Software market selloff: AI invstors banked every company would win, says Deutsche

Software and adjacent-sector stocks have undergone a significant repricing as investors reassess AI-driven disruption, with JPMorgan estimating roughly $2 trillion wiped off software market caps and companies in legal, IT, consulting and logistics also hit. Deutsche Bank frames the move as a correction of overly optimistic expectations, while market commentators warn that rapid obsolescence from large language models could accelerate sell‑offs and sustain volatile, sentiment‑driven leadership rotation despite an intact long-term secular AI thesis.

Analysis

Market structure: The $2tn-ish software repricing signals a bifurcation—winners are AI infrastructure owners and hyperscalers (NVDA, MSFT, GOOGL, AMZN) and data‑center/energy plays; losers are mid/small‑cap SaaS, professional services (legal/consulting) and logistics firms facing demand compression and margin tailwinds. Pricing power will concentrate: owners of compute, LLM training data and platform distribution can extract 200–500bp higher gross margins vs. service providers that compete on labor. Supply/demand: GPU scarcity and datacenter capacity keep hardware capex elevated near term, while demand for human‑delivered services may fall 5–15% in affected niches over 12–24 months.

Risk assessment: Tail risks include rapid regulatory action (EU AI Act + U.S. liability rules) forcing model retraining/labeling costs (+10–30% OPEX), catastrophic model failures causing litigation, or a GPU supply shock. Immediate (days) — elevated IV and sentiment swings; short term (weeks–months) — guidance resets and rotation; long term (quarters–years) — structural share shifts. Hidden dependencies: legacy SaaS revenue linked to professional services renewals and office occupancy; second‑order hit to commercial CRE and corporate travel if productivity gains persist. Key catalysts: earnings guidance (next 2–8 weeks), new chip releases, major model deployments, and regulatory announcements.

Trade implications: Favor concentrated exposure to NVDA (core AI compute) and hyperscalers; hedge broad software beta. Use pairs: long NVDA (2–3% NAV) vs. short IGV (1–2% NAV) to express infrastructure vs. mid‑cap SaaS divergence. Options: buy 3‑month NVDA call spreads (5–15% OTM) and buy 3‑month IGV put spreads (10% OTM) to play skewed volatility. Rotate 5–10% from professional services/logistics into data‑centre REITs (EQIX) and utilities with predictable cashflows.

Contrarian angles: Consensus underestimates durable capex by hyperscalers — even if software revenue shifts, cloud vendors lock in multi‑year spend; overreaction likely in high‑churn, low‑NRR SaaS. Historical parallel: post‑internet bubble winners consolidated after an initial wipeout; expect similar consolidation with 6–18 month time‑scale. Unintended consequence: accelerated AI adoption can raise cybersecurity and compliance budgets—look for mispricings in high‑quality security names with >120% NRR that were sold off indiscriminately.

Key Decisions for Investors

• Establish a 2–3% NAV long position in NVDA within 1–4 weeks, layering on 5–10% pullbacks; trim if position rises +25% within 3 months or if company guidance misses consensus by >10% revenue.
• Initiate a 1–2% NAV short position in mid‑cap SaaS via IGV (ETF) or select names with NRR <100% and >30% revenue from services; add if IGV rallies back within 10% of pre‑selloff levels, cover if IGV falls another 15%.
• Implement options hedges: buy 3‑month NVDA 5–15% OTM call spreads (funded by selling nearer‑dated calls) and buy 3‑month IGV 10% OTM put spreads sized to offset 30–50% of directional short exposure.
• Rotate 5–8% of equity allocation from consulting/logistics exposures into data‑centre REITs (e.g., EQIX) and electric utilities over 30–90 days, targeting names with >4% dividend yield and <8x EV/EBITDA.
• Monitor regulatory and guidance catalysts over next 30–90 days: reduce AI/infrastructure longs by 50% if (a) EU/US regulation materially restricts model deployment or (b) top hyperscaler cuts capex guidance by >20% YoY; increase allocation if guidance is maintained or capex cut <10%.