Aswath Damodaran on The AI Spending Spree: Bubble, Boom, or Both? | #619

What It Covers

NYU professor Aswath Damodaran analyzes the AI spending surge by Magnificent Seven companies, examining whether $600 billion in collective CapEx represents rational investment or dangerous overconfidence. He covers OpenAI's missing business model, software sector disruption, market trust erosion, sports franchise valuations, and why holding cash currently makes sense given richly priced equity markets.

Key Questions Answered

• •AI CapEx Risk Assessment: When evaluating Mag Seven AI spending, focus on the financing section of earnings reports, not just CapEx announcements. Companies funding AI infrastructure through private credit rather than cash flows create systemic risk. If investments fail, losses ripple beyond shareholders into lenders. The Mag Seven currently have sufficient cash flows, but smaller AI infrastructure companies borrowing through private credit markets represent the real danger zone.
• •Reverse-Engineer Market Cap Validation: Before investing in trillion-dollar companies, calculate what revenue growth rate their current market cap requires, then ask whether the total addressable market can ever support that revenue figure. At December 2025 valuations, most Mag Seven companies required 15–20% sustained growth to justify their prices — plausible but at the outer limits of what already-scaled businesses can realistically deliver.
• •Implied Equity Risk Premium as Market Gauge: Damodaran calculates a monthly implied equity risk premium by back-solving from current stock prices what return investors are actually pricing in. At the start of 2026, that figure was approximately 8.4%, translating to a risk premium of roughly 4.1% — in line with the 75-year historical average. This suggests the market is not in bubble territory but is fully priced with no margin of safety for a messy global economic transition.
• •Innovator's Dilemma in Software: Established software companies like Salesforce and Oracle face a structural trap: AI can replicate their core functions at a fraction of the cost, but their high-margin existing businesses prevent them from fully pivoting. Investors should identify which software companies are actively cannibalizing their own products with lower-cost AI offerings versus those in denial — the former group represents the more viable long-term investment within the sector's ongoing selloff.
• •Big Market Delusion Pattern: Every major technological disruption produces collective overinvestment because multiple well-funded competitors each believe they will be the winner in a winner-take-all market. Damodaran states with confidence that the AI infrastructure buildout represents aggregate overinvestment — historically, two winners emerge while three or more large investors write off tens of billions. The danger is not that innovation fails, but that debt-financed losers create economy-wide contagion rather than contained shareholder losses.
• •Business Model Before Scale: Companies scaling to near-trillion-dollar private valuations without articulating a coherent revenue model represent a structural risk. OpenAI's projected growth from $4 billion to $146 billion in revenue by 2029 requires a licensing or platform business model — not ChatGPT subscriptions — yet leadership has not defined that model publicly. Investors should treat any company at massive scale without a clear monetization narrative as uninvestable regardless of growth projections or market enthusiasm.