Masayoshi Son was supposed to be retired by now.
For decades, the billionaire founder of SoftBank stuck to a famous 50-year life plan he drew up as a teenager. The final chapter was simple. Pass the reins to a handpicked successor in his sixties and watch the empire from a comfortable distance. Meanwhile, you can read similar stories here: Why The Market Risk On Mood Is Tearing Through Global Trading Floors.
Instead, at 68 years old, Son just told his shareholders that he's staying at the helm for another 10 or 15 years. He isn't leaving. In fact, he's getting greedier.
The reason is a single-minded obsession with what he calls Artificial Superintelligence (ASI). Son isn't just looking to invest in the next wave of technology. He's actively dismantling the old SoftBank structure to rebuild the entire company around his personal vision of a machine-dominated future. If you think the tech landscape is undergoing a normal transition, you don't understand the scale of what SoftBank is planning. This isn't the Vision Fund era of throwing billions at ride-sharing apps and co-working spaces. This is a highly leveraged, hyper-focused bet on the literal infrastructure of reality. To see the bigger picture, check out the detailed report by CNBC.
The Death of the Vision Fund Model
The old SoftBank operated like an chaotic venture capital machine on steroids. The $100 billion Vision Fund bought minority stakes in everything from Uber to WeWork. It was an ecosystem of disparate startups, bound together mostly by Son's checkbook and a vague mandate to change the world.
That model is dead.
Son is moving SoftBank away from being a mere financial investor and turning it into an operating holding company. He doesn't want pieces of other people's companies anymore. He wants absolute control over the fundamental blocks of artificial intelligence. To achieve this, he's ruthlessly reshaping the organization around four pillars: chips, software models, data center infrastructure, and physical robotics.
If a portfolio asset doesn't fit into this specific stack, it's essentially irrelevant to SoftBank's future. The company's historic $32 billion acquisition of British chip designer Arm in 2016 was once mocked by Wall Street analysts. Today, with SoftBank still controlling roughly 90% of Arm, that stake is the crown jewel driving Son's entire thesis. Arm processors are celebrated for their power efficiency, a metric that matters immensely when AI data centers threat to overwhelm electrical grids. Son didn't get lucky. He looked 40 steps ahead while everyone else was looking at next quarter's earnings.
Ground Control Versus Space Cadets
This shift in strategy explains why Son has turned into a brutal critic of alternative tech theories. Look at his recent clash with Elon Musk.
When asked by shareholders about Musk's ambitious ideas to place AI data centers in Earth's orbit, Son dismissed the concept with a stark financial reality check. He pointed out that processors cost far more than kilowatts. Launching heavy hardware into space, dealing with extreme maintenance issues, and enduring transmission latency simply makes no economic sense when the AI race is being decided right now on solid ground.
"He who strikes first wins," Son told his investors.
SoftBank is striking first by anchoring its capital into massive terrestrial projects. In early 2025, SoftBank committed $19 billion alongside OpenAI to the Stargate project, a massive $500 billion infrastructure plan aimed at building unprecedented computing power within the United States. While others talk about theoretical configurations a decade away, Son is securing land, batteries, and power grids today.
The Physical AI Monopoly
The ambition doesn't stop at data centers. To understand how Son is remaking SoftBank in his own image, you have to look at his latest acquisitions in physical AI.
SoftBank recently agreed to buy ABB Ltd.'s robotics division for roughly $5.4 billion. It also bought Ampere for $6.5 billion to bolster its server processor capabilities. Why buy a traditional industrial robotics unit? Because Son believes the ultimate expression of superintelligence isn't a chatbot on a screen. It's a physical machine capable of autonomous labor.
Think about the global labor shortage, particularly in aging societies like Japan. Factories can't find workers. Farms can't find successors. Son's answer is an ecosystem where Arm-powered chips run OpenAI-level intelligence inside thousands of factory-floor robots. To accelerate this, SoftBank is quietly planning to establish and list a dedicated AI robotics venture called Roze in the United States as early as this year.
This creates a terrifyingly complete monopoly pipeline:
- Silicon: Arm architectures design the chips.
- Infrastructure: SoftBank's "neocloud" data centers power the compute.
- Intelligence: Large-scale investments in OpenAI provide the brainpower.
- Execution: ABB and Agile Robots hardware perform the physical work.
Dismantling the Corporate Guardrails
To pull off a transformation this massive, you need total corporate obedience. The classic institutional structure of SoftBank—complete with cautious board members and traditional financial officers—has been steadily eroded.
Executives who preferred a diversified, safer investment approach have gradually washed out of the system. In their place stands a leadership team completely aligned with Son’s risk appetite. Rene Haas, the CEO of Arm, now sits directly on the SoftBank Group board, bridging the gap between chip architecture and corporate capital. Son has effectively cleared the room of anyone who might tell him "no."
This lack of friction is dangerous, but it's exactly how Son likes to operate. He's aiming for a net asset value of 1 quadrillion yen—roughly $6.4 trillion—within the next 16 years. You don't hit numbers like that by listening to conservative risk managers who worry about an AI bubble. Son openly calls bubble talk an "insult" to a technology revolution that has barely entered its third year.
What This Means for Your Capital
If you're an investor, you can no longer view SoftBank as a proxy for global tech startups. It is now a highly volatile, massively leveraged bet on a singular vision of the future.
If Son is right, and the demand for ultra-efficient Arm processors and physical automation scales a hundred-fold, SoftBank will become one of the most powerful corporate entities on Earth. If he's wrong, or if the massive capital expenditures required for AI infrastructure fail to yield commercial returns before the debt matures, the collapse will be historic.
The takeaway for anyone watching this space is clear. Don't look at SoftBank's quarterly earnings or its remaining legacy venture stakes. Watch its energy procurement contracts. Watch its factory automation partnerships. Most importantly, watch Masayoshi Son's health. The company is no longer an institution; it is the financial extension of one man's mind.
To track whether this bet is paying off, keep a close eye on the upcoming U.S. listing of their robotics vehicle, Roze. If the market swallows Son's physical AI pitch at a premium, he will have the fresh war chest he needs to fund the next stage of his empire. If it flops, the cracks in the superintelligence thesis will show sooner than he thinks.
