J.P. Morgan Global Research remains positive on global equities for 2026, forecasting double-digit gains across both developed and emerging markets. This comes after a violent mid-year sell-off in technology and semiconductor stocks exposed how fragile leveraged hedge fund positioning in the AI trade had become.

The sell-off's mechanics came into focus this month. J.P. Morgan Asset Management's "Review of Markets over July 2026," published August 3, found that forced deleveraging by some hedge funds amplified the market decline. The MSCI World Semiconductors Index fell 13.2% during the period. Deleveraging here means funds had to rapidly sell assets to reduce their borrowed-money exposure, often because losses on those borrowed positions hit a threshold where lenders demanded repayment.

Reuters, citing a separate JPMorgan note, reported August 4 that global hedge funds gave up almost 3% of their gains during July due to the unwinding of technology-related trades. They remained up roughly 8% year-to-date.

This was not the first tremor of the year. J.P. Morgan Asset Management had already examined an abrupt sell-off in software stocks in early 2026, publishing its analysis on February 18, which explored AI disruption dynamics and what it meant for sector allocation in private markets. By late February, according to a JPMorgan note reported by Reuters on February 24, hedge funds were buying back the biggest technology stocks as well as those considered vulnerable to AI advances, after weeks of selling. The whipsaw was well established before July's more severe dislocation.

J.P. Morgan's "Mid-Year Outlook 2026: Promise and Pressure" noted that major stock markets experienced roughly 10% corrections during this period and explicitly flagged risks around hedge funds' use of leverage and speculative practices. The sequence, from the February software sell-off through the July semiconductor plunge, traces a clear arc: concentrated AI-related positioning, amplified by leverage, created a feedback loop where forced deleveraging turned ordinary risk reduction into something far more violent.

Against that backdrop, J.P. Morgan's more constructive voices are worth separating from the volatility narrative. The Private Bank's Q2 2026 investment review, published July 8, argued that a strong earnings outlook led by tech and AI-driven growth continued to underpin equities, while non-tech sectors could catch up on a U.S.-Iran peace-related dynamic. A week later, J.P. Morgan published a market takeaway titled "Is It All One Big AI Trade?" dated July 24, arguing that AI's reach extends beyond technology with diverse sectors showing strength.

The broader context here matters. That latter report reframes the AI narrative from a single-sector concentration risk into a broad-based productivity story, which, if sustained, would dilute the very positioning concentration that made July's deleveraging so damaging.

The tension between these threads is the central question for the remainder of 2026. J.P. Morgan Global Research's forecast of double-digit gains across both developed and emerging markets presumes that the earnings engine holds and that the AI thesis broadens rather than narrows. The July episode, however, showed that the path depends heavily on how leverage is distributed across the system. When hedge funds are crowded into the same thematic trades and forced to unwind simultaneously, index-level damage (the 13.2% semiconductors drop) is a function of position concentration and forced selling, not fundamentals changing overnight.

The year-to-date hedge fund return of roughly 8%, even after July's 3% give-back, suggests the AI trade has generated substantial returns for those positioned correctly. Alpha refers to returns above what the broader market delivers. But the round-trip (selling tech in February, buying it back, selling again in July) points to the cost of conviction in a crowded trade. Each reversal exacted slippage, the gap between the price you expect and the price you actually get, and for leveraged players, potentially margin-related compulsion to exit at unfavorable prices.

For institutional allocators, the software sell-off analysis from February and the July semiconductor plunge raise parallel questions about private market exposure. J.P. Morgan Asset Management's decision to examine what the software sell-off means for private markets signals that mark-to-market discipline in public equities can feed into private valuations with a lag. This is especially relevant for late-stage venture and growth equity portfolios with AI-adjacent exposure. Mark-to-market means valuing an asset at its current market price rather than what you paid for it.

The concern is straightforward. If public market multiples in AI-related sectors remain volatile, private market write-downs may follow. The illiquidity that protects against forced selling in the near term can become a valuation trap in the longer term, because private holdings are hard to sell quickly, so their paper values may not adjust downward until well after public markets have already moved.

The U.S.-Iran peace dynamic flagged in the Private Bank's Q2 review introduces a separate macro variable. If geopolitical de-escalation holds, the rotation from tech into non-tech sectors that J.P. Morgan anticipates would alleviate some of the concentration risk that made July so punishing. Broad-based earnings participation across sectors would justify the double-digit equity forecast without requiring the AI trade to carry the entire market on its own.

What remains unresolved is whether the July deleveraging was a one-time clearing event or the first of several. J.P. Morgan's mid-year flag on hedge fund leverage and speculative practices suggests the firm sees the structural risk as ongoing. The roughly 10% market corrections and 13.2% semiconductor index decline were not, in this framing, the cost of a bad fundamental call on AI. They were the cost of too much leverage chasing the same thesis, which is a different problem entirely, and one that does not necessarily resolve because the underlying thesis turns out to be correct.