SoftBank is rallying more than 30 top Japanese manufacturers—including Asahi Kasei, Fujitsu and Yaskawa Electric—to fund a new domestic AI unit, signaling a strategic shift toward “physical AI” and a bid to reduce reliance on foreign AI platforms.
SoftBank’s Homegrown AI Venture Attracts Over 30 Japanese Industry Leaders

SoftBank Group announced that a consortium of roughly 30 Japanese corporations from chemicals, robotics, automotive and electronics is evaluating a joint investment in its newly created AI subsidiary. The consortium includes heavyweights such as Asahi Kasei, Fujitsu, and Yaskawa Electric, each of which has signaled a willingness to commit capital and engineering resources to the project.
Market Context
- Domestic AI push: Japan’s AI market is projected to reach ¥15 trillion ($93 billion) by 2030, driven by demand for manufacturing automation, smart logistics and autonomous vehicles. Yet the country remains heavily dependent on foreign AI models, especially those from the United States and China.
- SoftBank’s AI pivot: After a 2024‑2025 period in which SoftBank doubled down on a partnership with OpenAI, the group reported a four‑fold increase in annual profit, largely from its Vision Fund’s stakes in AI‑related startups. The new venture marks a strategic diversification toward “physical AI” – AI that is tightly coupled with robotics and industrial hardware.
- Capital appetite: The potential investors collectively represent ¥2.5 trillion in annual revenue, indicating a sizable pool of financial and technical capital that could be mobilised quickly.
What It Means for the Industry
- Reduced reliance on foreign models – By developing AI capabilities in‑house, Japanese manufacturers can avoid licensing fees and data‑privacy constraints associated with overseas platforms. This aligns with the government’s “Society 5.0” agenda, which calls for domestically sourced AI to power critical infrastructure.
- Accelerated hardware‑AI integration – Companies like Yaskawa, a leader in industrial robotics, are likely to embed the new AI directly into its next‑generation controllers, shortening the time from algorithm development to factory floor deployment.
- Potential for new revenue streams – If the venture succeeds in commercialising AI‑driven services—predictive maintenance, quality‑inspection vision systems, or autonomous logistics—participants could capture recurring SaaS‑type income, diversifying away from traditional product sales.
- Competitive pressure on rivals – Domestic rivals such as Toyota’s AI lab and Hitachi’s Lumada platform will need to accelerate their own AI initiatives to stay relevant, potentially sparking a wave of M&A activity in the Japanese AI ecosystem.
- Supply‑chain implications – SoftBank has hinted at developing data‑center batteries that avoid lithium and cobalt, a move that could dovetail with the AI venture’s need for energy‑efficient compute. Manufacturers in the consortium could become early adopters of these battery technologies, creating a vertically integrated AI‑hardware‑energy ecosystem.
Strategic Outlook
The initiative reflects a broader trend among Japanese conglomerates: embedding AI at the core of physical products rather than treating it as an add‑on service. By pooling resources, the participating firms aim to achieve economies of scale that would be unattainable individually, while also sharing risk in a market that still faces talent shortages and high R&D costs.
If the consortium proceeds, SoftBank could raise ¥200 billion ($1.2 billion) in the first round, providing enough runway to recruit top AI talent, acquire niche startups, and build a dedicated compute infrastructure. Success would not only bolster SoftBank’s reputation as an AI catalyst but also set a template for industry‑wide collaboration in other high‑tech sectors.
*The article draws on data from Nikkei Asia and SoftBank’s recent financial disclosures. For further details on SoftBank’s AI strategy, see the company’s 2025 Vision Fund report and the Japanese Ministry of Economy, Trade and Industry’s AI roadmap.*

Comments
Please log in or register to join the discussion