Broadcom Is Down 20% From Its All-Time High. Here's Why It's a Screaming Bargain.

Broadcom (AVGO) is positioning 2027 as its growth inflection point, driven by custom AI chips expected to generate >$100B in 2027 revenue vs $10.8B in Q2. The article argues the market has not priced in the ramp in orders from major clients beyond Alphabet, and values Broadcom at ~20x 2027 earnings (or 30x in a base case). If AVGO can meet Wall Street estimates and trade near 30x fiscal 2027 earnings, the article suggests ~50% upside over the next 12–18 months.

Analysis

The setup is less about near-term earnings power and more about whether the market will extend duration on a 2027 story. If investors start treating AVGO as a multi-year AI infrastructure compounder, the stock can re-rate even before revenue arrives; if not, the name stays hostage to quarter-to-quarter capex scrutiny and multiple compression. The key question is not demand existence but order conversion: the first proof point of broad customer pull-through matters more than management commentary.

Second-order, custom silicon is a margin optimization tool for hyperscalers, not just an AVGO revenue stream. That should improve the economics of AI deployment for GOOGL and META, which can support more inference at lower unit cost, but it also raises the risk that AVGO’s growth is concentrated in a handful of very large programs with lumpy timing. NVDA is the obvious competitive pressure point only where workloads are stable and repetitive; where model iteration remains high, GPU share should hold better than the market fears.

The contrarian view is that consensus may be underestimating how much of the 2027 upside is already embedded in an 18-month-forward valuation lens. AVGO can absolutely work, but the stock is now sensitive to any slip in customer qualification timelines, gross margin mix, or AI capex cadence in 2026. The thesis is falsified if the company cannot show a clear step-up in backlog or if hyperscaler AI budgets flatten into next year, because then the market will de-rate the name before the 2027 revenue arrives.

Key Decisions for Investors

• Buy AVGO on weakness only, using a staged entry over the next 2-6 weeks; target 12-18 month upside, but require proof of order conversion in upcoming quarters. Falsify the thesis if AI revenue guidance or backlog fails to inflect by the next earnings cycle.
• Pair trade: long AVGO / short NVDA at a modest size as a relative-value expression on custom-ASIC share gain versus broad GPU spending. Best when hyperscalers emphasize inference efficiency over frontier-model training; exit if NVDA re-accelerates on stronger networking or platform demand.
• For a cleaner catalyst trade, use AVGO call spreads rather than outright stock to monetize the 2027 duration without overpaying for headline volatility. This is attractive only if implied vol stays elevated into earnings and the market keeps debating timing rather than demand.
• Watch GOOGL and META as second-order beneficiaries of lower compute cost; if either company raises AI capex while maintaining operating margin, that validates the custom-chip economics and supports AVGO’s multiple. If those firms cut spend or delay deployments, it is a warning that 2027 expectations are too aggressive.