AI leaders argue software will adapt – not die – but valuations are stretched

A sudden $1 trillion rout in U.S. software names has sparked debate at Web Summit Qatar, where AI founders and investors argued the software-as-a-service model will adapt rather than disappear even as AI valuations appear stretched. Speakers noted large private valuations (Glean at $7B, Miro at $17B), warned of a near-term correction, and flagged IPO timing for AI giants like OpenAI and Anthropic; Forbes data cited over $340B poured into startups in 2025 with >65% into AI, while reports suggest OpenAI may lose ~$14B this year. Market reactions were mixed—Microsoft (MSFT) and Salesforce (CRM) showed intraday gains—but investors expect a valuation normalization over the next 1–2 years amid continued investor interest and U.S.–China competitiveness on innovation versus scale.

Analysis

Market structure: AI accel benefits oligopolies that control chips, cloud and distribution—NVDA, GOOGL (cloud/AI stack) and AAPL (hardware demand) gain pricing power; mid‑to high‑multiple SaaS (CRM among them) face multiple compression if growth expectations reset. With ~65% of $340bn VC flow chasing AI, supply bottlenecks (chips, engineering talent) will sustain above‑trend hardware/software pricing for 6–18 months, advantaging incumbents with scale and renewable revenue.

Risk assessment: Tail risks include U.S./EU export controls on advanced GPUs, heightened model liability/regulation, or a rapid capital flight after a high‑profile IPO failure—each could erase 20–50% of market value in exposed names within 1–3 months. Immediate (days): elevated cross‑asset volatility; short (weeks–months): valuation normalization over 6–24 months; long (quarters–years): consolidation to 3–5 generational winners. Hidden dependency: broad exposure to Nvidia supply and China scale means geopolitical shocks multiply losses.

Trade implications: Favor concentrated exposure to NVDA (AI silicon) and GOOGL (AI infra/search ads) while trimming pure SaaS multiple risk (CRM). Implement convex options on NVDA (6‑month 20% OTM calls) sized 1–2% notional and hedge legacy SaaS with protective puts; rotate 3–6% portfolio weight into AAPL on confirmed iPhone sell‑through beats.

Contrarian angles: Consensus underestimates recurring revenue defensibility—MSFT and select SaaS with >70% gross margins and net retention >110% are buyable on >15% pullbacks. Beware private AI IPO froth; avoid participating in large unprofitable AI IPOs for 6–12 months unless S‑1 shows path to EBITDA break‑even within 24 months. Monitor OpenAI/Anthropic filings and Nvidia supply updates over next 30–90 days as primary catalysts.

Key Decisions for Investors

• Establish a 2–3% long position in NVDA (ticker NVDA) with a 6–12 month horizon; add on pullback >10%; target +20–40%, stop loss -15%.
• Initiate a 2% long position in GOOGL for AI infra exposure (6–12 months), and 2–3% long in AAPL on confirmed iPhone demand; trim other tech holdings if combined tech weight >25%.
• Reduce exposure to high‑multiple SaaS (example: CRM) by 30–50% over the next 30 days; hedge residual exposure with 3‑month ATM puts sized to cover 50% of position for downside protection to -30%.
• Buy NVDA 6‑month 20% OTM calls sized 1–2% of portfolio for asymmetric upside; avoid participating in unprofitable AI IPOs for 6–12 months unless S‑1 demonstrates EBITDA path within 24 months.