Masayoshi Son wants SoftBank to be the world's No. 1 in artificial superintelligence and rules out retiring

By Japan Today (via Kyodo) · June 24, 2026.

Masayoshi Son, chairman and CEO of SoftBank Group, declared on Wednesday at the company's annual shareholders' meeting held in Tokyo that his ambition is to make the company the dominant global supplier in the emerging field of artificial superintelligence (ASI). He is 68 years old and, although he had himself said in the past that he would retire while in his sixties, he has ruled out any retirement plans in the short or medium term: "I have no time to retire," he stated, adding that he intends to remain at the helm of the company for the next 10 to 15 years.

Son framed his strategy around four fundamental pillars in which he wants SoftBank to be "the absolute global number one": AI models, chips, infrastructure and robots. This roadmap is not new, but the shareholders' meeting served to publicly ratify it with concrete figures and with the backing of record financial results that lend credibility to the group's most ambitious bets.

**The bet on OpenAI: $64.6 billion before October**

One of the most notable data points in Son's address was the quantification of SoftBank's total investment in OpenAI, the company that created ChatGPT. According to the CEO, that investment will reach $64.6 billion before October 2026. This is an extraordinary figure that makes SoftBank one of the largest investors, if not the largest, in Sam Altman's company. This amount far exceeds the individual funding rounds that OpenAI has closed in recent years and places the Japanese conglomerate in a position of structural influence over the development of the world's most advanced AI models.

As sector context, OpenAI has gone from being a nonprofit organization to becoming one of the most sought-after technology assets on the planet. SoftBank's massive injection of capital grants it considerable bargaining leverage, but it also raises questions about the lab's independence and its original mission to develop AI in a safe and beneficial way for humanity.

**Infrastructure: data centers in the U.S. and Europe, and the alliance with TEPCO to bring them to Japan**

Beyond its stakes in AI companies, SoftBank is building or financing the construction of massive data centers in both the United States and Europe. This infrastructure strategy targets directly the most obvious bottleneck in the AI ecosystem: computing capacity. Without enough data centers with guaranteed power and cutting-edge chips, not even the best models can be deployed at scale.

In this regard, Son revealed that SoftBank Corp. —the group's telecommunications subsidiary— is currently negotiating a capital alliance with Tokyo Electric Power Company Holdings (TEPCO), Japan's largest utility. If the deal materializes, the resulting increase in electrical capacity would make it possible to bring large AI data centers to Japanese territory itself, something that has until now been a significant obstacle given the high cost and limited availability of power in the country. This negotiation, if confirmed, would imply an unprecedented convergence between Japan's traditional energy sector and the new AI economy.

**Physical robots already in mass production**

In the field of robotics —the fourth pillar of the strategy—, Son made a concrete revelation: the "physical AI" program, that is, robots powered by artificial intelligence models, has already begun mass production. "I hope to make an official announcement soon," he added, without providing further details at this time. This development, if confirmed, would mark a milestone in the transition of AI from purely digital environments to the physical world, one of the great frontiers of agentic AI.

As sector context, AI-powered robotics is one of the fastest-growing and most heavily invested fields globally. Companies such as Figure AI, 1X Technologies, Agility Robotics and Boston Dynamics are in a full-blown race to create humanoid robots capable of carrying out tasks in industrial and domestic environments. SoftBank already has experience in this field through its historic stake in Boston Dynamics and the Pepper robot. Son's statement suggests that the group is taking a decisive step toward commercial production at scale.

**Record financial results: 5 trillion yen in net profit**

The context in which these statements are made is especially favorable for SoftBank: the group posted a net profit of 5 trillion yen (equivalent to about $31 billion) in the fiscal year ended in March 2026, representing a more than fourfold increase over the previous year and constituting the company's all-time high. This spectacular recovery —recall that SoftBank suffered hefty losses in the years of the Vision Fund's collapse, with failed investments such as WeWork— is what gives real financial ammunition to Son's ambitions.

