Broadcom's 5.9% one-day slide is a reset in what the market demands from AI, not a verdict that AI demand has vanished. Our take is bullish: the company is still converting custom accelerators and networking demand into accelerating revenue, while the tape is reacting to a missed revenue expectation and a less aggressive 2027 setup. That distinction matters because expectation compression can punish a high-multiple stock without invalidating its operating engine. We’re buying the dislocation, with Sept. 3 earnings as the immediate test.
The clearest operating fact is Broadcom’s Q2 AI semiconductor revenue: $10.8 billion, up 143% year over year. Management attributed that surge to increasing demand for custom AI accelerators and AI networking, then guided to $16 billion of AI semiconductor revenue for Q3. That is not a demand air pocket; it is a business still accelerating after already reaching a large scale. The market can debate how much of that growth is priced in, but it cannot honestly describe the latest reported trend as broken.
Visibility reinforces the point. Management said it was very comfortable that Broadcom had secured supply for 2026 and 2027, while the broader order picture extends into 2028. That kind of planning horizon matters in custom silicon, where design wins, capacity, and deployment schedules are not rebuilt overnight. It does not guarantee every forecast will be met, and it says nothing about the eventual margin profile of each program. It does show why a single quarter’s expectation miss should not be mistaken for a collapse in customer commitment.
The tape looks more like multiple compression than demand destruction. Broadcom fell 5.9% even though its latest quarter still produced a strong AI print; the reported trigger was a revenue miss and management leaving its previous 2027 sales forecast unchanged. That is a harsh reaction, but it is exactly how a stock priced for perfection behaves when the future gets marked from extraordinary to merely excellent. AVGO has gained 13.1% year to date while technology has gained 31.7%, an 18.6-point lag that says this is already a relative de-rating, not an untouched momentum trade.
Fundamental quality gives the contrarian trade a foundation. The TickerSpark Score is 86, with its Profitability component at 100 and Growth component at 95. Analyst consensus is still Buy, with 51 buys, 8 holds and no sells. Those signals do not erase the premium, but they reinforce that the market is repricing expectations around a strong business rather than abandoning a weak one.
The bear case has teeth because AVGO is not cheap. A 63.49 trailing P/E and 24.78 price-to-sales multiple leave little room for a merely good AI outcome, let alone a stumble. The unchanged 2027 sales forecast gave that concern a concrete hook: the ramp may still be powerful, but perhaps not compounding as explosively as expectations demanded. Google’s supplier diversification is the sharper risk. Recent analyst commentary cited a possible decline in Broadcom’s share of Google TPU-related revenue from about 95% in 2026 to 80% in 2027 and 65% in 2028, alongside more in-house development and a shift toward another supplier. If that trajectory plays out, Broadcom’s headline AI growth could remain strong while the mix and concentration profile worsen.
Risk is not limited to Google. Nine recent insider transactions were sells totaling $82.30 million, with no buys, while the wider chip group has already suffered a sharp AI-trade de-rating. VMware-related EU antitrust document demands and a U.S. trade probe involving memory chips add sentiment overhangs even though neither is a direct read-through on custom-AI orders. These facts justify smaller sizing and a hard earnings checkpoint. They do not yet outweigh the central evidence: demand is still being described as effectively insatiable, supply is secured across the next two years, and the company is guiding AI revenue higher for the next quarter.
That leaves us buying, but not treating the selloff as a free pass. Exposure belongs in tranches because a 63.49 P/E can turn any guidance wobble into another violent repricing, and the Google concentration risk is real. The next earnings report, expected around Sept. 3, is the decision point: the bullish thesis gets stronger if the $16 billion Q3 AI target is defended and management maintains its supply and visibility language. A reduced AI outlook, a clear loss of key custom-accelerator programs, or evidence that networking demand is slowing would change our mind.
Price levels can keep the discipline concrete. AVGO closed at $392.99, just above its 50-day moving average at $390.46 and comfortably above its 200-day average at $369.20. We would respect a sustained break below the 200-day line as evidence that the market is no longer treating this as a routine reset, especially if it arrives with weaker AI guidance. Until operating data says otherwise, the 5.9% drop is the opportunity: expectations were cut faster than the demand evidence.
