‘Dean of Valuation’ Damodaran on Nvidia’s Multiple

Post-Nvidia earnings, Aswath Damodaran says AI is entering its “bar mitzvah moment,” with investors moving from promise-based valuation to demanding concrete evidence of progress. The commentary suggests a more evidence-driven risk assessment rather than a clear positive or negative catalyst for the broader market. As presented, the impact is likely limited to sentiment and positioning around AI winners, not an immediate market-wide repricing.

Analysis

The market is moving from paying for AI optionality to paying for audited monetization. That regime shift should widen dispersion inside the theme: dominant compute providers and equipment vendors with visible order books can keep earning a scarcity premium, while app-layer names that rely on future usage curves are more exposed to multiple compression if investors start demanding near-term ROI instead of TAM narratives.

Second-order, this favors the owners of real capacity over the storytellers. Hyperscalers can still defend capex because they can amortize spend across existing clouds, but smaller model builders, neo-clouds, and AI-native software names are more vulnerable to a higher effective cost of capital and slower customer conversion cycles. On the supply-chain side, networking, memory, and wafer-fab equipment should see the cleanest follow-through because proof-of-demand means actual deployment, not pilot budgets.

The contrarian risk is that this is not a bearish AI signal so much as a healthier one. If earnings continue to outrun the prior narrative, the cycle can extend for another 6-18 months with earnings growth doing the heavy lifting even as multiples compress modestly. The key falsifiers are a sequential slowdown in order growth, a pullback in hyperscaler capex guidance, or margin pressure that suggests the scarcity trade is already maturing; those would matter over the next 1-3 quarters, not just on the next print.

Key Decisions for Investors

• Long SMH / short ARKK for the next 1-3 months: cleaner exposure to the "proof over promise" regime, with semis likely to hold up better than duration-heavy AI software; risk is a broad rate rally that reflate growth multiples.
• Buy NVDA on a 3-5% post-event retrace with a 3-6 month horizon, preferably via call spreads rather than outright stock: upside remains intact if backlog/order commentary stays firm, but the reward/risk is poor if you chase strength after the gap.
• Short a basket of speculative AI software with limited current cash generation versus long NVDA or SMH: the market is likely to penalize names where revenue is still a story rather than a current operating metric; use this as a relative-value trade, not a directional AI short.
• Set an alert on hyperscaler capex and NVDA forward gross margin: if either rolls over, cut AI beta quickly; if both hold, the trade shifts from a sentiment squeeze to an earnings-driven rerating.