$AMZN is building AI demand without that dependence which makes AWS growth cleaner while the other two look like a leveraged bet on OpenAI. Source: Shay Boloor @StockSavvyShay FT

BNY just dropped a massive reality check. Everyone is celebrating the Magnificent 7's AI investments, but how are they paying for this revolution? The playbook is simple: Free Cash Flow OR... massive debt. (Look at Oracle's recent debt noise—it's a leading indicator). The Math That Makes Value Investors Sweat: The Mag 7 (ex-Tesla) forward P/E is nearly 30x. That's nowhere near "value investor comfort." The market is demanding a clear ROI by 2026. The pace of this AI buildout is directly tied to two things: Future Earnings and the Cost of Capital. The Domino Effect: If Margins Drop or Borrowing Costs Rise, the AI investment boom must slow down. That deceleration hits U.S. GDP hard. Hello, Federal Reserve and Government intervention? The feedback loop's timing will dictate how the entire equity market trades. 2026's Real Headwinds: Everyone expects lower rates, but don't forget the silent killers: Term Premiums Government Deficits (Crowding out private investment) Future Tax Risks The core question isn't whether AI is real—it's whether the current investment pace is sustainable until the returns finally justify the spending. This is the true AI bubble concern. Are we watching a self-sustaining cycle, or an investment spree built on borrowed time (and borrowed money)? Source: Neil Sethi

It's not about the current earnings—it's about the Capex future. And if that future changes, the market is in for a shock. Goldman Sachs just dropped a massive warning. The Core Risk: AI Spending Reversal They say that our current S&P 500 valuation is priced for an incredible, long-term AI Capex boom (spending on AI infrastructure). What if that boom stalls? If AI capex growth expectations revert to early 2023 levels, GS estimates the S&P 500 valuation multiple could see a 15-20% DOWNSIDE. That's a huge potential correction driven only by multiple compression. The Extreme Scenario (The Nightmare Fuel) -> Imagine the Hyperscalers slamming the brakes on spending. 🔴 The expected Capex for 2026 is approximately $433 billion. A reversion to the 2022 Capex level of $158 billion would result in a massive reduction of $275 billion—the "Lost Capex." 🔴This $275 billion shortfall represents a 30% reduction to the consensus estimate of $1 trillion in S&P 500 sales growth. Consequently, the expected S&P 500 revenue growth rate would drop sharply from the consensus of 6% to approximately 4%. ➡️ Ultimately, this decrease in spending would pose a substantial downside risk to both the AI investment trade and the broader S&P 500 market. The Takeaway for investors: This isn't about today's P&L. It's about the market's perception of tomorrow's AI-driven growth. A dramatic cut in capex would signal the long-term AI earnings thesis is broken, leading to a much steeper decline in stock valuations than a simple revenue reduction would suggest. 🔑 Don't miss this point: Near-term revenues might only drop modestly, but the hit to long-term earnings growth expectations will crush valuations. Source: Goldman Sachs, Neil Sethi @neilksethi