Masayoshi Son: Missing Anthropic Investment Was a Mistake — SoftBank Can't Bet Solely on OpenAI
nashnova research
Masayoshi Son publicly called his failure to invest in Anthropic a mistake and signaled SoftBank's AI strategy is shifting from a single OpenAI bet to a full-stack platform play — a redefinition of SoftBank's role across the AI supply chain.
Why does Son call missing Anthropic a mistake?
Son revealed he met Anthropic founder Dario Amodei early on and considered him "an extraordinary person," but ultimately passed on the deal.
His team's concern at the time: "You can't chase two rabbits at once." SoftBank concentrated its capital on OpenAI instead.
This means → SoftBank's early AI-model bet was a deliberate single-name concentration — and the cost was missing the company that became OpenAI's biggest rival.
Focus vs. diversification — how does Son weigh the trade-off?
Son acknowledged that spreading bets across multiple companies is a sound strategy in principle, but becoming a true strategic partner requires heavy commitment to a single name.
"You have to go big on certain investments — otherwise you can't really call yourself a partner," he said.
In plain terms = it is the classic venture trade-off: spread thin and you lose strategic leverage; go heavy and you risk missing the next winner.
Why is SoftBank's AI strategy shifting from "back a model" to "build a platform"?
Son stated explicitly: "If we only support OpenAI, we cannot become a platform."
He framed SoftBank's AI strategy around four pillars — models, chips, data-center infrastructure, and physical AI — likening them to controlling the four corners of an Othello board: holding key nodes matters more than dominating any single domain.
This means → SoftBank is repositioning from "bankroller of one AI model" to "neutral infrastructure supplier" serving the entire ecosystem, Anthropic included.
What role does Arm play in this chessboard?
Son emphasized that SoftBank's Arm — a chip-architecture design firm whose designs underpin over 99% of the world's smartphone processors — provides a differentiated ecosystem advantage.
Arm's positioning bridges two ends: high-performance data-center processors on one side, edge-device chips powering AI agents on the other.
In plain terms = Arm doesn't make chips itself, but nearly everyone builds on its blueprints. If AI moves from the cloud to phones, cars, and robots, Arm is the thread that ties the whole chain together.
How does Son respond to "AI bubble" concerns?
He said bluntly: "AI has only just begun" — those calling it a bubble "are not seeing the essence."
He invoked the 2000 dot-com crash as a cautionary tale: SoftBank ran short of cash back then and missed the chance to invest at post-crash lows. "Fortunately, this time we have some financial room to maneuver," he said.
This reflects Son's core logic: current market turbulence is not the end — it is the window to lay the foundation for future growth.
"SoftBank is just an investment company" — how does Son push back?
Responding to that criticism, he said: "My goal is not to make money through investing — it is to create the future through investing."
He framed capital as a tool for building industries, not an end in itself.
This means → Son is trying to pull SoftBank's narrative back from "financial investor" to "industry builder." Whether that logic holds depends on whether its chip and infrastructure bets can truly become a neutral, multi-party platform.
市场有风险,内容仅供研究参考,不构成投资建议。
