New York Times sues Perplexity AI for ’illegal’ copying of content

The New York Times has filed suit in the Southern District of New York accusing AI startup Perplexity of copying, distributing and displaying millions of NYT articles without authorization to power its generative AI products, alleging fabricated content and false attribution alongside NYT trademarks and seeking damages and injunctive relief. Perplexity — valued at roughly $20 billion and already facing related suits from the Chicago Tribune, Reddit, Encyclopedia Britannica and Dow Jones/New York Post — insists it indexes pages and provides citations rather than scraping for foundation-model training; NYT shares rose 1.8% on the news.

Analysis

Market structure: Publishers (NYT) gain bargaining leverage — successful suits or settlements could create recurring licensing revenue and raise marginal cost of training for independent LLM builders, concentrating economic rents with well-capitalized incumbents (Big Tech + legacy media). Startups that rely on unlicensed scraping (Perplexity and similar) are immediate losers; hardware vendors (SMCI, APP) remain indirect beneficiaries as absolute compute demand is likely delayed, not destroyed. Equity volatility should rise for small-cap AI names (expect 20–40% realized/IV expansion on headline risk) while credit spreads widen for private AI financings.

Risk assessment: Tail risks include injunctions requiring data deletion or retraining (10–30% downside for exposed model-builders) and legislated mandatory licensing or data taxes that compress margins. Timeline: immediate (days) — headline-driven 5–15% swings; short-term (weeks–months) — settlements, preliminary rulings; long-term (12–36 months) — regulatory frameworks and durable licensing markets. Hidden dependencies: cloud providers’ contracts, dataset provenance, and insurance coverage could transmit losses to larger vendors unexpectedly.

Trade implications: Favor liquidity and quality: consider long NYT as a defensive licensing play and long SMCI/APP as hardware-levered exposure to secular AI compute demand; underweight or hedge pure-play scraping/extraction names (small-cap AI/consumer data firms). Use options to buy downside protection on volatile AI names and structured call spreads on hardware to reduce premium spend; size to 1–3% of portfolio with clear stop-loss and catalyst windows (3–6 months).

Contrarian angles: Consensus assumes legal action kills AI growth; history (music, search) shows litigation often forces licensing and higher-margin incumbents — this benefits NYT/AMZN and entrenches cloud providers. The market may be underpricing hardware providers (SMCI, APP) which could rerate +25–50% over 12 months if compute demand persists despite legal noise. Unintended consequence: heavy licensing could accelerate closed, Big Tech–dominated models and raise barriers to new entrants, amplifying winner-take-most dynamics.

Key Decisions for Investors

• Establish a 2–3% long position in The New York Times Co (NYT) within 30 days to capture potential licensing settlements; target +25–40% in 6–12 months, use a 12% stop-loss and consider selling 10–15% OTM calls to finance carry.
• Allocate 3% to Super Micro Computer (SMCI) via a 3–6 month call spread (buy ATM call, sell 20% OTM) sized to equal a 3% delta exposure; take profits at +30–50% or cut if quarterly guidance shows >10% downside to server/bookings growth.
• Establish a 1–2% short exposure to a small-cap AI/scraping basket (include RDDT-sized positions) to hedge headline litigation risk; add up to 0.5% if a name gaps >15% on positive sentiment and remove on settlement favorable to publishers.
• Purchase 3-month puts equal to ~1% portfolio protection on a bespoke AI basket (or RDDT 15% OTM puts) to cap tail risk from adverse court rulings; if a preliminary injunction is filed within 90 days, increase protection to 2%.