Top Semiconductor Stock Outpacing Nvidia With a 36% Gain Over 6 Months

Broadcom is positioned to keep outperforming Nvidia as AI ASIC demand accelerates: AI semiconductor revenue grew 143% YoY and is expected to rise more than 200% in fiscal 2026 Q3, targeting $16B AI semiconductor revenue and $29B total revenue (implying 84% YoY total revenue growth). In fiscal 2026 Q2, Broadcom delivered 48% YoY revenue growth and net income nearly doubled, with AI driving nearly half of total revenue. The article argues ASIC specialization and energy-efficient inference workloads should help Broadcom gain further market share versus Nvidia.

Analysis

The real economic story is not that one chip is “better,” but that hyperscalers are trying to internalize more of the AI margin stack. That shifts bargaining power away from merchant GPUs toward custom silicon, but it also increases customer concentration risk for AVGO: a handful of capex plans now matter more than a broad installed base. If inference dominates the next leg of AI spending, ASIC share can expand faster than GPU share; if training remains the spend center, NVDA keeps the higher-quality growth profile.

The market may be underpricing how cyclical this custom-silicon buildout can be. AVGO’s upside is leveraged to each customer’s deployment cadence, so a single design win looks bigger in the numbers than it is in durable unit demand; any delay in hyperscaler rollout can create a sharp air pocket in expectations. Over 1-3 months, the key catalyst is next guidance from GOOGL/META/MSFT/AMZN capex commentary; over 6-18 months, the question is whether inference efficiency compresses total silicon spend per workload, which could cap the whole AI supply chain.

Contrarian view: the consensus is treating ASIC penetration as a clean structural win, but it may also be a sign that AI economics are normalizing from scarcity pricing to optimization. That is bearish for NVDA’s mix if custom chips truly displace general-purpose acceleration, but it can also be bearish for AVGO’s multiple if the market starts valuing AI semis more like a high-end component business than a monopoly-like platform. The thesis is falsified if NVDA’s next two quarters show sustained re-acceleration in data-center demand or if AVGO’s AI revenue guide slows below the implied >200% growth path.

Key Decisions for Investors

• Initiate a 1-3 month long AVGO / short NVDA pair trade on strength: thesis is mix shift toward inference ASICs. Target 10-15% spread capture; stop if NVDA’s next earnings/revision cycle shows renewed acceleration in data-center demand or Blackwell pull-through.
• Use AVGO pullbacks to add exposure only if hyperscaler capex remains intact; otherwise treat it as a watch item, not a blind long. The key risk is customer timing, not technology failure.
• Buy downside protection on NVDA into the next earnings window with a 6-10 week tenor if implied vol is reasonable; this is the cleaner way to express the risk that custom silicon continues to take share from merchant GPUs.
• Pair AVGO with a basket of hyperscaler capex proxies (GOOGL, META, MSFT, AMZN) as a second-order trade: long AVGO, short equal-weight capex basket only if their spend plans stay firm but they signal more internal chip substitution. Falsify if any of the four cuts AI capex or delays deployment.