The S&P 500 sits within striking distance of an all-time high while trading at a forward price-to-earnings ratio of 19.4, its lowest level since the tariff-driven sell-off of April 2025. The index has climbed 13.1 percent year-to-date and more than doubled since the start of 2023, yet the valuation multiple has compressed because earnings estimates have risen faster than prices, a dynamic that usually signals skepticism, not euphoria.
The mega-cap distortion
Nvidia, Amazon and Alphabet, three of the five largest companies by market value, now trade at forward multiples of 24.4, 20.1 and 17.2 times respectively. Each is cheaper than the index was a year ago. Nvidia’s stock has advanced 21.9 percent in 2026 and 38.9 percent in 2025, but its fiscal 2028 revenue guidance implies a 70 percent year-over-year jump, a pace that makes the current multiple look like a rounding error. The Vera Rubin platform began shipping in August, giving the company visibility well into the next fiscal year.
Hyperscalers funding their own competition
Amazon and Alphabet have turned free-cash-flow negative as they funnel capital into AI infrastructure, much of it paid to Nvidia for compute capacity. Amazon has historically reinvested rather than repurchased shares; Alphabet has been a consistent buyback story. Both have reversed those patterns in the same quarter. The market is pricing them at a discount to their own history, and at a slight premium to the broad index, despite balance sheets and moats that are materially stronger than the median S&P 500 constituent.
What the compression is pricing
A forward P/E at the low end of the recent range while the index prints highs means the market is not extrapolating the current earnings trajectory. It is demanding proof that hyperscaler margins inflect once the capital cycle peaks. Microsoft’s multiple has held steady; Apple’s has expanded on a rally that has not been matched by a comparable upgrade to growth forecasts. The dispersion tells you where the doubt lives.