AI's Public Image Is Losing Ground: 4 Defensive ETFs in Focus

AI sentiment is weakening amid concerns about elevated valuations, large capital spending, uncertain monetization, job losses, power costs and safety risks; Michael Burry says he is shifting from outright shorts toward put options, which he views as attractive amid exceptionally low volatility. The article highlights four defensive ETFs: QUAL gained 1.4% over the past month and 16.2% over six months, while VYM fell 4.7% and RSP fell 4.2% over the past month; JPST fell 0.3%. Treasury yields are described as being at multiyear highs, but no specific yield level or market-wide price reaction is reported.

Analysis

The key risk is not negative AI rhetoric by itself; it is a change in the hurdle rate for continued investment. If cloud providers’ AI revenue fails to validate infrastructure spending, capex expectations and supplier pricing power can reset together. For Micron (MU), expanding Chinese memory capacity would amplify the ordinary memory-cycle downside; watch pricing and inventory data, not executive commentary alone. Nebius (NBIS) is exposed to a different test: whether infrastructure utilization and customer commitments support its investment burden. The article provides no company-level figures to establish that either thesis is already in earnings.

The suggested “defensive” basket is not uniformly defensive. QUAL may reduce exposure to weaker balance sheets, but remains equity beta; RSP diversifies concentration rather than market risk and can lag if mega-cap AI leadership resumes. VYM can be rate-sensitive, while JPST still carries credit and liquidity risk and is not a cash substitute. Higher yields can pressure both equity multiples and dividend proxies, so the proposed rotation may disappoint if rates rise further.

Near term, positioning and volatility—not a proven collapse in AI demand—are the likely drivers. Over 1–3 months, monitor hyperscaler capex commentary, memory pricing, data-center/power permitting and election-related policy signals. Over 6–18 months, the decisive test is monetization versus installed capacity. A sustained revenue acceleration and stable memory pricing would falsify the bearish case; weaker utilization, capex cuts or inventory accumulation would strengthen it.

Key Decisions for Investors

• Do not treat the article’s sentiment shift as an earnings signal. For MU and NBIS, review exposure and set alerts for memory-price/inventory deterioration and utilization or customer-commitment weakness, respectively.
• If portfolio exposure is concentrated in AI hardware, consider a defined-risk put spread on MU or an appropriate semiconductor vehicle only after checking current implied volatility, skew, and event pricing. Avoid outright shorting on this evidence alone; cap premium at risk and reassess after earnings or capex guidance.
• Use QUAL or RSP only as a measured concentration rebalance, not as a hedge. Compare their factor and sector exposures with the portfolio first; RSP can increase exposure to rate-sensitive or cyclical constituents.
• Treat JPST as a short-duration credit allocation rather than cash. Verify portfolio credit quality, liquidity, and yield after fees; further rate increases or credit-spread widening could undermine its defensive role.