Sell in May and Go Away? 3 Stocks Where This Old Idea Doesn't Apply

The article is bullish on Nvidia, Broadcom, and Nebius, arguing that AI-driven growth is not fully priced into their stocks. Nvidia is cited at 24x forward earnings with Q1 revenue expected to rise about 79% year over year, while Broadcom’s custom AI chip business is projected to exceed $100 billion in annual revenue by the end of 2027 from $8.4 billion in Q1. Nebius is highlighted for expected growth of 523% this year and 206% next year, implying substantial upside if execution matches estimates.

Analysis

This is less a three-stock promotion than a read-through on second-order capex acceleration. The market is still pricing AI infrastructure as if hyperscalers are choosing between a few vendors, when the emerging reality is a broader stack expansion: GPU demand pulls through networking, power, packaging, memory, and purpose-built accelerators. That makes AVGO the cleanest lever on the migration from general-purpose GPUs to custom silicon, while NVDA remains the toll collector on the base layer; the real competitive damage is likely to hit smaller non-AI enterprise hardware names and legacy server vendors that are missing the next procurement cycle.

The more interesting setup is NBIS, where the upside is not just growth but scarcity value. If the company keeps early access to leading-edge Nvidia capacity, it can monetize constrained supply into premium cloud pricing and better utilization than generic neocloud peers, but that advantage only matters if demand stays ahead of buildout. The risk is execution and financing: any slip in power delivery, GPU availability, or customer concentration can turn “explosive growth” into a high-beta balance-sheet story within one or two quarters.

The consensus miss is that the AI trade is starting to bifurcate into “quality compounders” and “capacity optionals.” NVDA/AVGO are the compounders; NBIS is an option on sustained scarcity and capital discipline. If hyperscaler spend merely normalizes rather than re-accelerates, AVGO and NVDA likely grind higher on earnings visibility, but NBIS could re-rate violently in either direction because its valuation is much more dependent on narrative persistence than on near-term cash generation.

Key Decisions for Investors

• Long NVDA into any 3-5% pullback over the next 2-4 weeks; use a 3-6 month horizon and target outperformance if hyperscaler capex revisions continue, but respect downside if guidance implies a 2026 spending pause.
• Add AVGO on weakness as the preferred AI infrastructure expression over the next 1-2 quarters; pair it against a basket of legacy semi/hardware names to isolate custom-silicon share gains and reduce factor risk.
• Treat NBIS as a tactical long only if liquidity and financing runway remain intact; use call spreads or a smaller cash equity position for the next 1-3 months, since the upside is convex but drawdown risk is high on any execution miss.
• Short underexposed legacy server/networking vendors versus NVDA/AVGO for a 3-6 month relative-value trade; the second-order capex winners should continue to take share while non-AI incumbents face margin pressure.
• If AI capex commentary cools in the next earnings cycle, rotate from NBIS into NVDA/AVGO immediately; the former is the most fragile to sentiment reversal, while the latter two have the strongest earnings backstop.