Nicoletos: AI Makes a Case for Active Strategies

The article argues that passive investing may be creating an "index trap" as AI winners and losers are bundled together, with Nvidia cited as contributing 20% of the S&P 500’s 2025 return. It also notes that 79% of active strategies underperformed in 2025, underscoring the debate between active and passive management. Separately, U.S. oil rig counts have stayed flat even as crude prices nearly doubled during the Iranian conflict, suggesting drillers expect elevated prices to fade and one-year futures are only 10-15% higher.

Analysis

The bigger market implication is not simply that passive ownership is crowded, but that it is becoming a forced-flow amplifier in names where fundamentals are diverging fastest. When index-linked capital keeps allocating into both AI infrastructure winners and structurally challenged software/adjacent incumbents, the dispersion trade can persist longer than fundamentals alone would justify because marginal buyers are insensitive to valuation. That creates a setup where single-stock idiosyncrasy matters more, but only in the subset of names with real earnings power and pricing authority.

NVDA remains the cleanest beneficiary in the near term because it sits at the intersection of capital intensity, supply scarcity, and reflexive index ownership. The second-order risk is that the same passive machinery that supported the upside can deepen drawdowns in a growth scare if flows reverse, especially in crowded AI bellwethers and the lower-quality software cohort that lacks a balance sheet moat. In other words, the market may continue to reward infrastructure while simultaneously compressing the multiple of anything that looks like a non-essential application layer.

On the energy side, the signal from rig behavior argues the cash spike is not yet being validated by supply response, but the forward curve is already discounting mean reversion. That sets up a classic dislocation: spot strength can persist for weeks, while 6-12 month futures remain anchored by market skepticism. The tradeable takeaway is that geopolitical headlines may still move front-end oil, but the medium-term setup favors fading overstretched spot enthusiasm unless the conflict widens or exports are disrupted materially.

The contrarian point the market may be missing is that “active wins” is not a universal statement; it is a barbell statement. Most active managers will still underperform because dispersion is high but hit rates are low, while a concentrated basket of the right infrastructure and defense names can outperform dramatically. The opportunity is therefore not generic stock-picking, but very selective exposure to real operating leverage and away from crowded “AI-adjacent” business models that can’t defend margin.

Key Decisions for Investors

• Initiate a relative-value pair: long NVDA / short equal-dollar basket of lower-quality software beneficiaries of AI skepticism for 3-6 months; thesis is continued capex pull-through versus multiple compression in names with weaker product differentiation. Risk: if AI spend pauses abruptly, NVDA can de-rate faster than the short basket.
• Buy NVDA on 1-2 week pullbacks rather than chasing strength; use call spreads 3-6 months out to express upside with defined downside. Reward/risk is attractive while hyperscaler capex remains the dominant incremental demand driver.
• Fade spot oil spikes with a short-dated crude call spread or short front-end energy beta versus long integrated E&Ps only if the conflict headline fades; the rig data implies the market is already looking through the shock. Risk: renewed supply disruption invalidates the mean-reversion trade quickly.
• Add selective exposure to AI infrastructure suppliers with pricing power, but avoid broad tech ETFs for new capital. The index structure forces ownership of weak links, so ETF exposure is a diluted expression of the theme with less convexity than single names.
• Use any broad-market drawdown driven by passive de-risking to scale into quality compounders rather than index products; the key is to own names with self-funded growth and low dependence on multiple expansion.