Hock Tan Guided Broadcom Past $100 Billion of AI Revenue in 2027. The Stock Is 25% off Its High.

Broadcom CEO Hock Tan reiterated AI semiconductor revenue guidance of $100B+ by fiscal 2027 (from $56B expected this fiscal year). Despite strong AI momentum ($10.8B AI semiconductor revenue in fiscal Q2 2026, +143% YoY; $16.0B guided for Q3), the stock trades near $370, ~25% below the $495 52-week high, with the market valuing the same promise at lower earnings multiples (19x vs ~25x expected FY27 earnings at the peak). The key near-term catalysts are the Sept. 2 results—especially whether the $16B AI quarter and fiscal Q4 guidance support the back-half-loaded ramp toward the $100B target.

Analysis

The market is not debating whether AI demand exists; it is debating when the cash flow shows up. That matters because the stock was previously priced for a cleaner conversion path, and the current setup leaves little room for a back-half-loaded ramp to wobble. In the next 1-3 months, the important variable is not the long-dated revenue target but whether the company can keep the near-term guide strong enough to prevent another de-rating on “good story, slow proof.”

Competitively, the biggest winners are the hyperscale customers. Custom silicon shifts more of the AI value chain from merchant chips to vertically integrated economics, which improves unit cost control for GOOG/GOOGL and META even if it increases their capex intensity in the near term. The main loser is the supplier whose valuation is most tied to flawless execution on a concentrated customer base; if any one customer rebalances sourcing, the multiple can compress faster than the revenue base changes. Over 6-18 months, this also keeps pressure on NVDA’s scarcity premium in inference workloads if custom ASIC adoption keeps expanding.

The contrarian point is that the drawdown may still be mostly a valuation reset, not a thesis break. At a lower multiple, the stock is less fragile to minor timing slippage, but it is still vulnerable to any guide that suggests the ramp is slipping into next year. Falsifiers are simple: a September report that preserves the current AI cadence and shows no downgrade to fiscal 2027 visibility would argue for a relief re-rating; a guide below the current run-rate would likely take the stock to a new lower trading range rather than just a temporary dip.

Key Decisions for Investors

• Do not chase AVGO pre-earnings; wait for Sept. 2 confirmation. Buy only if AI revenue and forward guide hold while the stock remains at or below ~19x fiscal 2027 earnings; otherwise let it re-test lower before adding.
• For a relative-value expression of customer diversification risk, consider a small long MRVL / short AVGO pair into the print. Thesis: Google-style multi-sourcing can broaden custom-silicon winners, but AVGO still carries the sharper concentration premium. Use tight size because MRVL is more execution-sensitive.
• If you need event protection, use a limited-risk AVGO downside put spread for the September print rather than outright shorting. Best use is to hedge a long semiconductor basket against a guide disappointment that would compress the entire custom-ASIC group.
• Prefer GOOG/GOOGL and META on any AI-supply-chain weakness over AVGO. The market often underappreciates how custom silicon improves their long-run inference economics even as it temporarily raises capex.