4 Best-Performing ETFs of September

September markets were mixed-to-weak as the S&P 500 fell 0.2%, the Dow declined 4.1%, and the Russell 2000 lost 5.3%, while the Nasdaq gained 2.5%. The 10-year Treasury yield reached its highest level since 2007 after stronger economic data, hawkish Fed commentary, and a 25-bp September rate hike; PFIX gained 24.1% as a rate hedge. Supply disruptions tied to the Strait of Hormuz and Red Sea lifted BWET 82.0% and BNO 12.9%, while AI and data-center demand drove PSI up 20.7%.

Analysis

The common driver is a higher-for-longer inflation-risk premium rather than a clean growth acceleration. That combination favors scarce physical capacity (tankers, energy logistics) and duration-insensitive cash-flow businesses, while pressuring small caps, leveraged transport users, and long-duration software. The second-order inflation impulse should emerge over the next 1-3 months through fuel surcharges and longer transit times, with particular margin risk for retailers and import-heavy consumer names such as TGT, WMT and NKE; pricing power will determine dispersion more than headline sector exposure.

Tanker exposure is likely the most fragile trade after a near-vertical repricing: spot-rate gains do not translate one-for-one into NAV if disruptions normalize before vessels are re-chartered. Prefer liquid listed owners such as STNG and INSW, whose earnings sensitivity can be underwritten from fleet and charter data, over BWET until its current holdings, roll mechanics and premium/discount are verified. A credible shipping-lane reopening could compress tanker equities materially in days, whereas sustained rerouting for 1-2 quarters would support elevated estimates and capital returns.

The semiconductor breadth signal is more durable than a single-name AI rally, but it raises the bar for the next earnings season. Memory and equipment names—MU, LRCX, AMAT and KLAC—need upward 2027 capex and pricing revisions to justify further multiple expansion; otherwise rising real yields can overwhelm solid fundamentals. NVDA remains the clean AI demand proxy, but a relative long in equipment/memory versus NVDA better captures broadening demand while reducing dependence on one company’s valuation and supply cadence.

IVZ is not a material beneficiary of one strong ETF month: management fees on PSI are too small to move consolidated earnings absent sustained net inflows across the ETF platform. The contrarian read is that the crowded expressions are now oil and tanker beta, while the underappreciated risk is a supply-chain-driven inflation surprise forcing further real-rate repricing; that would challenge both cyclicals and high-multiple AI equities simultaneously.

Key Decisions for Investors

• Do not chase BWET after its outsized move. Establish a 1-3 month watchlist long in STNG or INSW only after confirming spot tanker rates remain elevated for 2 consecutive weeks and forward charter coverage is repriced; target 15-25% upside versus a 10-12% stop if freight indices fall more than 20% from entry.
• Implement a 3-6 month relative-value basket: long LRCX and MU, short a beta-adjusted amount of NVDA. The thesis requires continued data-center capex breadth and improving memory pricing; exit if either company guides capex/order demand below consensus or if the basket underperforms by 10%.
• Hedge import-cost and inflation spillover with a 1-3 month long XLE versus short XRT or a basket of TGT/NKE. Size modestly: a de-escalation or meaningful decline in Brent and tanker rates would reverse the relative trade quickly; take profits if the spread reaches 8-10%.
• Maintain PFIX or payer-swaption exposure as a portfolio hedge rather than a return-seeking trade over the next 1-6 months. Add only if long-end yields break prior highs on inflation data; reduce if core inflation decelerates for two consecutive prints or Fed communication shifts toward easing.
• Avoid treating IVZ as a direct ETF-flow beneficiary without evidence of persistent net inflows and fee-rate retention. Revisit after quarterly AUM data; a trade requires proof that flows are broad platform flows rather than performance-driven rotation into a few niche products.