Are Markets Still Worried About an AI Bubble?

What It Covers

Meta and Microsoft earnings reveal diverging AI strategies as markets evaluate bubble concerns. Meta increases CapEx to $135 billion with strong ad revenue growth, while Microsoft faces criticism for conservative infrastructure spending despite $625 billion cloud backlog. SoftBank pursues another $30 billion OpenAI investment as memory chip prices surge 120% amid unprecedented AI demand.

Key Questions Answered

• •Enterprise AI spending patterns: Meta commits $135 billion CapEx for 2025, nearly double prior year spending, while Microsoft increases CapEx 66% to $37.5 billion but faces criticism for underbuilding. Markets reward aggressive infrastructure investment when paired with revenue conversion, as Meta stock rises 8% versus Microsoft's 5% decline despite both beating earnings estimates.
• •AI wearables market opportunity: Meta Ray-Bans sales tripled year-over-year, positioning AI glasses as the smartphone moment for wearables. Zuckerberg frames billions of people wearing vision correction as an addressable market where most glasses will become AI-enabled within several years, validating hardware as a viable AI consumer entry point beyond software applications.
• •Memory shortage economics: Samsung and SK Hynix both doubled operating profits to $13.5-13.7 billion as DRAM prices expected to rise 120% and NAND 90% in 2026. Memory manufacturers maintain conservative CapEx despite demand, with analysts noting hyperscalers evaluate memory costs through different lens than consumer electronics, creating sustained pricing power for chip makers.
• •Microsoft competitive response velocity: Anthropic's Claude CoWork triggered emergency meetings at Microsoft, with product leaders noting superior Excel and PowerPoint capabilities versus Copilot. CEO Nadella now holds weekly standing meetings to demo competing lab features, with multiple divisions building CoWork-like prototypes in days, highlighting startup execution speed advantage over enterprise infrastructure constraints.
• •OpenAI infrastructure dependencies: OpenAI represents 45% of Microsoft's $625 billion cloud backlog with new $250 billion commitment, raising concentration risk questions about OpenAI's ability to fulfill financial obligations to Microsoft, Oracle, and other providers. This dependency shifts from Microsoft's early mover advantage narrative to concerns about customer diversification and revenue realization from AI investments.