Here's What I Think Is Going on With Broadcom Stock

Broadcom reported fiscal Q2 AI semiconductor revenue of $10.8 billion, up 143% year over year, and guided for roughly $16 billion in AI revenue in fiscal Q3, implying growth of more than 200%. Total revenue rose 48% to a record $22.2 billion and non-GAAP EPS increased 54%, but infrastructure software grew only 9% to $7.2 billion. Despite the strong results, the stock fell about 13% Thursday and another 8% Friday as investors reacted to a rich valuation and a high bar heading into earnings.

Analysis

The market reaction looks less like a fundamental downgrade and more like a positioning flush after a crowded AI winner became a consensus proxy for the entire infrastructure buildout. That matters because when a stock is carrying index-like ownership plus elevated implied perfection, the first miss is often not in the business but in the multiple: any sign of deceleration in the rate of acceleration can trigger de-grossing across adjacent AI beneficiaries, even if end demand remains intact.

The bigger second-order issue is customer concentration. A handful of very large programs means the revenue stream is not just cyclical, it is lumpy and binary at the project level; that creates a wider dispersion of outcomes than the headline growth rate suggests. If hyperscaler capex stays strong, the next leg likely accrues to networking, opticals, and foundry capacity before it fully benefits the “platform” names, because those are the bottlenecks that convert demand into shipped volume.

The underappreciated tell is the software segment’s slower growth relative to the chip arm: it signals that investors are no longer willing to pay a blended multiple for a business where the high-growth AI piece is doing all the narrative heavy lifting. In that setup, the stock can remain volatile for weeks even if the fundamental story is fine, because holders are forced to decide whether they own an AI pure play or a diversified compounder. The most likely catalyst for a reversal is not another beat, but evidence that the next custom silicon design win expands the customer set or shortens the revenue ramp, which would directly address the concentration discount.

Key Decisions for Investors

• Maintain a tactical short-term hedge against AVGO with a 1-3 month collar or put spread; the stock is now in a valuation-sensitive zone where good news may only stabilize, not re-rate, and downside can extend 10-15% if AI spend rotates elsewhere.
• Relative value: long NVDA vs short AVGO on a 6-12 week horizon. If AI capex stays robust, the more diversified demand base and broader ecosystem participation should reduce single-name execution risk versus AVGO's concentrated program exposure.
• Long the bottleneck suppliers rather than the headline winner: consider a basket of AI networking/optical exposure over AVGO for the next 2-4 quarters, since the first-order beneficiaries of continued hyperscaler spend are often the infrastructure choke points, not the most crowded AI equity.
• For new money, wait for either a 15-20% deeper pullback in AVGO or a clear upward revision to the FY27 AI revenue target before initiating a long. At current levels, risk/reward is asymmetric only for existing holders managing tax and portfolio constraints.
• If using options, express bullishness via call spreads rather than outright shares; the implied growth path is already expensive, so capped upside is preferable to paying full multiple exposure for incremental confirmation.