Broadcom collapsed 19.5% between June 3–5, 2026, despite raising its Q3 FY26 AI semiconductor guidance.1 NVIDIA dropped 6% and Meta fell 6% on June 5 alone. The leveraged AI ETF FNGU lost 22% across the five-day window.1
The selloff coincided with the US 2-year Treasury yield spiking to its highest level in 16 months.1 Rate-sensitive growth equities bear the heaviest valuation impact from rising short-term yields. Discounted cash flow models compress sharply when the risk-free rate rises, punishing stocks priced on future earnings — precisely the profile of AI infrastructure plays.
Critically, AI demand signals did not deteriorate. Dell reported Q1 AI server order bookings of $24 billion and raised its full-year guidance to approximately $60 billion.1 Broadcom's own guidance increase confirms the AI capital expenditure cycle remains active. Earnings revisions stayed positive while stock prices fell — a divergence that points to rate sensitivity, not fundamental weakness, as the dominant driver.1
For portfolio managers, the episode sharpens a key risk: AI-exposed equities are now dual-beta instruments. They carry both AI adoption risk and duration risk. As long as the Fed's terminal rate path remains uncertain, high-multiple tech names will exhibit bond-like sensitivity to yield moves — regardless of underlying business momentum.
Investors overweight in AI semiconductors and infrastructure stocks face a reallocation question. Short-duration value names and dividend-paying industrials that benefit from AI buildout — without the valuation premium — may offer a more rate-resilient expression of the same thesis. The AI trade is intact; the risk premium attached to it has repriced.
The divergence between fundamentals and price action creates both a risk and an opportunity. If yields stabilize, the selloff gap between strong earnings and depressed prices may close quickly. If rate pressure continues, the multiple compression has further to run even as AI revenues grow.
