New York Times reporter files lawsuit against AI companies

Veteran reporter John Carreyrou and five other authors sued xAI, Anthropic, Google, OpenAI, Meta and Perplexity alleging the companies trained large language models on copyrighted books without authorization, marking the first suit to name xAI and intentionally brought by individual plaintiffs rather than a class. The filing highlights broader industry exposure — referencing Anthropic’s recent $1.5 billion settlement and noting class members there would receive only ~2% of the Copyright Act’s $150,000 statutory ceiling — signaling continued legal and potential licensing costs that could raise compliance and financial risk for AI firms.

Analysis

Market structure: Copyright litigation increases bargaining power for content owners (NYT, DIS) and raises marginal training costs for LLM vendors (GOOGL/GOOG, META, Anthropic, xAI). Expect a near-term shift toward licensed datasets and proprietary data, favoring deep-pocketed incumbents who can absorb licensing fees, while smaller AI players face higher entry costs and price competition compression.

Risk assessment: Tail risks include injunctions forcing model retraining or takedown (>$1B remediation costs for a major model) or statutory damages per work (up to $150k) creating outsized liabilities; probability low but impact high over 6–24 months. Immediate (days–weeks) effects: elevated IV and modest share weakness; medium term (3–12 months): settlements, licensing markets forming; long term (1–3 years): structural OPEX increase and consolidation.

Trade implications: Favor media/IP owners for licensing upside (DIS, NYT) and hedge or reduce pure-play generative-AI exposure (GOOGL, META). Use option hedges to control downside—expect increased implied volatility; pair trades (long media/short ad-tech AI) exploit relative re-rating risk while awaiting legal clarity.

Contrarian angle: Consensus discounts potential for normalized licensing revenue akin to music/streaming—this could become recurring revenue (10–20% incremental gross margin for large media holders). Also, Big Tech’s balance sheets and ability to negotiate standardized, amortized licenses or acquire content (M&A premium) are under-appreciated; initial selloffs may be overdone if settlements mirror Disney–OpenAI style deals.

Key Decisions for Investors

• Establish a 2% long position in DIS and a 1.5% long in NYT (total 3.5%); target absolute upside of 12–25% over 6–12 months driven by licensing revenue/M&A optionality; set stop-loss at -8% and trim 50% on any >20% run-up.
• Initiate defensive protection for large-cap AI exposure: buy Mar-2026 5% OTM puts on GOOGL and on META sized to cap portfolio downside to ~2% (aggregate premium spend ~0.5–1% of portfolio); increase if implied vol spikes >20% vs 30‑day average.
• Execute a pair trade: long 1% NYT funded by short 1% GOOGL (or equivalently short call spread on GOOGL) — hold 3–9 months and unwind on either a settlement announcement or a court ruling favoring plaintiffs.
• Reallocate 3–5% from high-velocity AI/software names into media/IP and cloud providers with enterprise AI controls within 30 days; revisit after 90 days or upon a major judicial/legislative catalyst (class cert, injunction, or Disney‑OpenAI style license).