A profit of that magnitude not only demonstrates that the technology investment model is producing returns, but also gives Son the freedom to keep betting heavily on AI without the need for defensive rationalizations before shareholders.

**Implications for agentic AI**

SoftBank's strategy is directly relevant to the agentic AI ecosystem for several reasons. First, the AI models that underpin autonomous agents require massive infrastructure: SoftBank's investment in data centers and in OpenAI translates directly into more computing capacity available for the most advanced agents. Second, the bet on robotics as "physical AI" represents exactly the extension of agentic AI to the physical world: robots that perceive, reason and act autonomously in unstructured environments are, in essence, AI agents with a body.

Third, the combination of chips (SoftBank has a stake in Arm Holdings, the processor designer), models (via OpenAI and other holdings), infrastructure (its own data centers) and robots (stakes in robotics companies) configures a vertical technology stack that, if successfully completed, would give SoftBank an unrivaled position in the deployment of agentic systems on an industrial scale.

**Comparison with competitors in the race for ASI**

Son's ambition to be "number one in ASI" places him in direct competition with the major AI labs (OpenAI, Anthropic, Google DeepMind) and with the tech giants that fund them (Microsoft, Amazon, Alphabet). The difference is that SoftBank is neither a research lab nor a big tech company with mass-consumer products: it is an investment conglomerate seeking to be the principal capital allocator of the ASI ecosystem.

In this sense, Son's strategy resembles that of a sovereign wealth fund or an industrial holding company more than that of a traditional technology company. The idea is that, if ASI is the most valuable asset of the 21st century, whoever controls the stakes in the key players and the infrastructure that supports them will have a structural advantage that no individual product or technology can replicate.

**The Son factor: who else bets so long-term?**

The statement that Son plans to remain at the helm of SoftBank for the next 10 to 15 years is, in itself, a relevant signal for the market. At 68, the Japanese CEO projects a horizon of action that would extend his leadership into the late 2030s, when —if the industry's most optimistic predictions come true— ASI could already be an operational reality.

This continuity in leadership eliminates, at least rhetorically, the uncertainty over succession and reinforces the strategic coherence of the long-term bets. An investor who knows he will once again sit in that chair ten years from now makes structurally different decisions than one who plans to leave in two or three. The question, of course, is whether the pace of technological change will allow a strategy set in 2026 to remain relevant in 2036.

**Regulatory perspective**

SoftBank's profile does not fit directly within the scope of the EU AI Act in its strictest form, but its investments do. As sector context, European regulation classifies AI systems according to the risk they pose to fundamental rights. General-purpose AI models (GPAI) such as those of OpenAI are subject to transparency obligations and, in the case of models with greater systemic impact capacity, to more rigorous risk assessments. A shareholding exposure of $64.6 billion in OpenAI makes SoftBank an actor with a direct interest in how that regulation evolves.

Likewise, the construction of data centers in Europe will be subject to GDPR data protection standards and to EU energy and environmental rules, which adds layers of regulatory complexity to the group's infrastructure ambition.

**Outlook**

If SoftBank manages to execute the strategy described by Son —investing massively in OpenAI, deploying data center infrastructure across three continents, bringing mass-produced AI robots to market and securing energy supply in Japan via TEPCO—, it will have built a singular position in the artificial superintelligence economy. It would not be the creator of ASI, but it could be the largest owner of the capital and infrastructure surrounding it.

SoftBank's recent history, however, invites caution: Vision Fund 1 was a large-scale investment experiment that ended with multimillion-dollar losses in companies such as WeWork, Greensill and Katerra. The difference now, optimists argue, is that the bet is more concentrated and in a sector —AI— that is producing real and measurable returns. Skeptics, by contrast, recall that the dot-com bubble also seemed different until it stopped being so.

What is beyond doubt is that Son's statements at this shareholders' meeting mark a milestone of strategic intent: SoftBank is not merely investing in AI; it intends to lead the era of artificial superintelligence from a position of capital, infrastructure and hardware. Whether that amounts to being the "number one in ASI," time will tell.