Semiconductor stocks hit record highs in mid-June 2026. JPMorgan is now asking a pointed question: what happens when everyone tries to exit the same trade at once?

In a note dated June 18, the bank’s strategists warned that volatility spikes in chip stocks like Nvidia and AMD could breach portfolio Value-at-Risk limits at institutional funds. When that happens, the selling isn’t optional. It’s mechanical, automatic, and indifferent to whether the underlying business is thriving.

The crowding problem

JPMorgan has been sounding this alarm for nearly a year now. Back in July 2025, the firm published a note describing AI-linked momentum stocks, including Nvidia (NVDA), Palantir (PLTR), and Coinbase (COIN), as reaching the most extreme crowding levels in 30 years.

What’s notable is that JPMorgan’s concern isn’t really about the AI chip business itself. The firm’s recent reports don’t point to falling demand for AI chips or deteriorating fundamentals. The risk they’re flagging is structural: it’s about how markets behave when volatility rises in stocks that institutions have crowded into with historically unusual intensity.

VaR, or Value-at-Risk, is a risk management tool that banks and hedge funds use to measure how much a portfolio could lose in a given time period. When a stock’s volatility spikes, the VaR calculation can breach internal limits, forcing portfolio managers to reduce exposure. This creates a feedback loop. Selling begets more volatility, which triggers more VaR breaches at other firms, which triggers more selling. JPMorgan’s strategists have described this dynamic as a “VaR shock,” and they see the current setup in semiconductor stocks as particularly vulnerable to one.

Record highs meet record fragility

Chip stocks rebounded to all-time highs in mid-June 2026, buoyed by continued enthusiasm around AI infrastructure spending. JPMorgan itself has upgraded its S&P 500 projections, partly because of anticipated AI-driven earnings growth across multiple sectors.

The bank’s analysts have pointed to inflated valuations and the likelihood of rapid reversals in speculative segments of the market. Sharp corrections could leave investors exposed, particularly those running leveraged strategies or momentum-chasing algorithms that amplify moves in both directions.

What this means for investors

Traders heavily concentrated in momentum names like Nvidia, AMD, Palantir, and Coinbase face the highest risk of forced liquidation cascades. These are the stocks where 30-year extremes in crowding have been documented. When institutional risk models start firing, these names would likely see the most aggressive selling pressure, regardless of their earnings trajectories.

JPMorgan’s own position captures the tension perfectly: the firm is bullish enough on AI earnings to raise its S&P 500 target, yet cautious enough to warn that the stocks most directly tied to that thesis are sitting on a volatility powder keg.