SoftBank's Son calls AI bubble question 'extremely stupid' in 2026

Masayoshi Son dismissed concerns about an AI bubble as "extremely stupid," doubling down on his bullish bet as SoftBank deepens its AI investment push.

SoftBank Group Corp. Chief Executive Officer Masayoshi Son called the question of whether artificial intelligence is a bubble "extremely stupid," reinforcing his aggressive investment stance as the Japanese conglomerate pours capital into AI through its Vision Fund and Arm Holdings.

"Asking whether AI is a bubble is extremely stupid," Son said during a shareholder meeting, according to a report. The SoftBank founder has been one of the most vocal AI bulls, directing billions into AI startups and infrastructure.

Son's comments come as SoftBank continues to deploy capital through its Vision Fund, which has backed companies including OpenAI rival Anthropic and AI infrastructure providers. Arm, in which SoftBank holds a roughly 90% stake, has emerged as a key beneficiary of the AI chip boom, with its architecture powering processors from Nvidia, Apple and Qualcomm.

The statement reinforces SoftBank's AI strategy at a time when skeptics point to mounting losses across the sector. OpenAI's net loss widened to $38.53 billion in 2025 from $5.09 billion in 2024, according to leaked audited financials verified by the Financial Times, raising questions about whether AI model providers can achieve profitability.

The Bull Case vs. The Numbers

Son's unwavering optimism stands in contrast to growing scrutiny of AI economics. While he argues the technology's potential justifies any valuation, the data shows a more complicated picture. OpenAI's losses surged more than sevenfold in a single year, and the era of "tokenmaxxing" — companies capping AI usage because of cost — has begun to weigh on demand. The Wall Street Journal reported that OpenAI is considering drastic price cuts as a result.

SpaceX, which positioned itself as an AI company in its IPO filing, estimated its total addressable market at $28.5 trillion, with $26.5 trillion — or 93% — expected to come from AI. Yet Morningstar equity analyst Nicolas Owens values the stock at $63 per share, a fraction of the $153 trading price, calling the bull case a "Moonshot scenario" that requires unproven technologies.

Bank of America analyst Ronald Epstein set a $235 price target on SpaceX but flagged "investment negatives" including unproven Starship technology, regulatory risk, and the need for significant investment to support AI buildout. The analysts noted that most of SpaceX's long-term AI opportunities depend on Starship successfully commercializing full reusability — an outcome that is far from certain.

A Divided Market

The gap between AI bulls and bears has rarely been wider. Major Wall Street banks that underwrote the SpaceX IPO — Goldman Sachs with a $205 target, Morgan Stanley at $300, and Bank of America at $235 — have issued bullish price targets. But even these banks acknowledged significant risks.

Son's dismissal of bubble concerns may boost sentiment among retail and institutional investors already heavily allocated to AI. But with AI model providers burning cash and the competitive landscape intensifying — Chinese open-weight models like GLM are closing the performance gap — the question of whether AI valuations reflect reality remains unresolved.

For investors, the stakes are clear. SoftBank's bet on AI through Arm and the Vision Fund ties the conglomerate's fortunes directly to the sector's trajectory. Arm's architecture is embedded in nearly every AI chip, from Nvidia's H100 to Apple's A-series processors, giving SoftBank exposure to AI demand regardless of which model provider wins. But if AI spending slows or valuations correct, SoftBank's concentrated bet could face significant headwinds.

This article is for informational purposes only and does not constitute investment advice.