TD Cowen sees software stock rebound after worst quarter since 2008

IGV (iShares Expanded Tech-Software) fell 1.9% in March, completing a Q1 cumulative decline of -24% (Jan -14.5%, Feb -9.7%, Mar -1.9%)—the worst three-month stretch since Q4 2008. TD Cowen cites negative investor sentiment with a long-semiconductor/short-software positioning, AI model announcements (Anthropic Claude, leaked Mythos) pressuring software, and booking misses at Adobe, Salesforce, Workday and SAP. IGV outperformed SOXX by 5.0% in March after deep underperformance earlier in the quarter; TD Cowen’s broader software basket trades at ~4.1x EV/NTM sales (down from 4.5x in Feb and 5.6x in Jan). Historical data shows modest average 3-month rebounds after large IGV drawdowns (avg +1.9%, median +5.2%).

Analysis

Market action appears driven more by positioning and narrative volatility than by broad demand deterioration. A crowded long-semiconductor/short-software trade can amplify intramonth moves: transient de-risking in semis forces software-covering, then re-grossing in semis flips the cross again, creating outsized short-term dispersion that will dominate P&L over weeks rather than quarters.

Proliferation of new AI models is creating a bifurcated competitive landscape: infrastructure and chip suppliers win from increased inference and training spend, while application vendors face compressed differentiation and faster marginal churn as customers experiment with cheaper model embeds. That dynamic increases idiosyncratic risk — best-in-class SaaS with high net retention will decouple from lower-quality names, and cloud providers become the implicit clearinghouses for model distribution and cost volatility.

Key catalysts to watch that will change the trend are concrete ARR/NRR inflection (quarterly), visible enterprise renewals or large multiyear contracts (45–90 days), and any coordinated re-rating from PE M&A activity (3–12 months). Near-term tail risks include another wave of model announcements that reset expectations and liquidity-driven deleveraging in macro weak spots; technical squeezes can however produce 1–6 week reversals that are tradable.

The consensus is treating the software selloff as homogeneous; it's not. The opportunity is dispersion: trade the cross and use options to harvest asymmetric payoffs around earnings/guidance windows while scaling fundamental longs in high-retention, free-cash-flow-positive names for a 6–12 month recovery.

Key Decisions for Investors

• Pair trade (1–3 months): Long SOXX vs Short IGV — enter on intraday pullback in semiconductors or on IGV weakness. Target 15–25% gross return on reversion; hard stop 6–8% on the pair. Rationale: exploit positioning flips that historically create squeezes and mean reversion.
• Options asymmetric (3–9 months): Buy NVDA 3–6 month call spread (buy near-the-money, sell 10–15% OTM) sized for 2:1 upside vs max loss. Use as leveraged exposure to AI-driven hardware spend while capping premium decay.
• Short-volatility/credit event (1–3 months): Buy 3-month put spreads on 2–3 large-cap, low-NRR software names (example tickers: CRM, ADBE) ahead of next earnings if guidance risk remains elevated. Max loss = premium paid; target 2–3x payoff if guidance/booking miss recurs.
• Fundamental long (6–12 months): Accumulate high-retention, free-cash-flow-positive mid-cap SaaS on 10–15% intraday capitulation (example tickers: ZS, DDOG as screening candidates). Size modestly with conviction laddering; target 30–60% return on visible ARR recovery while stopping out on secular NRR decline.