Howard Marks Flags Uncertainty in AI Investing

Oaktree co-founder Howard Marks said market optimism has prevailed since late 2022, with investors increasingly looking past adverse news, warranting greater caution. He said enthusiasm for AI is clear, but unresolved questions around profitability, valuations and employment effects make it impossible to judge whether current AI-related exuberance is irrational.

Analysis

The actionable signal is not a directional AI call but a regime warning: when broad risk appetite is conditioned to discount negative surprises, valuation dispersion and downside convexity matter more than index exposure. The most vulnerable cohort is unprofitable or low-visibility AI-adjacent software and infrastructure names whose multiples embed sustained 30%+ growth and rapid margin scaling; any guide-down can trigger simultaneous earnings and multiple compression. Near term, this favors quality and cash-flow screens over wholesale de-risking, since positioning can remain resilient absent a concrete earnings or macro disappointment.

Over the next 1-3 months, the catalyst path is AI capex monetization evidence: hyperscaler spending must translate into incremental cloud revenue, enterprise software bookings, or margin leverage rather than merely sustaining semiconductor demand. A widening gap between capex growth at MSFT, GOOGL, AMZN and their disclosed AI-related revenue contribution would pressure second-order beneficiaries first, especially richly valued application software. Conversely, continued upward revisions to hyperscaler cloud growth and broadening enterprise deployment would invalidate a bearish valuation thesis.

The contrarian view is that skepticism focused on AI profitability may be premature for the platform owners: they can absorb investment cycles through diversified cash flows while using AI to defend search, cloud, advertising and productivity moats. The cleaner hedge is therefore long profitable incumbents with distribution and free-cash-flow support versus short high-duration software, rather than shorting AI infrastructure leaders outright. Structural risk over 6-18 months is labor substitution and IT-budget reallocation: automation beneficiaries may gain operating leverage, while services, BPO, and commoditized SaaS vendors face pricing and seat-count pressure.

Key Decisions for Investors

• Maintain core AI exposure but rotate toward quality: favor MSFT and GOOGL over long-duration application software; reassess after the next earnings cycle based on cloud growth, AI revenue disclosure, and capex guidance.
• Consider a 3-6 month relative-value basket: long MSFT/GOOGL versus short IGV or a basket of cash-burning, premium-multiple SaaS names. Thesis is multiple dispersion, not an outright Nasdaq decline; exit if IGV earnings revisions stabilize and valuation premiums re-expand despite weak bookings.
• Use QQQ put spreads rather than outright shorts to hedge a sentiment reversal: target 3-6 month maturities and strikes roughly 8-12% below spot, where convexity is more attractive than paying for near-spot protection. The hedge is falsified by broad upward 2027 EPS revisions and easing real yields.
• Watch hyperscaler capex-to-monetization metrics as a trigger, not a trade by itself: if aggregate MSFT/AMZN/GOOGL capex guidance rises while cloud growth or operating-margin guidance falls for two reporting periods, increase the software short leg; if cloud growth accelerates with stable margins, remove it.