When consensus prices in perfection, an in-line result reads as a disappointment. AI infrastructure spend expectations are embedded so deeply in current valuations that only materially above-consensus guidance prevents institutional selling. Meeting expectations is no longer enough.
The Broadcom episode illustrates the mechanism. A guidance release that would once have confirmed AI demand momentum instead triggered a near-20% two-day collapse.1 The market has recalibrated: the burden of proof for AI chip makers is no longer growth, it's acceleration.
The setup heading into Q2 earnings season is unchanged. NVIDIA, AMD, and Marvell each face the same dynamic.1 Options implied volatility on semiconductor names is expected to expand as earnings approach, reflecting institutional hedging against guidance disappointment.1
For growth-dependent portfolios, the risk structure is specific. The reward for beating consensus is incremental. The penalty for meeting it is severe. That asymmetry compresses the risk/reward ratio for holding AI semiconductor names through earnings events.
Investment thesis risk runs deeper than short-term price action. Portfolios built on AI infrastructure spend as a multi-year secular growth theme face compressing multiples if guidance momentum decelerates. The valuation premium embedded in these names assumes compounding beats, not steady growth.
Institutional investors with concentrated AI semiconductor exposure face a narrowing window before Q2 reporting. Holding through earnings requires either conviction that guidance will materially exceed consensus or tolerance for drawdowns matching Broadcom's June magnitude.
The early June moves in Broadcom and NVIDIA are not isolated. They are a data point describing the current risk structure for the sector. Until valuations reprice or guidance expectations reset, the asymmetry remains.
