The AI Debate Gets More Complicated: Microsoft Has a Win, Meta Stumbles | The Weekly Wrap

Summary

Steve Eisman argues the market’s AI enthusiasm has shifted from a simple capex-up = stocks-up narrative to a far more selective debate about moats, pricing power, and capital intensity. He distinguishes hyperscalers such as Microsoft, Amazon, Google, and Oracle from LLM providers like OpenAI and Anthropic, saying the former have real infrastructure businesses while the latter may face shallow moats and price pressure from cheaper Chinese open-source models. On earnings, he likes Microsoft and Amazon’s cloud acceleration, but says Meta’s huge AI spending is crushing margins and free cash flow, making the business model look much less attractive. He also points to AI-driven power demand benefiting Quanta and Bloom Energy, while noting credit risk is rising for names like CoreWeave. Outside AI, he says Charter’s broadband losses worsened, Visa and Mastercard showed solid consumer payment growth, and FICO, Apple, and PayPal presented a mixed picture.

Notes

• Investors sold tech and AI-related names as the debate turned more risk-aware.
• The AI trade is no longer uniformly bullish.
• Eisman frames capital intensity as the central AI bottleneck.
• He warns cheaper Chinese models could trigger price competition.
• He separates hyperscalers from LLM providers.
• Data-center scale itself functions as a moat.
• He sees only a handful of hyperscalers surviving at scale.
• Google, Amazon, Microsoft, and Oracle are the names he treats as having durable AI infrastructure businesses.
• He is more skeptical of Anthropic and OpenAI economics.
• Enterprise model-switching weakens LLM pricing power.
• Oracle’s OpenAI concentration is a monitoring risk.
• He says a weakness in OpenAI or Anthropic could hit the broader AI ecosystem.
• AI may lower software costs and spur startup formation.
• He argues AI is more likely to dislocate jobs than destroy net employment.
• He speculates job-loss narratives could be used to justify favorable regulation.
• Regulation could become a moat if incumbents shape the rules.
• SOXX has fallen sharply from its June 22 peak.
• The semiconductor group is under pressure alongside broader AI trades.
• CoreWeave is tapping debt markets for growth capital.
• Its debt is priced above 9%, signaling expensive financing.
• Fixed-income buyers are splitting large credits from smaller AI borrowers.
• Visa posted 20% EPS growth and 14% revenue growth.
• Visa’s payment volume growth suggests consumer spending remains resilient.
• Mastercard also delivered strong payment growth.
• Procter & Gamble showed flat organic growth and lower EPS.
• Bloom Energy sells distributed generators for data centers.
• Bloom’s growth was extraordinary, with EPS up 680% and revenue up 166%.
• Bloom crossed $1 billion in revenue for the first time.
• He notes Bloom’s valuation still compresses quickly as growth compounds.
• Quanta is benefiting from power-buildout demand tied to AI.
• Quanta raised full-year EPS and revenue guidance.
• Charter’s broadband losses worsened again.
• Eisman is selling Charter because the operating trend is still deteriorating.
• Meta’s AI costs overwhelmed revenue growth.
• Meta’s R&D rose from $13 billion to $22 billion.
• Free cash flow at Meta was only $784 million.
• Meta narrowed capex guidance to $130 billion-$145 billion.
• Meta disclosed $279 billion in future lease agreements tied mostly to AI.
• Meta lacks the cloud business that is cushioning Microsoft.
• Microsoft’s Azure and cloud strength contrasts with Meta’s thinner AI economics.
• Apple beat EPS but missed in services and China.
• Apple’s services and China exposure were the weak spots.
• Amazon’s AWS growth accelerated to 37%.
• Amazon rallied after the report, reflecting confidence in cloud demand.
• He says a $20 billion AI hedge fund was hit hard by the selloff.
• Leverage amplified losses in the AI correction.