Apollo's Sambur says software's AI troubles will persist, noting the 'very large unknowns'

IGV Software ETF has rebounded ~3% in March but remains down ~20% YTD. Apollo's David Sambur warns the software selloff is "very early," forecasting historic, rapid displacement from AI that will pressure revenue models, gross margins and valuations. Several large software names (Intuit, HubSpot, Salesforce) are announcing buybacks, which Sambur views as selective opportunities but RBC warns may signal companies "waving the white flag" and reduce capital available for future M&A and innovation.

Analysis

AI-driven cost-to-compete is a margin reallocation force, not just a feature wedge: as foundational models continue to lower the incremental cost to deliver intelligent features, smaller and horizontally-focused SaaS vendors face a two-front squeeze — lower price per feature and higher engineering spend to differentiate. Expect the economics to play out over 6–24 months as customers test prototypes, force-vendor comparisons, and then consolidate spend onto suppliers with proprietary data access and integration depth.

The primary beneficiaries are platform and infra owners (cloud, model-hosting, data platforms) that capture the incremental compute and deployment spend; secondary winners are companies that turn unique first-party data into defensible ML advantages. Conversely, point-solution vendors targeting SMBs or selling easily replicated modules should see churn and slower new ARR — a structural bifurcation that will amplify dispersion across the software cohort.

Key reversals are plausible: if enterprise integration costs, security requirements, or regulatory constraints materially slow deployments, the market could re-rate incumbents within 3–9 months. Tail risks include rapid open-source model advances that abruptly compress value for many modules (a 1–2 quarter acceleration scenario) or antitrust action that reshapes distribution economics — both can trigger >30% repricing episodes for exposed names.

The market appears to be over-indexing on binary outcomes; not every vendor is a casualty. Names with durable revenue fungibility (deep transaction or tax data, critical workflow lock-in) will buy time to reprice and monetize AI, so selective ownership with explicit stress tests for 20–40% margin compression is the prudent path forward.

Key Decisions for Investors

• Pair trade (3–12 months): Long INTU (1.25% NAV) / Short HUBS (1.25% NAV). Rationale: take a relative-value view that data-moat incumbents will preserve monetization vs SMB-focused, easily-replicated offerings. Risk management: set pair stop at 8% adverse move and size to limit portfolio tail to 2% NAV. Target relative return 20–40% if thesis unfolds.
• Hedge/insurance (6–12 months): Buy CRM put spread (cost-limited structure sized 0.5–1% NAV) to protect against a downside re-rating driven by feature commoditization or revenue churn. Reward: asymmetric payoff if CRM re-prices >25%; cost capped, preserves upside if integration lifts ARR. Exit on signs of sustained re-acceleration in AI-driven net-new ARR growth.
• Income/convexity trade (12–24 months): Long INTU equity (1–2% NAV) and sell 9–12 month covered calls at ~10–15% OTM to monetize elevated option premia while retaining upside exposure to durable data moats. Expect to collect yield that offsets near-term multiple compression; risk is forced buyback of calls on unexpected positive prints.
• Opportunistic event allocation (6–18 months): Allocate a small sleeve (0.5–1% NAV) to private-equity style deals or funds (use APOS or similar exposure) that can buy distressed or mispriced software assets if M&A/innovation stalls. Rationale: buy-side capital arbitrage when strategic buyers pull back; downside: deal competition from other PE sponsors and financing cost volatility.