The man who lost $70 billion on WeWork is now betting more than that on a single AI company, and this time his own team is nervous.
The Summary
- SoftBank's Masayoshi Son has committed over $60 billion to OpenAI, raising red flags among insiders worried about concentration risk and the founder's personal devotion to Sam Altman
- This represents Son's largest single bet in SoftBank's history, dwarfing even the Vision Fund's notorious portfolio companies
- Internal concerns center on whether Son's "starstruck" relationship with Altman is clouding judgment on what could be a company-defining wager
The Signal
Masayoshi Son built his reputation on bold bets. The $20 million he put into Alibaba in 2000 turned into $60 billion. But his $60 billion commitment to OpenAI isn't just big. It's existential. For context, SoftBank's entire market cap hovers around $90 billion. Son isn't diversifying into AI. He's going all-in on one company, one CEO, and one thesis about who wins the agent economy.
Inside SoftBank, people are worried. Not about OpenAI's technology. About the math and the man making the decision. Son's devotion to Sam Altman has some insiders spooked, according to Bloomberg reporting. This isn't due diligence. It's something closer to faith.
"Son's personal relationship with Altman is driving a bet that represents two-thirds of SoftBank's market value."
The WeWork comparison is unavoidable. Son famously fell for Adam Neumann's vision, pouring billions into a real estate company masquerading as a tech platform. That cost SoftBank $70 billion in losses. The OpenAI bet is structurally different, OpenAI actually builds technology that matters, but the pattern is familiar:
- A charismatic founder Son personally admires
- A valuation built on future dominance, not present cash flow
- A check so large it becomes impossible to walk away from
Son has always believed the biggest wins require the biggest conviction. Alibaba proved him right. WeWork proved the model has a failure mode. OpenAI will settle which lesson matters more.
The Implication
If Son is right, SoftBank becomes the primary capital partner to the company that builds the infrastructure layer for Web4. Every agent, every enterprise AI deployment, every business rebuilt around autonomous systems, OpenAI takes a cut and SoftBank rides along. If he's wrong, or even just early by three years, SoftBank faces a balance sheet crisis that makes the Vision Fund writedowns look like a rounding error.
For the rest of the AI investing world, this is the new baseline. Competition for OpenAI alternatives, Anthropic, Gemini, open-source models, just got a $60 billion subsidy to exist. No one else can afford to match Son's bet. But they also can't afford to let OpenAI win uncontested. Watch for a wave of defensive capital into every credible GPT alternative over the next 18 months.