Short positions hit record highs—can the U.S. stock market bull run hold up?

TL;DR

• S3 data shows that the short interest ratio in U.S. stocks has risen to a record level, while reductions in technology stock holdings are also increasing.
• This is more akin to hedging at high levels rather than signaling a bull market reversal; the direction of volatility will be determined by earnings results.
• Related tickers: SPX, NDX, XLK, SMH, NVDA, Mag 7.

Even as the S&P 500 trades near highs, the short interest ratio in U.S. stocks and hedge fund reductions in technology holdings are both heating up, leading the market to reassess the margin of safety in AI-related trades.

This set of signals makes investors nervous because it runs counter to the prevailing trend of the past two years. The AI narrative has been driving index gains, yet institutions are buying more insurance on this trade. The issue isn't just whether the U.S. stock market has peaked, but whether bad news will be amplified when prices already reflect a high degree of optimistic expectations.

Let's clarify the concepts first. The short interest ratio represents the size of borrowed shares sold short relative to total tradable shares. It can represent a direct bet on a decline, or it can simply be a hedge. Funds may still hold long positions but use short selling, options, or position reductions to mitigate downside risk.

Therefore, a record short interest does not automatically mean institutions are universally bearish. A more accurate description is that while the U.S. stock market is still trading on the long-term benefits of AI, institutions have begun to reprice short-term volatility.

Short Interest Can Increase Even as Indices Rise

Rising prices and increasing short interest may seem contradictory, but they often occur simultaneously. Especially when valuations are stretched, positions are crowded, and earnings season approaches, funds will maintain long positions while increasing their protective hedges.

According to data from S3 Partners cited by media, the short interest in S&P 500 components was roughly 3.79% of free-float shares, a new high since S3 began tracking in 2010. For Russell 3000 components, this ratio was around 6.3%, also at a record high.

These figures cannot be simply aggregated with other metrics. Different institutions have varying statistical scopes, and exchange disclosures are often about the total number of shorted shares rather than the same proportional indicators. Data from the New York Stock Exchange for July, for example, showed that total short interest on NYSE Group increased compared to the prior period as of June 30, 2026, only indicating an increase in the scale of short positions.

Client reports from Goldman Sachs Prime Brokerage also point in a similar direction. According to a Reuters report from July 6, U.S. hedge funds net sold technology hardware and semiconductors for the fourth consecutive week, with the Information Technology sector also being the most net sold U.S. sector for the fourth straight week.

The key takeaway here is not a precise percentage, but that large players are reducing their net exposure to tech. The market hasn't made a collective retreat; overall prices are still supported by retail buying, corporate buybacks, and trend-following capital. However, the protective cushion during the rally has thickened.

AI Trade Enters a Phase of Bad News Amplification

Institutions aren't worried that AI has no value, but that current prices have already pre-emptively discounted a significant amount of future returns.

Over the past few years, the core explanation for the U.S. stock market rally has been AI. Cloud providers and tech giants have increased capital expenditures, Nvidia and the semiconductor supply chain have benefited from the spillover of orders, and the market has believed these investments would eventually translate into revenue, profit margins, and productivity gains.

For investors, capital expenditure is not just a story; it's an investment that needs to generate returns. If AI infrastructure investment continues to rise, but terminal revenue, enterprise spending, and profit contributions do not accelerate in tandem, valuations will come under pressure first.

Semiconductors are the most likely to become volatility amplifiers. They sit at the front end of the AI investment chain, reacting most quickly to orders and expectations, and their valuations are most sensitive. If earnings reports show slowing order growth, margin pressure, or concerns about customer concentration, the market tends to compress semiconductor valuations first, which then spills over to the Nasdaq and S&P 500.

Geopolitical risks serve as an external catalyst. Conflicts in the Middle East, energy prices, and supply chain uncertainties may not change the long-term demand for AI, but they do change the multiples the market is willing to pay for high-valuation assets. What a high market fears most is not a single piece of bad news, but overcrowded positions when that news arrives.

This explains why the increase in short interest is more like a rise in insurance premiums. Institutions may not necessarily believe the AI bubble will burst, but they are unwilling to maintain significant unhedged exposure ahead of earnings and external risks.

Morgan Stanley's Dual-Track Assessment

Morgan Stanley's recent strategic framework perfectly illustrates the core of this divergence. The index could still have upside medium-term, but tech and semiconductors need to digest their gains in the short term.

On July 14, Morgan Stanley strategist Mike Wilson mentioned on the firm's official podcast that semiconductors could pull back, and there would still be volatility and corrections before the next leg of the bull market. On July 15, a Morgan Stanley Wealth Management article noted that its Global Investment Committee expects the S&P 500 could rise to between 8000 and 8300 points over the next year, while also recommending taking some profits in the semiconductor sector.

This isn't simply bullish or bearish; it's a common dual-track assessment for a market at highs. Long-term, if earnings continue to be revised upward and AI investments generate real revenue, the index can move higher. Short-term, if valuation expansion runs ahead of earnings, a pullback can occur.

For investors, it's important not to interpret positioning signals as one-way predictions. An increase in short positions could, after positive earnings, turn into a short squeeze, forcing short sellers and hedgers to cover and pushing prices higher. Conversely, it could amplify a decline if bad news emerges.

What determines the direction is not the level of short interest, but whether the market finds its prior valuation assumptions too conservative or too optimistic when the catalyst lands.

Earnings Will Determine if Shorts Are Fuel or Pressure

The upcoming earnings reports from tech giants and semiconductor companies will serve as a stress test for the AI trade. The market isn't looking for a simple statement of "strong demand"; it needs to see that cloud revenue, AI orders, gross margins, and returns on capital expenditure all align.

If earnings show accelerating cloud revenue, still-strong AI orders, and stable gross margins, the short positions could turn into fuel for a rally. Short sellers or hedgers would need to cover, and momentum-driven capital would reconfirm the AI theme.

If earnings only prove that capital expenditure continues to expand, but fail to demonstrate a concurrent payoff in returns, the market will re-evaluate paying high valuations for future growth. In that case, short positions aren't the cause of the decline but will act as a volatility amplifier.

The currently more reasonable assessment is neither that the bull market is over nor that shorts must be squeezed. The U.S. stock market is entering a more discerning phase. The AI narrative remains valid, but valuations require earnings to deliver. Semiconductors remain a core theme, yet they are also the first area to bear the brunt of risk repricing.