The S&P 500 Has Fallen in 56% of Septembers Since 1928 — What It Means for AI Stocks

September has a reputation on Wall Street, and it isn’t a flattering one. Since 1928, the S&P 500 has finished lower in roughly 56% of Septembers — making it the only month of the year that has historically been more likely to decline than to rise. For investors who have ridden the artificial intelligence trade higher, that statistic is worth pausing over as the calendar turns.

What the September statistic actually says

A coin flip that lands on tails 56% of the time is still very close to a coin flip. Nearly half of all Septembers since 1928 have been positive, and the ones that were negative include everything from mild single-digit dips to the panics that punctuated the Great Depression and the 2008 financial crisis. Averaged over almost a century, the month’s losing record is real but modest — and it says nothing definitive about what any single September will do.

There is also no widely accepted explanation for the pattern. Popular theories point to fund managers returning from summer vacation and rebalancing portfolios, mutual funds selling losers ahead of fiscal year-end, and investors raising cash after a typically quiet August. None of those explanations is provable, which is part of why seasonality is best treated as trivia rather than a trading signal.

Why AI stocks feel more exposed

If the broad market wobbles, high-multiple growth names usually wobble more. That is the nature of stocks whose valuations rest heavily on earnings expected years into the future: when sentiment sours or interest-rate assumptions shift, the discount applied to those distant profits rises, and prices fall faster than the market average.

AI-linked companies — chipmakers, data-center suppliers, cloud platforms and software firms selling AI features — have been among the market’s biggest winners of the past few years. That success cuts both ways. Crowded, richly valued trades tend to see sharper drawdowns during periods of indiscriminate selling, simply because so many investors are sitting on gains they may decide to protect.

AI stocks also carry a second layer of risk that has nothing to do with the calendar: the sheer size of the capital spending commitments underpinning the theme. Investors are watching closely for evidence that the enormous outlays on computing infrastructure translate into durable revenue and profit growth. Any quarter that raises doubts on that front is likely to hurt far more than a seasonal slump.

The practical takeaway

For long-term investors, the September statistic is not a reason to sell. Attempting to sidestep a month with a 56% chance of a decline means accepting a 44% chance of missing a gain, plus taxes and trading costs — a losing proposition over time. Historically, the market’s long-run advance has rewarded people who stayed invested through unremarkable pullbacks.

What the statistic can do is set expectations. If September brings volatility, it should not come as a surprise, and it should not be mistaken for a verdict on the AI investment thesis. Investors who want to be prepared can focus on the things they control: position sizes that won’t force a panic sale, diversification beyond a handful of correlated AI names, and a shopping list of companies they would be glad to own at lower prices.

Seasonal weakness, when it arrives, tends to be temporary. Business fundamentals are what ultimately determine returns — in September and every other month. Read More