Why U.S. Stocks *Aren’t* In A Bubble (Yet) | Aswath Damodaran
Speakers
Aswath Damodaran
— Professor of Finance, NYU Stern
Jack Farley
— Host, Monetary Matters
Summary
Aswath Damodaran argues the S&P 500 is richly priced but not in a bubble, with cash flows from dividends and buybacks supporting valuation. He sees a big market delusion in AI, especially private LLMs, while viewing the public AI architecture buildout as more protected. He explains portfolio moves in Nvidia, Microsoft, Tesla, and Oracle and discusses cloud, Meta, Alphabet, Apple, China, and gold.
- Damodaran says U.S. equities are overvalued but not bubble-like, and overvaluation alone has been a poor market-timing signal.
- He distinguishes AI architecture spending as more certain and protected from LLM product/service and private-startup risk.
- He sold Nvidia and Tesla, holds Microsoft, and views Oracle as overextended on AI spending.
- Cloud is described as utility-like and sticky, with AI a smaller portion of hyperscaler revenue than feared.
- Meta's AI ROI should be tracked through ecosystem engagement and advertising revenue.
- AI adoption may lower margins collectively by intensifying competition and passing savings to consumers.
- He is underweight China due government involvement and views gold as catastrophe and trust-loss insurance.
- Valuation should focus on cash flows including buybacks rather than P/E alone.
Ideas
S&P richly priced, not bubble
The S&P 500 is richly priced and discounting a benign economy, but it is not in bubble territory: earnings have been resilient, the equity risk premium is around historical median, and dividends plus buybacks produce a healthier cash yield. He warns that using overvaluation or high P/E as a market-timing sell signal has been costly, and a bubble is plausible but not probable.
AI infrastructure protected by spent capex
The AI architecture buildout is the lower-uncertainty half of the AI trade: chip companies, data centers, and power providers such as Constellation Energy are already receiving spending, and customer mistakes become sunk costs, so they are relatively protected even if the LLM/product side corrects. Growth would level off if new AI money slows.
Nvidia too rich; sold and avoid
Nvidia benefited from the AI architecture buildout, and chip suppliers are protected because customers have already spent the cash and cannot reclaim it. However, Damodaran considers Nvidia richly priced, requiring too much to go right to buy at current prices, and he sold his last stake after scaling out over four years.
Hold Microsoft; path more reachable
He has owned Microsoft since 2014 and keeps it fully in his portfolio because its plausible path to justify valuation is more reachable than Nvidia's, its cloud business is a sticky, utility-like necessity, and AI-related revenue is a smaller piece than feared. He would not buy at today's price but can tolerate the overvaluation given taxes and long-term holding.
Cloud utility-like; AI bust slows only
Cloud is a utility-like necessity and sticky for customers. AI is only a small part of cloud revenue for hyperscalers, so even an AI spending bust would slow cloud growth rather than cause revenue declines; the longer-term risk is margin pressure as cloud matures and competition increases.
Alphabet capital allocation key; watch
Alphabet remains heavily dependent on advertising and has failed to make most side bets stand alone; its corporate VC structure may prevent ruthless loss-cutting. Gemini is a better AI execution and YouTube is a hugely valuable ecosystem asset, but it remains to be seen whether other bets like Waymo can deliver standalone value.
Meta AI lifts engagement and ads
Meta's AI capex is for its own ecosystem rather than external cloud customers, so ROI should show up through greater user engagement in its apps. If AI lifts average time spent from around 57 minutes toward 90 minutes, advertising revenue should benefit.
Oracle's trillion-club AI bet overvalued
Oracle has made an outsized AI spending bet, separating itself from other large-cap tech spenders, and appears overvalued because management wants to join the trillion-dollar club. It resembles a gambler putting all chips on one number, creating high execution and return risk.
Alibaba's US exposure low; watch
Alibaba's direct US revenue exposure is around 10% or less, so blanket tariff and China bearishness may overstate the direct earnings impact. The point is to distinguish narrative from actual revenue exposure.
Tesla political, narrative-dependent; avoid
He exited Tesla because its valuation requires a shift into higher-margin robotaxi, robotics, and ride-sharing businesses he cannot underwrite with confidence, and it has become a political stock where politics can affect product demand. Those reasons still hold despite the stock rising after his exit.
Underweight China due state involvement
He is deliberately underinvested in China because the government is a player in every industry narrative, making politics and state involvement unavoidable risks. He avoids business narratives where politics is part of the investment case.
Apple better without huge AI capex
Apple is in better shape than hyperscalers spending tens of billions on uncertain AI infrastructure because it is not making the same huge AI capex bet. If AI capabilities prove cheaper or different than expected, its more capital-light position is an advantage.
Gold as trust-loss catastrophe insurance
Gold is the asset investors turn to when institutional trust erodes, and the current catastrophe-insurance demand base is broader than historically, including investors who would not normally buy it. This suggests markets see higher catastrophe risk, making gold useful insurance.
Up Next
This Monetary Matters video, published January 25, 2026, features Aswath Damodaran, Jack Farley discussing SPY, DTCR, CEG, NVDA, MSFT, SKYY, GOOG, META, ORCL, BABA, TSLA, FXI, AAPL, GLD. 13 trade ideas extracted by AI with direction and confidence scoring.
Speakers: Aswath Damodaran, Jack Farley · Tickers: SPY, DTCR, CEG, NVDA, MSFT, SKYY, GOOG, META, ORCL, BABA, TSLA, FXI, AAPL, GLD