Research Roundup: Tech stock prices down sharply, profit growth outlook the same

Scotiabank says the S&P 500 Tech sector is now down to 21.5x forward earnings from 31x last October, while still showing stronger forward EPS momentum and expected Q1/26 earnings growth of 26% y/y. The note sees AI-related semiconductor demand still strong, citing March export surges of 150% in Taiwan and 46% in South Korea, and reiterates a small overweight in tech. BofA’s fund manager survey turned more bearish, with growth expectations plunging and inflation expectations rising, while Morgan Stanley remains constructive on equities, favoring cyclicals and quality growth.

Analysis

The set-up is less about “AI enthusiasm” and more about the market finally paying for earnings durability with a lower multiple. If semiconductor export data is still accelerating while the sector’s valuation premium has compressed, the asymmetry shifts toward names with visible capex-to-revenue conversion rather than the broad software basket; the market is likely underestimating how quickly incremental AI spend can translate into upward revisions for infrastructure, network, and fab-equipment beneficiaries. The key second-order effect is that accelerating demand in the hardware layer can coexist with multiple compression in software if buyers become more selective about who can monetize AI and who is just funding it.

Positioning is a bigger driver than fundamentals in the next few weeks. The fund manager survey suggests investors are still long risk, but crowded in semis and oil, which means the cleaner trade is not chasing the obvious winners outright but expressing the view through relative value and timing around earnings. In a tape where growth fears are rising yet recession odds remain low, sectors with self-funding earnings momentum should outperform when macro noise fades; that favors quality tech and hyperscalers over capital-intensive cyclicals that need a perfect macro backdrop.

The main tail risk is that geopolitics re-price inflation and rates faster than earnings can re-rate tech. If oil stays elevated for several weeks, duration-sensitive software and speculative AI beneficiaries can lag even if the semiconductor chain remains strong, because the market will demand proof of monetization before rewarding long-duration cash flows. Conversely, if energy spikes reverse quickly, tech can catch a two-stage bid: first on lower discount rates, then on earnings revisions as capex orders and cloud demand stabilize.

Consensus appears to be over-anchored to the idea that AI capex is a future margin headwind. The more important read-through is that capex revisions are still being outpaced by earnings revisions, which implies the market may be underestimating the near-term operating leverage in the winners and overestimating the competitive threat to incumbents with distribution, data, and enterprise relationships. That argues for owning the providers of picks-and-shovels and the hyperscalers, while fading lower-quality software where AI narrative support has outrun monetization visibility.

Key Decisions for Investors

• Add a tactical long in XLK or SMH into any 1-2 day weakness; target a 4-6 week holding period into earnings with a 2:1 upside/downside if revisions continue to outpace capex concerns.
• Pair trade: long SMH / short IGV for the next 1-3 months. The thesis is hardware revenue visibility and AI order strength versus software multiple risk; stop if software starts delivering clear AI monetization beats.
• Own MS as a relative-benefit financial/market-structure expression if equity issuance, trading volumes, and institutional activity remain elevated; use 6-8 week horizon, with upside from volatile markets rather than direction alone.
• Buy call spreads on QQQ or XLK expiring after Q1/26 earnings for a defined-risk expression of the “earnings beats beat geopolitics” view; size small because oil-driven rate shocks can blunt the move.
• If semis rally hard into earnings, fade the most crowded names via a modest hedge in the highest-multiple, lowest-free-cash-flow AI software cohort; the better risk/reward is owning the picks-and-shovels, not the story stocks.