Musings on Markets is the personal blog of Aswath Damodaran, professor of finance at NYU Stern and widely known as the "Dean of Valuation." Running since 2008, it publishes hands-on intrinsic-value teardowns of headline companies (SpaceX, Tesla, Nvidia) using his narrative-and-numbers DCF framework, plus periodic market-wide reviews.
This piece explains how the AI investment boom is driving U.S. stocks. Author Damodaran says stocks held up despite rising interest rates because earnings beat expectations and AI spending surged. He is 'bullish but cautious'—most of his portfolio is in AI-heavy stocks like Amazon and Meta, but all new money goes into short-term Treasury bills (safe government bonds). Key holdings: Nvidia (biggest AI chip winner, but he doesn't own it), Meta (spending heavily on AI, hurting near-term profits), and Amazon (both spends on AI and profits from cloud services).
One-sentence summary of the author's market view this period: U.S. equity resilience stems from earnings beats and AI capex, but the author personally chooses to be "bullish but cautious," hoarding cash with new funds. [Cautious]
The article notes that from September 2026, the 10-year Treasury yield surged from 4.75% to 5.29% (a monthly increase of 54 basis points), ranking among the top 10% of monthly gains since 1962. Yet, the total market capitalization of U.S. stocks still rose by $2.5 trillion, with the technology sector contributing nearly $1.5 trillion. The author attributes this resilience to higher-than-expected corporate earnings in 2026, alongside upward revisions to earnings expectations for 2027 and 2028. For the full year, the technology, energy, and materials sectors were the primary market drivers, though approximately 50% of companies within each sector posted gains year-to-date, and about 65% of listed companies saw share price declines in the third quarter.
The author states: "The resilience of stocks during the year could be traced to higher-than-expected earnings being reported by companies in 2026, and a concurrent increase in expected earnings in 2027 and 2028."
Using an implied equity risk premium model, Damodaran calculates that as of early October 2026, the expected return on stocks stood at 8.99%, with an equity risk premium of 3.70% (relative to the 10-year Treasury yield of 5.29%). The author points out that even as interest rates rose from 4.18% to 5.29% over the year, the expected stock return increased from 8.41% to 8.99%, indicating that market pricing remains supported.
The article emphasizes that AI capital expenditure is the central narrative behind corporate behavior and the market's resistance to rising interest rates, yet its presentation on financial statements is complex, with two major limitations: data gaps and off-balance-sheet entities. The author categorizes AI capex participants into two groups:
The author highlights a gray area: some AI builders are also suppliers (e.g., Amazon both invests in AI capex and benefits from its cloud space for data storage; Nvidia sells chips while also directly or indirectly investing in data centers).
The article warns that public market data underestimates the full scope of AI capex, as two key components are excluded. First, some AI giants are private companies, including Anthropic and OpenAI (with an estimated market value exceeding $1 trillion), whose revenue, earnings, cash flow, and invested capital do not appear in public financial reports. Second, some AI capex takes the form of joint ventures and off-balance-sheet entities, where accounting treatment (especially regarding debt) may not fully reflect the true risk to companies. Consequently, public company financial data understates the overall impact across all enterprises.
Despite the underestimation, public data still shows that U.S. listed companies' capital expenditures increased by $133.4 billion year-over-year in Q2 2026 (the most recent complete earnings quarter), representing a growth rate of approximately 36%. The growth is highly concentrated: the technology, communication services, and consumer discretionary sectors each posted quarterly year-over-year increases exceeding 50%. The table provided by the author shows total capex of approximately $505 billion in Q2 2026, compared to about $371.6 billion in Q2 2025. This surge is primarily driven by AI-related investments, with the top ten hyperscalers contributing over $2 trillion in AI capex.
The article points to earnings growth in the U.S. technology, energy, and materials sectors as the core market support, with AI capex serving as the main thread driving corporate behavior and cash flow changes. Investors should focus on: 1) the capex pace and earnings inflection points of AI builders (Meta, Alphabet, Amazon, Microsoft); 2) the revenue conversion efficiency of AI suppliers (Nvidia, TSMC, Broadcom, Micron, Marvel, Constellation Energy, Equinix, Digital Realty); and 3) the potential market impact from the listing or fundraising of private companies (Anthropic, OpenAI). As a valuation model builder, the author's implied equity risk premium model suggests current market pricing is within a reasonable range (ERP 3.70%), but readers should note this is an estimate based on public data, and the model itself does not constitute timing advice.
From Q2 2025 to Q2 2026, total book equity for U.S. stocks rose by nearly 13%, while total debt grew by approximately 8%. Specifically, book equity increased by $1.8 trillion, and book debt rose by $1.9 trillion. The technology sector, as the primary driver of AI capital expenditure, saw even more pronounced gains: book equity climbed about 30%, and total debt grew by roughly 18.8%. The author states, "Technology, being the most active player in AI cap ex, has seen much bigger increases in both numbers."
In 2026, AI capital expenditure is impacting earnings through supply chain effects: spending by companies building AI infrastructure becomes revenue and earnings for companies supplying AI components. The author notes that Nvidia has been the biggest beneficiary of the AI capex boom so far, with its chips (after significant markups) powering data centers. Additionally, beneficiaries range from electrical equipment manufacturers to utility companies and real estate developers. While AI builders incur higher amortization expenses, longer amortization schedules mitigate the current impact on earnings.
