness that competitors cannot replicate. CFO Goto expressed high financial expectations that this large-scale portfolio will mutually enhance values, leading to the growth of SoftBank Group’s Net Asset Value (NAV). Please refer to the diagram below.

The Physical Walls Blocking Construction and Future Outlook

Severe Delays Triggered by Power Procurement and Equipment Lead Times

Despite SoftBank Group’s ambitious vision to accelerate AI data center construction, physical constraints are emerging as significant barriers. Behind the supply shortage Goto pointed out are delays in power grid connections and a critical lack of major equipment. Analyses of the PJM region, the largest power market in the United States, indicate that while a data center building can be completed in about 1.5 years, obtaining permission to connect to the power grid takes an average of over seven years, with some recent projects reporting delays of up to eight years.

Furthermore, lead times for essential equipment have extended dramatically:

• Large power transformers: As of 2026, average lead times exceed 160 weeks (approx. 3 years)

• Step-up transformers for generators: Wait times are as long as 210 weeks (approx. 4 years)

• Gas turbines: Backlogs have surged, with wait times of around 3 years

The absence of even one of these components prevents a data center from operating. This temporal gap is the real reason many announced plans remain stalled. SoftBank’s adoption of a vertical integration model—building its own power plants—is an expression of urgency; they cannot afford to wait for utility-led infrastructure developments if they are to remain competitive in the AI race.

Future Outlook: Balancing Financial Health with Infrastructure Hegemony

The key focus for the future lies in SoftBank Group’s balancing act: how to maintain financial health while continuing these massive investments. CFO Goto has stated that the financial policy for the 2025 fiscal year remains unchanged, adhering strictly to the “ironclad” rule of keeping the Loan-to-Value (LTV) ratio below 25% during normal times. However, credit rating agencies such as S&P Global Ratings have pointed out that the increase in private equity investments is reducing financial flexibility and could put downward pressure on the company’s creditworthiness.

The future prospects of AI infrastructure investment include:

• Investment Returns: Whether these colossal capital outlays will generate sufficient future cash flow

• Power Procurement Capacity: The ability to secure independent power sources without relying on the public grid will determine a company’s competitive advantage

• Return to the Physical Layer: The primary battlefield of the digital economy is shifting back to physical infrastructure like power and heavy electrical equipment

CFO Goto has stated that SoftBank Group will deploy financial activities both cautiously and boldly under highly uncertain conditions to lay the foundation for NAV growth. The infrastructure hegemony over who will control the next-generation foundation of AI will be a crucial test that determines the structure of the global economy for years to come.

References

• [CFO Message (Yoshimitsu Goto) — SoftBank Group Report 2025 | SoftBank Group Corp.] https://group.softbank/en/ir/financials/annual_reports/2025/message/goto

[#SoftBankGroup #AIDataCenter #ASI #MasayoshiSon #InfrastructureInvestment #EnergyStrategy #ScienceAndTechnology #OpenAI]