Broadcom Shares Tumble Despite Surging AI Revenue. Should Investors Buy the Dip?

Broadcom reported Q2 revenue of $22.19 billion, up 48% year over year, with adjusted EPS of $2.44 versus $2.40 consensus and AI semiconductor revenue surging 143% to $10.8 billion. The stock fell because management left its fiscal 2027 AI chip sales target unchanged, even though it still expects AI semiconductor revenue to reach $56 billion this fiscal year and more than $100 billion in fiscal 2027. Fiscal Q3 guidance calls for revenue of $29.4 billion versus $28.5 billion expected, supported by a projected 200% increase in AI semiconductor revenue to $16 billion.

Analysis

The market is reading this as a guidance miss, but the more important signal is that the AI revenue trajectory is still being constrained by execution capacity rather than demand. That matters because in this phase of the cycle, the scarce asset is not customer intent but manufacturing, packaging, and network integration; Broadcom appears to have the more defensible bottleneck position versus companies selling commoditized compute. The softer software growth is a near-term multiple drag, but it also creates a cleaner segmentation: AI hardware is now the valuation anchor, while software becomes the swing factor for whether the stock re-rates or just compounds.

Second-order, Broadcom’s customer concentration is no longer just a risk; it is a financing-like advantage. If a handful of hyperscalers and model builders are effectively pre-committing demand several years out, the visibility profile starts to resemble a contracted infrastructure platform, which should compress the earnings discount rate over time. The counterpoint is that any slowdown in capex from one or two key accounts would hit the stock harder than the headline growth numbers imply, so the next 1-2 quarters matter more for sentiment than the next 2 years do for fundamentals.

The market’s reaction looks more like disappointment that upside was not revised, not evidence that the growth curve is breaking. That sets up a classic post-earnings reset: the stock can de-rate further if investors decide the AI guide is already “fully known,” but the path back higher is likely if management simply continues to validate supply readiness and margin stability. The bigger contrarian read is that the stock may be cheap not because the business is misunderstood, but because the market is underestimating how long these customers will keep spending once model training and inference infrastructure become embedded operating necessities.

Key Decisions for Investors

• Buy AVGO on a 2-6 week post-earnings drift, targeting a move back toward the pre-print multiple if AI revenue visibility is reaffirmed; use a tight stop if management commentary turns to customer digestion rather than supply constraints.
• Pair trade: long AVGO / short a basket of higher-beta AI enablers with weaker pricing power over the next 1-3 months; Broadcom’s contracted demand and margin support should outperform names whose AI revenue is still more narrative than realized cash flow.
• Add GOOGL or META on AVGO weakness as a secondary AI-capex hedge; both remain core demand nodes, and any investor overreaction in AVGO is likely to be a read-through to hyperscaler spend that is too aggressive in the short term.
• Avoid chasing NVDA on this print; Broadcom’s guide implies AI infrastructure spending remains broad, but the cleaner trade is the company with visible conversion of demand into cash flow rather than the name already priced for perpetual acceleration.
• For options, consider AVGO call spreads 3-6 months out to express a rebound in sentiment without paying full upside for a stock that may remain range-bound until the next guide refresh.