Markets are running a live experiment on which approach to AI exposure carries the least downside when spending and credit concerns take center stage. As several infrastructure-focused stocks plunged in intraday trading, the most diversified AI exchange-traded funds posted noticeably smaller losses, according to intraday moves.
Today's moves in infrastructure names
Dell Technologies (DELL) tumbled roughly -10.2% intraday, Vertiv (VRT) slid about -7.4%, and Arm (ARM) dropped near -7.9% - all companies closely linked to AI infrastructure and data-center spending. Those sharp moves have put concentrated AI hardware and semiconductor strategies under pressure.
ETF performance in the stress test
Several candidate funds saw materially different outcomes in the same market environment:
- KOMP (SPDR Kensho New Economies Composite): AUM $2.62B, intraday -0.99%, one-year +17.9%, expense ratio 0.20% - lowest volatility among the listed funds.
- AIQ: AUM $9.26B, intraday -1.37%, one-year +29.9%, expense ratio 0.68% - balanced AI exposure with less pure hardware concentration.
- ARTY: AUM $3.47B, intraday -3.62%, one-year +52.7% - higher recent performance but greater intra-day drawdown.
- BAI: AUM $12.8B, intraday -5.28%, one-year +39.5% - the largest fund and the most exposed to capex-sensitive names.
Why BAI proved vulnerable
BAI is the largest AI ETF in the group at $12.8 billion, but its concentration in capex-dependent technology names has made it susceptible to the current rout. Its top holdings include SK Hynix (6.6%), Micron (MU, 6.4%), AMD (5.5%), and NVIDIA (NVDA, 5.0%). Those positions amplify exposure to the question the market is asking now: whether the anticipated surge in hyperscaler capex can generate matching returns. The ETF is 91% allocated to Technology and has significant Asian semiconductor weightings; SK Hynix was down around -11% in Seoul during the selloff. Active management has not insulated BAI from the downturn.
AIQ: a measured AI exposure
AIQ, with $9.26 billion in assets, appears to strike a middle ground for investors seeking AI exposure without an all-in on infrastructure. Its leading holdings include Cisco Systems (CSCO) at 4.2%, Apple (AAPL) at 3.7%, and NVIDIA kept to around 3%. The fund is 77.3% allocated to Technology versus BAI’s 91%, and it includes a meaningful Communication Services weighting (10.8%), capturing companies that monetize AI rather than build most of the hardware. Among passive AI ETFs, AIQ had the highest financial health momentum score cited here, at 4.24 out of 10.
KOMP: diversification as defense
KOMP, which carries the lowest expense ratio at 0.20%, is the least narrowly focused "AI ETF" among the group. It spreads risk across multiple new-economy themes - AI, genomics, robotics, clean energy, and cybersecurity - reducing reliance on infrastructure spending. That broad mix helped KOMP limit losses to under 1% while more concentrated AI funds declined 3% to 5% in the same session. The trade-off is lower AI intensity; KOMP is better suited for investors prioritizing risk control rather than maximal AI upside.
Comparing fits by priority
- Lowest volatility - KOMP: Broad diversification, low fees, down -0.99% in the session cited.
- AI exposure with risk balance - AIQ: Delivers meaningful AI exposure while avoiding an overconcentration in capex-driven hardware.
- Performance over safety - ARTY: Strong one-year return (+52.7%) but contains volatile names such as CoreWeave and SMCI, increasing drawdown risk.
Macro backdrop and conclusion
The broader macro context weighs on the assessment. Fitch flagged AI-related capital expenditure as a top global credit risk and the market is pricing in a near-term surge in hyperscaler spending projected to reach roughly $700 billion this year before there is broad revenue validation. In this environment, the investment approach that limits bets on the infrastructure buildout - and instead focuses on adopters and software enablers - has been the less risky path. The intraday divergence between software names that rallied (for example, Adobe and ServiceNow) and the hardware group that sank provides a clear illustration of that dynamic.
For investors, the choice between funds should map to risk tolerance: hold diversified adopters and enablers for lower volatility, or accept larger drawdowns for concentrated hardware-orientated AI exposure.