Scholarlink

<p>We study AI investment vehicles by comparing AI-investing ETFs (thematic exposure to AI adoptors), AI-powered selection ETFs (robo/algorithmic stock pickers), and the “Magnificent Seven” tech giants (individually and as equal-weight and optimized portfolios). Using daily returns net of stated ETF expense ratios, we evaluate risk-adjusted performance (Sharpe, Sortino, Treynor, Omega and Information) and Downside tail/risk, (semivariance, downside deviation Value at Risk, Expected Shortfall (ES), maximum Drawdown). We also employ to supplement tail risk estimation. We use the Technology Select Sector SPDR Fund (XLK) as a benchmark. To synthesize results, we employ a performance heatmap and unsupervised learning (K-means and K-medoids on principal-component scores). Benchmarks include the technology sector (XLK) and the broad market (S&amp;P 500). Three findings emerge. First, NVIDIA and Tesla form a distinct high-risk, high-reward group with the largest tail losses. Second, among AI ETFs, XAIX exhibits the most balanced risk-adjusted performance, while other AI thematic and AI-powered ETFs underperform both sector and broad-market benchmarks. Third, factor-model robustness (Fama-French five-factor, excess returns with Newey-West errors) shows no abnormal return for AI ETFs, whereas the equal-weight Magnificent-Seven portfolio delivers a statistically significant positive alpha that strengthens post-COVID. These results suggest that concentrated exposure to leading AI innovators dominates packaged AI ETF products over our sample (May 2019-Sep 2024), and they provide a practical map of the AI investment landscape for investors with different risk tolerances.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

The Rise of AI in Exchange-Traded Funds: Assessing Robo-Advisors and AI-Exposed ETFs

  • 期刊论文

  • 发表日期：

    2026年2月21日

查看全文
  • Monia Antar

摘要

We study AI investment vehicles by comparing AI-investing ETFs (thematic exposure to AI adoptors), AI-powered selection ETFs (robo/algorithmic stock pickers), and the “Magnificent Seven” tech giants (individually and as equal-weight and optimized portfolios). Using daily returns net of stated ETF expense ratios, we evaluate risk-adjusted performance (Sharpe, Sortino, Treynor, Omega and Information) and Downside tail/risk, (semivariance, downside deviation Value at Risk, Expected Shortfall (ES), maximum Drawdown). We also employ to supplement tail risk estimation. We use the Technology Select Sector SPDR Fund (XLK) as a benchmark. To synthesize results, we employ a performance heatmap and unsupervised learning (K-means and K-medoids on principal-component scores). Benchmarks include the technology sector (XLK) and the broad market (S&P 500). Three findings emerge. First, NVIDIA and Tesla form a distinct high-risk, high-reward group with the largest tail losses. Second, among AI ETFs, XAIX exhibits the most balanced risk-adjusted performance, while other AI thematic and AI-powered ETFs underperform both sector and broad-market benchmarks. Third, factor-model robustness (Fama-French five-factor, excess returns with Newey-West errors) shows no abnormal return for AI ETFs, whereas the equal-weight Magnificent-Seven portfolio delivers a statistically significant positive alpha that strengthens post-COVID. These results suggest that concentrated exposure to leading AI innovators dominates packaged AI ETF products over our sample (May 2019-Sep 2024), and they provide a practical map of the AI investment landscape for investors with different risk tolerances.