In Q1 and Q2 2026, total U.S. stock earnings grew by 35% and 57% year-over-year, respectively, yet more than half of the companies reported a decline in net profit. The specific data is as follows:
| Metric | Q1 2026 | Q2 2026 |
|---|---|---|
| Total Earnings (vs. same period 2025) | Increased from $511B to $692B | Increased from $576B to $904B |
| Year-over-Year Growth Rate | 35% | 57% |
Sector divergence is clear: Technology, Financials, and Communication Services saw the largest earnings gains; Healthcare, Utilities, and Real Estate lagged behind. The author emphasizes, "As with stock prices, the earnings benefits are not broad-based."
Earnings growth has not translated synchronously into shareholder cash returns. Over the trailing twelve months in 2026, dividends grew by approximately 5% year-over-year (+$43.3 billion), and share buybacks increased by about 9% (+$106.7 billion). S&P 500 companies returned only 63% of their earnings to shareholders over the past twelve months, the lowest level since 2004. The author points out that those who previously criticized companies for excessive buybacks are now criticizing them for reducing buybacks and increasing AI capital expenditure.
If AI business outperforms expectations, investing companies will generate incremental earnings exceeding their cost of capital; if it falls far short, accounting write-downs and market crashes will follow. The author analyzes two extreme scenarios in detail:
Intermediate scenarios include moderate AI success (earnings insufficient to cover the cost of capital) or moderate failure (some companies can justify the investment, but most cannot). The author judges that the market will react before accounting: "if AI is working better (worse) than expected, you should see stock prices at AI-centered businesses rise (fall) before you see accountants respond."
The current market and economy have become a massive bet on AI, leaving investors with four choices, but the AI debate is unavoidable. The author does not explicitly recommend specific actions but notes that success will validate companies and investors betting on AI, while failure will manifest as large-scale write-downs and stock price declines. Readers should be aware that this analysis comes from a position-holder's perspective—the author holds a positive view on the AI theme, and their discussion may understate the uncertainty of capital expenditure returns.
The author categorizes current AI investment strategies into four types and explicitly chooses to "follow the market consensus." These four paths are: going all-in on AI winners, following the market consensus, being an AI skeptic, and shorting the AI narrative.
The author allocates most of the portfolio to the "market consensus" path, but nearly all new funds over the past one to two years have been placed in cash. He holds five of the "Magnificent Seven" U.S. stocks (Amazon, Alphabet, Meta, Microsoft, and presumably Apple? The original text explicitly excludes Tesla and Nvidia), all of which are heavy investors in AI capital expenditure. However, over the past one to two years, new funds have primarily been allocated to short-term Treasury bonds (yielding 4%), resulting in a higher-than-usual cash ratio in the portfolio.
The author admits: "I have left money on the table undoubtedly by doing so, but it has helped me sleep better at night." He advises readers to find a path in the AI jungle that passes the "sleep test."
The author reminds that career income (human capital) is an unaccounted component of the portfolio, and young investors should adjust their financial capital allocation accordingly. If one believes AI disruption is about to impact their own career, they should shift financial capital away from businesses most vulnerable to AI to balance the portfolio. The author quips: "I am old enough not to care much about this component, since I have fewer working years left." But for younger investors, this could reshape the investment landscape.
The author explicitly outlines personal actions: core holdings remain aligned with the "market consensus" direction (holding AI-heavy stocks in the Mag Seven), but all incremental funds are hoarded as cash (short-term Treasury bonds)—a "bullish but cautious" compromise strategy. Institutional Perspective Bias: As an academic value investor, the author is naturally inclined toward conservatism and risk control, and his "sleep test" advice may underestimate the compounding effect of holding AI assets long-term.
| Ticker | Direction | Author's One-Sentence View | Key Data |
|---|---|---|---|
| Nvidia | Hold & Observe | The biggest beneficiary of AI capex, but the author personally does not hold it | Chips power data centers, sold at significantly marked-up prices |
| Meta | Hold & Observe | AI architecture builder, capex drags down current earnings | One of the hyperscale cloud providers; capex pace affects earnings inflection point |
| Alphabet | Hold & Observe | AI architecture builder, the author personally holds it | One of the hyperscale cloud providers; capex pace affects earnings inflection point |
| Amazon | Hold & Observe | AI architecture builder and supplier, the author personally holds it | Both invests in AI capex and benefits from cloud data storage |
| Microsoft | Hold & Observe | AI architecture builder, the author personally holds it | One of the hyperscale cloud providers; capex pace affects earnings inflection point |
| TSMC | Hold & Observe | AI architecture supplier, chip manufacturer | Leads chip manufacturing; revenue conversion efficiency is key |
| Broadcom | Hold & Observe | AI architecture supplier, networking equipment manufacturer | Revenue conversion efficiency is key |
| Micron | Hold & Observe | AI architecture supplier, networking equipment manufacturer | Revenue conversion efficiency is key |
| Marvel | Hold & Observe | AI architecture supplier, networking equipment manufacturer | Revenue conversion efficiency is key |
| Constellation Energy | Hold & Observe | AI architecture supplier, power plant developer | Revenue conversion efficiency is key |
| Equinix | Hold & Observe | AI architecture supplier, data center real estate developer | Revenue conversion efficiency is key |
| Digital Realty | Hold & Observe | AI architecture supplier, data center real estate developer | Revenue conversion efficiency is key |
| OpenAI | Not specified | Private company, expected market cap over $1 trillion; IPO or fundraising will disrupt market landscape | Expected market cap over $1 trillion |
| Anthropic | Not specified | Private company; IPO or fundraising will disrupt market landscape | No specific data provided |
| SpaceX | Not specified | Market consensus sees it as one of the AI winners; a public-market large language model company | No specific data provided |