The surge of AI-driven mega-cap stocks over 2023–2026 has reshaped the landscape for many smart‑beta and single‑factor ETFs. As a handful of companies outperformed, rule-based momentum, quality and low‑volatility ETFs—designed to tilt into the best recent performers or highest‑quality businesses—became implicitly concentrated in those same names. For ETF investors this concentration raises two linked questions: how material is the crowding, and what should portfolio managers and retail investors do about it?

Why factor ETFs concentrated around AI winners

Factor ETFs apply mechanical screens or weighting schemes: momentum uses past returns, quality uses profitability metrics, and minimum‑volatility favors companies with lower realized volatility. In the AI era, companies producing or deploying foundational models and chips scored highly on multiple metrics—rapid returns, rising margins, and lower volatility as investors bid them up. The result: different factor strategies converged on the same set of mega‑cap winners.

Two structural dynamics accelerate this effect:

  • Index and ETF rules are blind to narrative. They optimize objective metrics and thus can independently identify identical companies as “factor leaders.”
  • Large-cap dominance. Many factor products remain U.S.-large-cap centric, so when a few U.S. mega caps drive market returns they disproportionately influence factor weights.

How to measure concentration and crowding in factor ETFs

Investors need standardized metrics to move beyond impression. Here are practical diagnostics you can run using ETF holdings files (available daily or monthly from issuers):

1. Top-holdings weight

Compute the combined weight of the top 5 and top 10 holdings. If multiple factor ETFs in your portfolio have the same company within their top holdings, that indicates single-stock concentration risk across the sleeve.

2. Herfindahl‑Hirschman Index (HHI)

HHI = sum of squared market weights of holdings (expressed as decimals). HHI quantifies concentration: higher values mean more concentration. Benchmarks and internal policies can set thresholds (e.g., HHI > 0.05 signals elevated concentration for an equity ETF).

3. Overlap and ownership correlation

Measure pairwise overlap among ETFs in your allocation: percent of AUM that overlaps by holdings or a correlation of returns over the last 12 months. High overlap across different factor ETFs indicates herd behavior rather than distinct diversifying sources.

4. Factor loading decomposition

Run a simple factor regression (market, value, momentum, size, quality) on ETF returns. Rising exposure to a “single-stock beta” or to a style correlated with a handful of names is a red flag. Principal component analysis (PCA) can reveal whether returns are driven by one dominant component—often the mega caps.

5. Liquidity and creation/redemption risks

Assess whether an ETF’s trading volumes and authorized participant (AP) mechanisms can handle stress. Concentration in a few mega caps typically increases on‑screen liquidity, but it can create price gaps in fast downside moves if multiple ETFs attempt to redeem simultaneously.

Evidence from common factor ETFs (practical examples)

Many widely used factor ETFs—momentum (e.g., MTUM), quality (e.g., QUAL), low‑volatility (e.g., USMV) and equal‑weight alternatives (e.g., RSP)—showed overlapping top holdings during the AI rally. The overlap is not a failure of construction; it’s an emergent property of different rules selecting for companies benefiting most from AI adoption or semiconductor demand.

For investors that combined multiple factor ETFs to diversify factor risk, the overlap often undermined the intended diversification. A portfolio built with three different “factor” sleeves can end up concentrated in the same five names, leaving it exposed to idiosyncratic shocks to those companies.

Risks that matter

  • Idiosyncratic single‑stock risk: If investor sentiment or regulatory action reverses for a dominant AI winner, concentrated factor sleeves can experience amplified drawdowns.
  • Factor crowding & correlation spikes: Crowding increases the chance that historically low correlations between factors break down during stress, reducing the hedging value of multi‑factor allocations.
  • Liquidity & tracking risk in stress: In sharp selloffs, ETFs may exhibit wider spreads and tracking error if creations/redemptions are slowed, especially for niche factor products with smaller AP networks.
  • Concentration tax consequences: High turnover to maintain factor weights can realize capital gains at the ETF level; though ETFs are tax efficient, concentrated creations/redemptions can complicate the tax picture for non‑ETFs in taxable accounts.

Portfolio responses: a framework for investors

There is no single correct response. Your choice depends on horizon, constraints and whether factor exposure is core or satellite. Below are practical responses graded by investor sophistication.

For retail investors

  • Run a holdings check: don’t assume “factor” equals diversification. Look at top‑10 holdings and overlap with your core portfolio.
  • Prefer multi‑factor or blended factor ETFs if you want a single product with built‑in diversification. These often rebalance to avoid excessive drift toward a handful of names.
  • Consider complementing factor ETFs with equal‑weight or small‑cap exposures to offset large‑cap concentration.

For tactical allocators and institutions

  • Set quantitative concentration limits (e.g., cap top‑5 single‑stock exposure across the sleeve to X% of portfolio). Use overlays if necessary.
  • Employ hedges: single‑stock put positions or index puts can be sized to account for concentrated exposures.
  • Use custom baskets or separate account implementations to exogenously cap exposures while maintaining factor tilts.

For ETF portfolio managers

  • Disclose concentration metrics in factsheets. Investors are increasingly sophisticated and expect visibility on single‑stock risk.
  • Consider guardrails in index construction (sector caps, max single‑name weights) when concentration becomes elevated, or offer a capped‑weight share class.

Scenario planning: stress tests investors should run

Run two stress scenarios to understand vulnerability:

  1. Single-name shock: Assume a 30–40% drawdown in a dominant AI name. Measure portfolio and each ETF’s expected drawdown, and whether correlated holdings amplify losses.
  2. Correlation regime shift: Simulate a spike in cross‑section correlations (e.g., factors moving together). Evaluate diversification erosion in a multi‑factor sleeve.

These scenario analyses don’t predict timing, but they reveal how concentrated factor allocations behave in stress and inform position sizing and hedging costs.

Conclusion: treat factor ETFs like active bets

Factor ETFs remain valuable tools: they offer inexpensive exposure to systematic risk premia and can improve long‑term risk‑adjusted returns. But in the AI era, many factor products have behaved less like passive diversifiers and more like active bets on a handful of mega‑cap winners. The practical implication for investors is straightforward: measure concentration, quantify overlap, and decide whether the factor exposure fits your risk budget. If it does, size it and, where appropriate, hedge it. If it does not, seek alternative constructions—multi‑factor, capped, or equal‑weight—that deliver intended diversification.

In short: don’t assume “factor” means diversified. In 2026, factor exposure often meant concentrated exposure. The only defensible response for prudent investors is transparency, measurement, and intentional portfolio design.