Why ‘sell in May and go away’ could be the wrong move for investors

Nvidia remains the most selected stock in Trade Off portfolios, with its market cap up nearly $1 trillion since the season began, underscoring continued AI-driven positioning. The newsletter argues that the old 'Sell in May' strategy is less relevant, citing S&P 500 gains in 25 of the past 33 May-to-October periods and a 22% rise last summer. It also highlights five AI software 'plumbing' names—CrowdStrike, Palo Alto Networks, Datadog, Cloudflare and Rubrik—as alternative ways to gain AI exposure.

Analysis

The key market implication is not just that AI remains a crowded trade, but that the crowd is still underestimating the breadth of the monetization cycle. If the semiconductor leg has already been repriced, the next marginal dollar of AI capex increasingly accrues to security, observability, and network-control layers where spend is less discretionary and gross retention is stickier. That favors the “plumbing” names as second-derivative winners: they can capture AI budgets without needing consumer-facing product adoption to accelerate immediately.

The competitive dynamic is important: the more AI workloads migrate into production, the more exposure rises to model abuse, data exfiltration, uptime risk, and cloud cost sprawl. That creates a defensive growth pocket for CRWD, PANW, DDOG, and NET that should be less sensitive to valuation resets than the headline AI leaders. RBRK is the highest-beta way to express the same thesis, but it is also the most vulnerable if budget scrutiny intensifies or if buyers decide backup/restore can be deferred.

The contrarian miss is that “AI exposure” is not synonymous with NVDA-style concentration. If portfolios remain overloaded into the obvious winners, the biggest risk over the next 1-3 months is not an outright AI unwind but a rotation: semis can pause while software infrastructure continues to grind higher on durable spend. Seasonality is a weak catalyst for de-risking; relative-performance investors risk underperforming by trying to be too cute during a still-positive tape.

The main catalyst to fade this theme would be a capex digestion phase or evidence that enterprise AI pilots are not converting to recurring spend fast enough. Near term, any broad market pullback would likely hit the crowded mega-cap AI names harder than the security/observability basket because positioning is more one-sided there. The setup favors buying dips in the infrastructure layer rather than chasing the leader after extended moves.

Key Decisions for Investors

• Long CRWD / short NVDA pair for the next 4-8 weeks: expresses AI monetization breadth versus crowded semiconductor ownership; risk is a sharp reacceleration in AI capex that reasserts NVDA leadership.
• Buy DDOG on a 2-3% market pullback and hold 1-3 months: best pure-play on AI workload monitoring and cloud sprawl; upside is multiple expansion if enterprise AI usage normalizes into production.
• Add PANW on weakness over the next 2-6 weeks: lower-beta way to own AI security demand with stronger balance-sheet support; risk/reward skews favorably if the market rotates from growth to quality.
• Use NET call spreads 3-6 months out rather than outright stock: higher torque to AI traffic/security use cases, but position size should reflect execution and valuation risk.
• Treat RBRK as a tactical satellite position only: prefer a small starter or option structure ahead of earnings/catalysts, since it offers the most upside if AI governance spend rises but also the most drawdown if budgets tighten.