Current Affairs Topics Archive
International Relations Economics Polity & Governance Environment & Ecology Science & Technology Internal Security Geography Social Issues Art & Culture Modern History

SoftBank plans massive U.S. power plant for AI data centres

March 23, 2026 5 min read Science & Technology Economics GS3 (Science & Technology Infrastructure Economy)
On This Page ▼

What Happened

  • SoftBank Group announced plans to build a $33.3 billion gas-fired power plant at the US Department of Energy's Portsmouth site in Piketon, Ohio, to power AI data centres.
  • The plant will have a capacity of 9.2 gigawatts (GW), making it the largest single-location power generation facility in the United States.
  • Masayoshi Son, SoftBank's Chairman and CEO, announced the project at a ceremony in Piketon on March 20, 2026, describing it as part of a vision to "develop the smartest intelligence in the world."
  • A consortium of 12 Japanese firms (including Toshiba, Hitachi, Mizuho Bank, and Sumitomo Mitsui Banking Corporation) and 9 US companies (including Goldman Sachs) will invest a total of $500 billion in the data centre infrastructure.
  • The project is part of a broader $550 billion Japanese investment commitment to the US, negotiated as part of trade tariff discussions between the US and Japan.

Static Topic Bridges

AI Infrastructure and Energy Demand

Artificial Intelligence — particularly large language models (LLMs) and generative AI systems — is extremely compute- and energy-intensive. A single AI training run for a frontier model can consume as much electricity as hundreds of thousands of homes for months. Data centres that host these models require not only massive computing hardware (GPUs/TPUs) but also continuous, high-capacity power supply for both computation and cooling. This has created a global race to secure dedicated power infrastructure.

  • The IEA (International Energy Agency) projects that global data centre electricity consumption could double by 2030, driven largely by AI workloads.
  • SoftBank's 9.2 GW plant would alone exceed the combined installed power capacity of several small countries.
  • India's total data centre capacity stood at approximately 1.5 GW in 2025, projected to grow to 9 GW by 2030.
  • India's IndiaAI Mission has awarded contracts for over 34,000 GPUs (as of mid-2025), a small fraction of what large US and Chinese AI clusters deploy.

Connection to this news: SoftBank's massive investment exemplifies the infrastructure race that generative AI has triggered — where compute leadership requires co-locating energy generation with data centres at unprecedented scale.

SoftBank Group and the Vision Fund Model

SoftBank Group is a Japanese multinational conglomerate and technology investment company. Its Vision Fund — the world's largest technology-focused venture fund — has made it the dominant private financier of AI infrastructure globally. SoftBank's Vision Fund 1 (launched 2017, $100 billion) and Vision Fund 2 invest in AI, semiconductors, robotics, and platform companies.

  • SoftBank invested $41 billion in OpenAI (the maker of ChatGPT), acquiring approximately 11% stake — later raised to 13% with an additional $30 billion follow-on investment in February 2026.
  • SoftBank is also a key investor in Arm Holdings (chip architecture), which powers most smartphones globally.
  • The Vision Fund model aggregates capital from sovereign wealth funds (Saudi Arabia's PIF, Abu Dhabi's Mubadala) and institutional investors, then deploys at scale into technology bets.
  • Japan's investment in the US AI ecosystem reflects a broader "technology diplomacy" dynamic — investment as a tool of trade negotiation.

Connection to this news: The Ohio power plant project is consistent with SoftBank's strategy of building AI infrastructure end-to-end — not just investing in AI companies but creating the physical substrate (power, data centres) they depend on.

India's Data Centre and AI Energy Policy

India is rapidly building its AI and data centre ecosystem. The IndiaAI Mission (2024) provides compute access, datasets, and skilling. However, India faces a power-data centre nexus challenge: data centres need reliable, clean power, while India's grid still has significant renewable intermittency issues.

  • India's data centre capacity grew from ~375 MW in 2020 to ~1,500 MW in 2025 — a fourfold increase in five years.
  • The SHANTI Act (Sustainable Harnessing and Advancement of Nuclear Energy for Transforming India) was passed to enable small modular reactors (SMRs) to power data centres.
  • The Union Budget 2025-26 set a target of 100 GW nuclear capacity by 2047 (up from ~7-8 GW currently) under the 'Nuclear Energy Mission for Viksit Bharat.'
  • India's National Data Centre Policy is under the Ministry of Electronics and Information Technology (MeitY).
  • Data centres' load variability (rapid ramp-up and withdrawal) poses grid stability challenges distinct from conventional industrial demand.

Connection to this news: While SoftBank builds the world's largest AI power plant in the US, India faces a parallel challenge of matching AI infrastructure ambition with reliable energy provision — making nuclear and renewable co-location policies critical for India's AI competitiveness.

Geopolitics of Technology Investment

Large-scale cross-border technology investment is increasingly intertwined with trade diplomacy. The US-Japan investment package demonstrates how capital flows can serve as tools of bilateral negotiation — Japan's $550 billion US investment was linked to securing relief from US tariff increases under the Trump administration's trade policy.

  • The broader US-Japan framework echoes the CHIPS Act strategy — using investment commitments to secure supply chain partnerships and market access.
  • Foreign Direct Investment (FDI) in critical infrastructure (power plants, data centres, semiconductors) increasingly triggers national security review in many countries (e.g., CFIUS in the US, the National Security and Investment Act in the UK).
  • India has similar FDI screening under the Press Note 3 (2020) framework, which requires government approval for investments from countries sharing a land border with India.

Connection to this news: The $550 billion Japan-US investment deal illustrates how technology infrastructure is now a currency of geopolitical negotiation — a model India must navigate as it seeks foreign investment in its own digital infrastructure.

Key Facts & Data

  • Power plant capacity: 9.2 GW (gas-fired), at Portsmouth site, Piketon, Ohio
  • Estimated cost: $33.3 billion for the power plant alone
  • Total data centre investment by consortium: $500 billion
  • Broader Japan-US investment commitment: $550 billion
  • SoftBank's OpenAI stake: ~13% (after $64.6 billion cumulative investment as of February 2026)
  • India's current data centre capacity: ~1.5 GW (2025), target ~9 GW by 2030
  • IEA projection: Global data centre electricity demand to double by 2030
  • Consortium composition: 12 Japanese firms + 9 US companies (including Goldman Sachs)