NYU Stern School of Business professor Aswath 'The Dean of Valuation' Damodaran sees a basic split running through corporate America in 2026. Soaring headline earnings have kept stock indexes buoyant despite surging yields.

At the same time, massive artificial intelligence spending is draining free cash flow and turning capital-light technology giants into heavy infrastructure plays.

How Capex Moves Through the Books

In his latest market reflection, Damodaran notes that capital spending at U.S. public companies reached $507.4 billion in the second quarter, up 35.66% from a year earlier.

Technology led the gains, up 79.88%, followed by communication services at 62.01%. Invested capital climbed $3.69 trillion, or 13.24%, as debt rose $1.91 trillion and book equity $1.77 trillion.

This capital boom operates through an asymmetric accounting loop between builders and suppliers. Hyperscaler "builders" such as Alphabet Inc. (GOOGL), Meta Platforms Inc. (META), Microsoft Corp. (MSFT), and Amazon.com Inc. (AMZN) capitalize outlays directly onto balance sheets, while longer amortization schedules being used by many of them are reducing the current hit to earnings.

“There is no way the year that you do capex that your earnings are going to go up,” Damodaran said.

But that spending becomes revenue for suppliers such as NVIDIA Corporation (NVDA), Taiwan Semiconductor Manufacturing Company Limited (TSM) and Broadcom Inc. (AVGO), and their rich margins carry it through to profit.

“That’s why AI capex will increase earnings,” he said. “It’s not at the hyperscalers, but it’s collectively across companies.”

Payout Ratios at 20-Year Lows

Since capex represents a direct, upfront cash deduction, aggregate cash burn among builders dwarfs net cash generation by suppliers.

“No matter how much you downplay the amortization effect on earnings, capex is capex,” Damodaran said. Over 12 months, dividends rose about 5% and gross buybacks about 9%, or roughly 7% net of share issuance.

Both trailed far behind a 49% surge in net income. S&P 500 companies returned about 63% of earnings to shareholders, the lowest since 2004, compared with a historical 80% to more than 100%.

Alphabet, Meta, Microsoft and Amazon are becoming asset-heavy companies. Their returns must now clear a much higher cost-of-capital hurdle to avoid market repricing. A decade ago, Damodaran said, several of them were “capital light, high growth businesses.”

Four Paths for Investors

Going forward, Damodaran sees four actionable paths, none of which lets investors sit out the debate.

• The Maximum Bull (All-In Disruption): Investors can pursue the thesis articulated by Leopold Aschenbrenner by going long AI architecture builders and suppliers while shorting vulnerable legacy industries. Investors must strictly avoid leverage, as excessive debt triggers forced liquidation long before secular cash flows materialize.
• The Passive Realist (Market Consensus): Maintaining broad equity exposure via holdings like the State Street SPDR S&P 500 ETF (SPY). Yet, passive indexing assumes aggregate corporate AI investments will generate returns that clear the cost of capital.
• The Defensive Skeptic (Capital Reallocation): Pivoting toward insulated niches like food processing and leisure – or allocating into regulatory-shielded regions like Europe or labor-competitive Asia. Damodaran advises preserving liquidity in short-term Treasuries yielding above 4% to withstand extended market exuberance.
• The Direct Contrarian (Crash Out / Reversal): Investors expecting imminent failure can liquidate equities into cash or even short overextended AI darlings while buying disrupted incumbents – the reverse of Aschenbrenner’s playbook. However, this approach demands strict discipline against market momentum.