BCA upgrade equities, downgrades cash as AI boom continues By Investing.com

BCA Research is upgrading equities and downgrading cash, arguing the AI investment boom remains the dominant driver of growth even as the Strait of Hormuz closure squeezes consumers. The firm cited over $400 billion of hyperscaler data-center capex in the past 12 months and a surge in Alphabet, Microsoft and Amazon remaining performance obligations to $1.5 trillion in Q1 2026 from $596 billion in Q1 2025. It upgraded Communication Services to overweight, downgraded EU equities to underweight and Australia to neutral, and said reduced investor exposure plus sensitivity to falling markets skews returns higher.

Analysis

The market is increasingly treating AI capex as a self-funding growth regime rather than a discretionary spend cycle, which is why the usual “consumer deterioration = tech multiple compression” linkage is breaking. That favors the largest platform owners and model infrastructure suppliers because they sit closest to the budget source, while second-order beneficiaries are the picks-and-shovels layers that monetize every incremental training and inference dollar without needing end-demand to accelerate. The key implication is that the revenue visibility premium should widen for names with multi-quarter backlog and dense installed ecosystems, while mid-cap software and semis with weaker strategic entrenchment may lag even if the thematic tape stays hot.

The geopolitical shock matters less through direct energy P&L than through policy constraints: if inflation is re-energized, the odds of a near-term rate-cut narrative collapse, which mechanically supports long-duration growth leaders and penalizes economically sensitive cyclicals. Europe is the cleanest relative loser because it has the weakest mix of energy dependence, slower fiscal transmission, and less exposure to the AI spend impulse; that creates a regional dispersion trade rather than a broad equity-risk-off call. Australia also faces a nuanced headwind: a stronger dollar and softer China-linked demand can offset any commodity insulation, so the move there is more about earnings revisions than headline macro.

The consensus may still be underestimating how much positioning can amplify the next leg up. If institutional exposure has been de-risked into the event, then a squeeze higher in the AI complex could be violent over days to weeks, especially if management teams keep validating capex plans into earnings season. The main reversal trigger is not sentiment exhaustion but evidence that incremental AI revenue monetization fails to catch up to spend within 2-3 quarters; absent that, the tape can stay bid longer than valuation skeptics expect.

Key Decisions for Investors

• Go long NVDA vs. a basket of lower-quality semiconductor names for 1-3 months; thesis is that capital spending concentration favors the highest-attach-rate infrastructure provider, with upside from multiple expansion if AI capex remains endogenously funded.
• Buy MSFT and GOOGL on pullbacks over the next 2-6 weeks; use them as core exposure to AI monetization durability, with the risk that any disappointing cloud/AI booking commentary would compress multiples first in these two.
• Pair trade: long META / short a Europe ex-tech equity proxy for 1-2 quarters; META has the strongest operating leverage to ad + AI productization, while Europe is the most exposed to the macro squeeze and least leveraged to the AI spend cycle.
• Consider call spreads on AMZN into the next earnings window; the asymmetry is favorable if capex intensity is read as strategic rather than wasteful, but trim if fulfillment/retail margins show signs of being sacrificed to fund cloud buildout.
• Avoid adding to cash and defensives here unless real rates reprice materially higher; the risk/reward over the next 4-8 weeks still favors remaining long the AI complex because the combination of low positioning and policy sensitivity can extend the upside blow-off.