Broadcom: The Market Overreacted To Great Earnings, And I Started A Position

Broadcom posted record Q2 results, with revenue up 48% year over year and AI semiconductor revenue up 143%, while shares still fell more than 20% after earnings. Management guided Q3 AI semiconductor revenue to $16B and reiterated a $100B+ AI revenue target for FY27, backed by multi-year AI silicon and networking deals with Google, Meta, Anthropic, and OpenAI that extend visibility through 2028. Despite the selloff, the fundamental setup remains strong and the stock now trades at 15x 2028 earnings.

Analysis

The market is treating the print like a peak-growth event, but the more important signal is that AVGO has effectively converted AI capex from a spot market into a contracted annuity through 2028. That shifts the competitive battleground from demand uncertainty to execution and supply allocation, which should favor the few vendors able to ship custom silicon plus networking as a bundle; commodity AI infrastructure names and slower ASIC rivals are likely to see multiple compression as investors reprice durability.

The post-earnings selloff looks more like positioning and valuation reset than a thesis break. A 15x 2028 multiple on a business with visible multiyear growth is not expensive if management can keep converting design wins into revenue at anything close to plan; the bigger risk is that customers delay deployment or stretch qualification cycles, which would matter most over the next 2-3 quarters, not years. In other words, the stock is vulnerable to near-term digestion, but the business remains insulated unless there is evidence that hyperscaler AI capex is pausing or migrating to in-house alternatives faster than expected.

Second-order effects matter: every incremental dollar of custom AI silicon spend at GOOGL and META is likely displacing merchant GPU exposure, not reducing total AI spend. That is a relative negative for firms whose economics depend on generalized accelerator demand and a relative positive for networking, optical, and power-delivery suppliers tied to cluster buildouts. The market may be underestimating how sticky these custom programs become once software stacks, validation, and supply chain integration create switching costs.

The contrarian setup is that the decline may be overdone if investors are anchoring on headline AI growth deceleration instead of backlog visibility. If AVGO can hold the AI revenue trajectory through the next two quarters, the stock should regain support as the market moves from "is AI real?" to "how much of the next wave is already contracted?" The key tells will be order cadence, packaging capacity, and whether customer concentration turns from a risk into a moat.

Key Decisions for Investors

• Buy AVGO on weakness over the next 1-2 weeks; prefer staged entry rather than chasing a rebound. Risk/reward favors a 12-18 month hold if management continues to show contracted AI revenue visibility; invalidate if AI order cadence slows materially in the next two earnings cycles.
• Pair trade: long AVGO / short a basket of merchant AI compute exposure (or semis most dependent on open-market accelerator demand) for 3-6 months. The thesis is that contracted custom silicon captures share from more fungible AI hardware economics.
• Use a call spread in AVGO rather than outright shares if volatility remains elevated: 6-12 month bull call spread to express mean reversion with defined downside, targeting a recovery as the market digests guidance and multiple resets.
• For GOOGL and META, treat the announcement as a strategic positive but not an immediate earnings catalyst. Add only on pullbacks over 1-2 quarters; the upside is from better AI capex efficiency and faster infrastructure deployment, not from near-term P&L lift.
• Watch for a catalyst-driven re-rate if AVGO confirms no slippage in Q3-Q4 AI revenue conversion. If that happens, add to long AVGO and trim any short exposure in adjacent AI infrastructure names, as the market may have over-discounted execution risk.