Nikhil Kamath announces upcoming podcast with Anthropic CEO Dario Amodei: ‘No script conversation about the tech…’

Zerodha co-founder Nikhil Kamath will host Anthropic CEO Dario Amodei on the 'People by WTF' podcast on Feb. 22, 2026 for an unscripted discussion about AI’s economic, geopolitical and safety implications. Amodei framed AI as bringing enormous opportunities alongside serious risks — autonomous behavior, misuse by actors including states, and potential economic displacement — highlighting regulatory and operational uncertainties investors should monitor, though the episode contains no company-specific financial data.

Analysis

Market structure: Demand is concentrating on AI compute, cloud inference, and safety tooling—clear winners are NVIDIA (NVDA), TSMC suppliers, Microsoft (MSFT) and Alphabet (GOOGL) for cloud+model ops, and cybersecurity vendors for mitigation services. Losers in a multi-year scenario include labour-heavy outsourcers and ad-monetization dependent digital incumbents as automation compresses margins; pricing power will shift to GPU/accelerator providers and cloud infra where scarcity of high-end silicon sustains >20% gross margins for suppliers.

Risk assessment: Tail risks include rapid regulatory limits on high-capability models (US/EU export controls, or China data localization) and high-profile misuse that triggers immediate procurement freezes; these are low-probability but could shave 20–40% off sector multiples in 3–6 months. Near-term (days–weeks) market reaction to a podcast is negligible; medium-term (months) is driven by model releases, legislation and earnings; long-term (years) is structural displacement of jobs and value capture by platform+compute owners. Hidden dependency: concentrated GPU supply (NVIDIA/TSMC) and power/cooling bottlenecks in data centres.

Trade implications: Favor concentrated long exposure to NVDA (core long), diversified cloud names (MSFT, GOOGL) and security plays (CRWD/PANW); pair trades: long NVDA vs short legacy IT services (CDW/CTSH) to express automation spread. Use 3–9 month call spreads on NVDA to capture upside while financing premium, and buy protective put spreads on SOXX (or NVDA) to hedge a regulatory shock. Rotate portfolio into semis, cloud and cyber over next 1–3 months, trimming consumer ad-exposed names by 5–10%.

Contrarian angles: Consensus focuses on runaway growth; underappreciated is near-term regulatory tightening that benefits controllable/safety-first firms (Anthropic-like partnerships) and regional champions (BIDU) with local data moats. NVDA is increasingly priced for perfection—look for entry on 8–15% pullbacks or shift incremental capital to underowned cybersecurity and Chinese AI leaders if onshore compute capacity increases. Historical parallel: 1990s enterprise software consolidation—winners captured persistent SaaS rents after a painful regulatory and investment cycle.

Key Decisions for Investors

• Establish a 2–3% long position in NVDA within 1 month; add on dips of 8–15%; target +30% in 6–12 months, set a stop-loss at -20% from entry. Rationale: dominant GPU pricing power and >50% share of training/inference capacity.
• Allocate 1.5% long MSFT and 1.5% long GOOGL to capture cloud AI monetization over 12 months; add on pullbacks >10%, target 15–25% upside as enterprises shift to paid inference services.
• Initiate a 1–2% position in BIDU as a China-AI growth play within 3 months; reduce to zero if China announces export controls or new AI model-content restrictions within 60 days. Target 20–30% upside if domestic AI adoption accelerates.
• Add 1–2% exposure to cybersecurity (CRWD or PANW) within 2 months, buying on next 5–10% pullback; expected 15–25% upside in 9–12 months as AI misuse drives spend.
• Buy a 3-month put spread on SOXX (cost <2% portfolio) as insurance against a regulatory-driven semiconductor drawdown; widen strike selection to keep premium under 2% while capping downside between -15% and -40% sector moves.