Up Nearly 40% Since April Began, Is Broadcom Stock Still a Buy?

Broadcom's custom AI chips are seeing strong hyperscaler adoption, and CEO Hock Tan said the business could generate more than $100 billion in revenue by 2027. Wall Street estimates call for revenue to rise from $63.9 billion in fiscal 2025 to $158 billion in fiscal 2027, though the stock already trades at 38x 2026 earnings and 24x 2027 earnings. The article is bullish on Broadcom's AI growth outlook, but notes valuation is no longer cheap after a nearly 40% rally since early April.

Analysis

The market is starting to re-rate AVGO not as a cyclical semis supplier but as a toll road on hyperscaler capex, and that changes the earnings durability debate. If custom silicon keeps taking share from merchant GPUs in training/inference, the mix shift should widen Broadcom’s margins and reduce the volatility investors normally assign to hardware names. The bigger second-order winner is GOOGL: external TPU commercialization turns an internal cost-optimization project into a monetizable platform, which could improve cloud economics and help defend share against the fastest-growing AI-native workloads.

The key framing mistake is to anchor on the next quarter instead of 2027 realization. The market is discounting a multi-year revenue step-up, but the path is lumpy: design wins today do not become revenue until tape-out, validation, and deployment cycles roll through, so headline upside can get pulled forward while actual cash conversion lags by several quarters. That creates a setup where AVGO can keep grinding higher on guidance revisions, but the first real disappointment would likely come from timing slippage rather than outright loss of demand.

NVDA is the quiet competitive loser at the margin, not because it is being displaced in absolute terms, but because every dollar of hyperscaler spend diverted to custom silicon reduces the top-end growth rate multiple investors are willing to pay. INTC is even more exposed indirectly: if the market decides custom accelerators are the preferred path for scale AI workloads, it reinforces the view that the value pool is in design/IP integration rather than general-purpose x86. The contrarian read is that the move may already reflect a lot of the 2027 upside, while consensus is underestimating how quickly customer concentration and chip-program execution risk can cap re-rating potential if one large design win slips.

Key Decisions for Investors

• Stay long AVGO, but favor call spreads over outright stock for the next 3-6 months; the stock has likely pulled forward part of the 2027 story, so upside remains but convexity is now the better expression.
• Add GOOGL on any post-earnings or AI-spend-related weakness over the next 1-2 months; TPU monetization is an underappreciated margin lever and gives cloud a differentiated cost curve versus GPU-heavy peers.
• Pair trade: long AVGO / short NVDA into the next AI capex read-through if you want to express the custom-silicon share gain; risk is a rebound in merchant GPU pricing or a surprise acceleration in NVDA enterprise demand.
• Avoid chasing INTC on this headline; if anything, use strength to fade, since the article reinforces a market structure where specialized accelerator IP captures value faster than legacy CPU-centric franchises.
• Set a catalyst watch for hyperscaler capex commentary over the next 1-2 quarters; any language implying budget discipline or deployment delays is the first signal that AVGO’s 2027 revenue bridge needs to be cut.