Jane Street’s Trading Haul Juiced by Surging Bet on Anthropic

Jane Street Group told investors that its investments in private companies and funds boosted third-quarter trading revenue by roughly $830 million, helping drive a record haul for the firm this year. People familiar with the matter said Anthropic PBC — whose valuation has surged — accounted for the vast majority of those gains, highlighting how rising AI valuations can produce large mark‑ups for market‑making firms with private stakes and potentially influence allocation and risk posture toward AI venture assets.

Analysis

Market structure: Jane Street’s roughly $830m Q3 uplift concentrated in Anthropic signals a bifurcated market where liquidity and mark-to-model private gains reward market-makers, large VCs and AI infrastructure vendors. Direct winners: GPU and cloud providers (NVDA, MSFT, GOOGL) and prop/prime brokers capturing private-round flow; losers: undercapitalized small-cap AI developers and legacy CPU incumbents losing share. The immediate effect (days–weeks) is higher equity and options skew on AI names; over months the supply of public AI product will be constrained as companies stay private longer, supporting higher private valuations.

Risk assessment: Key tail risks include regulatory intervention (EU AI Act, U.S. hearings) and an investor sentiment snapback that could cause >30–50% private valuation markdowns in 3–12 months. Hidden dependency: Jane Street’s mark gains are concentrated (Anthropic “vast majority”) — a liquidity mismatch could force large write-downs if secondary markets dry up. Catalysts: Anthropic fundraises, NVDA/MSFT/GOOGL earnings, and policy moves in the next 30–180 days will materially re-rate exposures.

Trade implications: Prefer concentrated exposure to AI infrastructure over speculative model plays. Tactical: overweight NVDA and cloud operators via defined-risk options to capture 20–80% upside over 6–12 months while trimming speculative small-cap AI. Use pair trades (long NVDA, short INTC) and sell premium on crowded retail/ETF AI plays to monetize elevated implied vols and skew.

Contrarian angle: The market is understating liquidity risk in private marks — consensus may be over-allocating to “AI” without distinguishing revenue-generating infra from research bets. Historical parallel: 2018–2022 private-tech multiple compression shows private gains can reverse sharply when funding tightens. Opportunity: short or sell term premium on thematic AI ETFs/low-revenue AI names and reallocate to cash-flowing infra names before potential re-pricing within 3–9 months.

Key Decisions for Investors

• Establish a 2–3% portfolio long in NVIDIA (NVDA) via a 9–12 month call spread sized to risk ~1% of portfolio (buy ~30–40% OTM calls, sell ~60–80% OTM calls) to capture 20–80% upside while capping downside; add if NVDA corrects >10% intraday.
• Add 1–1.5% longs in Microsoft (MSFT) and Alphabet (GOOGL) each via equity or covered-call (6–12 month) structures to play cloud model deployment; reduce delta if either misses revenue guidance by >3% on next earnings.
• Implement pair trade: long NVDA (1.5–2% notional) and short Intel (INTC) (1% notional) to express GPU vs CPU secular divergence over 3–12 months; rebalance if spread outperforms/underperforms by >25% relative.
• Trim 30–50% of positions in AI-themed ETFs and low-revenue AI small caps (e.g., ARKK/BOTZ exposure) this quarter; redeploy proceeds into NVDA/MSFT/GOOGL or cash. Monitor Anthropic fundraising/earnings and EU/US AI regulatory moves over next 30–90 days — if funding stalls or legislation tightens, increase hedges and reduce private-equivalent exposure by another 20–30%.